Transplant

ABSTRACTS FRIDAY, FEBRUARY 21, 2014, 8:00AM–9:00AM PRESIDENT’S PLENARY PP.01 THE NEAR INFRARED TECHNIQUE FOR CHOLANGIOGRAPHY: 10 REASONS THAT SUPPORT THE SYSTEMATIC UTILIZATION OF THE METHOD F. Dip, M. Roy, C. Simpfendorfer, E. Lo-Menzo, S. Szomstein and R. Rosenthal Cleveland Clinic Florida, Weston, FLORIDA Introduction: We looked at the reasons why fluorescent cholangiography (FC) should be used routinely in laparoscopic cholecystectomy (LC). Method: A single dose of 0.05 mg/kg of Indocyanin Green (ICG) was administered intravenously one hour prior to the surgery to perform fluorescent cholangiograhy. Results: FC could be performed in all 45 (100%) patients whereas intra-operative cholangiography (IOC) could be performed in 42 out of 45 (93%) patients (p < 0.078). Individual median cost of performing FC was cheaper than IOC (13.97 ± 4.3 vs 778.43 ± 0.4 US dollars per patient, p = 0.0001). The mean operative time was 64.95 ± 17.43 minutes. FC was faster than IOC (0.71 ± 0.26 vs 7.15 ± 3.76 minutes, p < 0.0001). The cystic duct was identified by FC in 44 out of 45 patients (97.77 %). The residents were able to identify the extrahepatic structures in all 45 cases (100%) with FC. No complications were detected related to surgery and the use of FC. Learning curve was not necessary to identify structures using FC. X-ray leads were only used for IOC. FC could be performed by all residents at different level of training in 100% of the cases. Smooth dissection, transection and resection could be safely performed in 45 cases (100%). Conclusion: Fluorescent cholangiography seems to be feasible, cheap, expeditious, useful, an effective teaching tool, safe, no learning curve is necessary, does not require x-ray and easy to perform. It can be used for real time surgery to delineate the extrahepatic biliary structures.

PP.02 IN SEARCH OF THE BEST RECONSTRUCTIVE TECHNIQUE IN THE WHIPPLE OPERATION PANCREATICOJEJUNOSTOMY VERSUS PANCREATICOGASTROSTOMY. A RANDOMIZED CLINICAL TRIAL J. Grendar1, J. F. Ouellet2, F. Sutherland1, O. Bathe1, C. Ball1 and E. Dixon1 1 Department Of Surgery, University Of Calgary, Calgary, ALBERTA; 2Department Of Surgery, Quebec City University Hospital, Quebec, QUEBEC Background: Objective of this study is to compare rates of pancreatic fistulas and complications following the Whipple operation between pancreaticojejunostomy (PJ) and pancreaticogastrostomy (PG). HPB 2014, 16 (Suppl. 1), 1–99

Methods: 98 patients undergoing Whipple resection were randomized to either PG (48) or PJ (50) reconstruction. T-test and Chi-square tests were used for intention to treat data analysis. Logistic regression was used to measure the influence of surgical technique, preoperative ASA score and soft pancreatic gland on overall complications, severe post operative complications and overall fistula rates. Results: The rate of pancreatic fistula formation was 18% (Grade A = 6%, B = 10%, C = 2%) in the PJ arm and 25% (Grade A = 8%, B = 13%, C = 4%) in the PG arm, p = 0.399. The rate of postoperative complications was 48% (Clavien 1 = 14%, 2 = 36%, 3 = 10%, 4 = 0%, 5 = 2%) in the PJ and 58% (Clavien 1 = 21%, 2 = 38%, 3 = 25%, 4 = 6%, 5 = 4%) in the PG arm, p = 0.306. There was a significant difference in severe complications (Clavien 3–5) with 12% in the PJ and 31% in the PG arm, p = 0.02. In the multivariate analysis randomization (together with ASA) was only predictive of severe complications (OR 0.10, p < 0.005 for randomization to PJ reconstruction; OR 11.58, p < 0.05 for ASA 2 and OR 30.89, p < 0.05 for ASA 3 compared to ASA 1). Conclusion: Results of the study suggest that while there are no overall differences in rates of pancreatic leak/fistula and overall complications between PG and PJ arms, pancreaticogastrostomy is associated with a higher rate of severe post operative complications following the Whipple operation. This was also confirmed in the multivariate analysis.

PP.03 PROGNOSTIC ROLE OF PLASMA VASCULAR ENDOTHELIAL GROWTH FACTOR (VEGF) IN PATIENTS WITH HEPATOCELLULAR CARCINOMA UNDERGOING LIVER TRANSPLANTATION: EXPERIENCE OF A SINGLE WESTERN CENTER W. Zhang, C. Quintini, K. Hashimoto, M. Fujiki, T. Diago, D. Kelly, B. Eghtesad, C. Miller, J. Fung and F. Aucejo Hepatobiliary & Liver Transplant Surgery, Cleveland Clinic Foundation, Cleveland, OH Background: Vascular endothelial growth factor (VEGF) has an important role in hepatocellular carcinoma (HCC) development by fostering tumor cell proliferation and new tumor vessel formation. Recent studies demonstrated the prognostic value of serum/plasma VEGF levels in patients undergoing liver resection or locoregional therapies for HCC. Most studies have been generated in Asian countries, and limited data correlating serum/plasma VEGF in the setting of liver transplantation (LT) for HCC is available. This study investigated the prognostic significance of pre-LT plasma VEGF levels in patients with HCC undergoing LT in a large western transplant center. Methods: From January 2007 to December 2011, pre-LT plasma VEGF levels were measured by an enzyme-linked immunoassay in 113 patients with HCC. Median VEGF level of the entire cohort was used as the cutoff value to determine high and low plasma VEGF levels. Plasma VEGF levels were © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

2

Abstracts

correlated with clinicopathological characteristics and overall and recurrence-free post-LT survival. Results: Median pre-LT plasma VEGF levels were significantly associated with total tumor diameter >5 cm (p = 0.005), tumor vascular invasion (p = 0.000), pre-LT locoregional therapy (LRT) (p = 0.001) and beyond Milan criteria (p = 0.016) on univariate analysis. Median pre-LT plasma VEGF level >35 pg/ml was independently associated with vascular invasion (p = 0.006), worse overall (hazard ratio 2.693; p = 0.036) and recurrence-free patient survival (hazard ratio 2.590; p = 0.049) in comparison with median pre-LT plasma VEGF level <35 pg/ml on multivariate analysis. Conclusion: In patients with chronic end stage liver disease undergoing LT with HCC, median pre-LT plasma VEGF level >35 pg/ml appears to be an independent predictor of vascular invasion, overall and recurrence-free post-LT survival. Due to the current absence of accurate prognostic biomarkers regarding LT in the setting of HCC, further studies are warranted to confirm this observation.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

PP.04 THE EFFECT OF PREOPERATIVE RENAL INSUFFICIENCY ON POSTOPERATIVE OUTCOMES FOLLOWING MAJOR HEPATECTOMY: A MULTI-INSTITUTIONAL ANALYSIS OF 1170 PATIENTS M. H. Squires1, N. L. Lad1, S. B. Fisher1, D. A. Kooby1, S. M. Weber2, A. S. Brinkman2, C. R. Scoggins3, M. E. Egger3, K. Cardona1, C. S. Cho2, R. C. Martin3, M. C. Russell1, E. R. Winslow2, C. A. Staley1 and S. K. Maithel1 1 Department Of Surgery, Division Of Surgical Oncology, Winship Cancer Institute, Emory University, Atlanta, GA; 2 Department Of Surgery, University Of Wisconsin, Madison, WI; 3Division Of Surgical Oncology, Department Of Surgery, University Of Louisville, Louisville, KY Background: Renal insufficiency is known to adversely affect outcomes after cardiac and vascular surgery. The effect of preoperative renal insufficiency on postoperative outcomes following major hepatectomy is unknown. Methods: All patients who underwent major hepatectomy, defined as resection of ≥3 hepatic segments, at three academic institutions from 2000–2012 were identified. Resections were performed utilizing low CVP technique unless contraindicated. Renal function was analyzed by preoperative serum creatinine (sCr) level. Primary outcomes were major complications (Clavien Grade III-V), respiratory failure, acute renal failure (ARF) requiring new-onset hemodialysis, and 90-day mortality. Multivariate models for each endpoint were constructed using all variables with a p-value < 0.05 on univariate analysis. Results: 1170 patients with preoperative sCr levels available for analysis were identified. Renal function was analyzed using sCr dichotomized at 1.8 mg/dL, one standard deviation above the mean value (0.97 +/− 0.79 mg/dL) for the cohort. Twenty-three patients had sCr ≥ 1.8 mg/dL. Major complications occurred in 279 patients (23.8%), respiratory failure in 62 (5.3%), and ARF in 31 (2.6%). Ninety-day mortality rate was 5.4%. On univariate analysis, sCr ≥ 1.8 was significantly associated with each endpoint. On multivariate analysis, patients with sCr ≥ 1.8 remained at significantly increased risk for major complications (HR 3.94;95%CI:1.48–10.49; p = 0.006), respiratory failure (HR 4.43;95%CI:1.33–14.80; p = 0.015), and ARF (HR 4.63;95%CI:1.16–18.48; p = 0.030). Serum Cr ≥ 1.8 was not independently associated with 90-day mortality on multivariate analysis (p = 0.27). Conclusion: Preoperative serum creatinine ≥1.8 mg/dL identifies patients at significantly increased risk of postoperative major complications, respiratory failure, and acute renal failure requiring dialysis. Patients are well-selected for major hepatectomy, and few patients with significant renal insufficiency are deemed operative candidates.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

3

bile duct coursed over the right posterior bile duct before inserting separately from the common hepatic duct confluence into the choledochal cyst. The common bile duct was dissected to the level of the pancreatic parenchyma and then ligated using hem-o-lok clips (Teleflex Medical, Research Triangle, NC). The hepaticoduodenostomy was constructed using 3-0 vicryl sutures with a running back wall and an interrupted front wall. The patient progressed well postoperatively and was discharged home post-operative day four after an uncomplicated hospital course. Conclusion: Robotic resection of choledochl cyst with hepaticoduodenostomy is a feasible technique that confers the advantages over open technique such as fine movements, superior visualization and improved post-operative pain and return to function.

VL.02 ROBOTIC SEGMENTAL DUODENAL RESECTION R. M. Seshadri, D. J. Niemeyer, R. Z. Swan, D. Sindram, D. A. Iannitti and J. B. Martinie Carolinas Medical Center, Charlotte, NORTH CAROLINA

FRIDAY, FEBRUARY 21, 2014, 1:00PM–2:00PM LUNCH VIDEO PRESENTATIONS VL.01 ROBOTIC RESECTION OF CHOLEDOCHAL CYST WITH HEPATICODUODENOSTOMY S. G. Warner and K. T. Nguyen University Of Michigan, Division Of HPB And Advanced GI Surgery, Ann Arbor, MI Introduction: Resection of the common bile duct is the treatment of choice for type I choledochal cysts found in adults. The procedure can portend a significant morbidity because of the large incision required for adequate exposure. Robotic surgery provides excellent visualization and fine manipluation of the biliary system with minimized morbidity to the patient. Methods: A 26 yr old female required excision of large type I choledochal cyst. The Da Vinci Si Robot System (Intuitive Surgical, Sunnyvale, CA) was utilized to assist resection. Results: Port placement was similar to standard robotic cholecystectomy ports, with slight alterations to facilitate possible roux-en-Y hepaticojejunostomy in case a hepaticoduodenostomy was not possible. Operative time was less than 210 minutes, estimated blood loss was 100 mL. Aberrant anatomy was identified in that the right anterior HPB 2014, 16 (Suppl. 1), 1–99

Background: Duodenal polyps can have malignant potential and hence warrant a biopsy and subsequent resection if needed. Most of them can be done endoscopically but some warrant a surgical resection. Introduction: We present a 69 year old female with a 3 cm intramural lesion on the lateral wall of the 2nd portion of the duodenum. Her primary clinical presentation was upper GI bleeding. The endoscopic biopsy was consistent with a myofibroblastic lesion with ulceration. We decided to proceed with a robotic sleeve resection of the lesion. Method: After trocar placement, the hepatic flexure of the colon was taken down using monopolar scissors to visualize the duodenum. A complete kocker maneuver was performed and the lesion was identified. Intraoperative ultrasound was performed to confirm the location. We used the monopolar electrocautery scissors to make a duodenotomy along the anteromedial aspect. We then proceeded to create an ellipse around the polyp until it completely free of the remaining duodenum. Frozen section was consistent with a benign inflammatory fibrous polyp. The segmental duodenal resection was closed using running lambert sutures using two 40 V LOC suture. At completion we checked for hemostasis and a 19 French blake drain was placed over the area of the repair. The patient tolerated the procedure well and was discharged 3 days later. Conclusion: We present a novel minimally invasive approach for segmental duodenal resection that would decrease the overall morbidity and mortality of the procedure and promote early recovery.

VL.03 LAPAROSCOPIC RIGHT HEPATECTOMY WITH HYDROJET AND HARMONIC SCALPEL DISSECTION J. B. Conneely, R. Smoot and S. Cleary Toronto General Hospital, Tornto, ONTARIO The increasing availability of laparoscopic energy devices for vessel sealing and division has been a boon to laparoscopic surgeons worldwide. Liver resection in particular has been © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

4

Abstracts

aided by the improved performance and evolving ergonomics of these devices. The Hydrojet dissector has proven to be of enormous benefit in open liver resection. It allows precise parenchymal dissection and clear exposure of even the smallest vessels and biliary radicles. Its adoption in open surgery has been somewhat limited, with the established technologies of CUSA, bipolar electrocautery and ultrasonic technology enjoying more widespread adoption. In laparoscopic liver surgery, few centres routinely employ the Hydrojet. Our institution routinely employs Hydrojet for live donor hepatectomy and as such we have extensive experience with the device. We have employed Hydrojet regularly for laparoscopic hepatectomy and describe herein our technique of parenchymal division during major hepatectomy, combining the benefits of the Hydrojet and the Harmonic Scalpel. We demonstrate the precise parenchymal dissection made possible by the laparoscopic Hydrojet. We propose that this technique of liver resection is safe, efficient and precise.

FRIDAY, FEBRUARY 21, 2014, 2:55PM–4:55PM PARALLEL LONG ORAL ABSTRACT A LIVER LO-A.01 LAPAROSCOPIC HEPATECTOMY SIGNIFICANTLY SHORTENS THE TIME TO POSTOPERATIVE CHEMOTHERAPY IN OBESE PATIENTS UNDERGOING MAJOR HEPATECTOMIES S. C. Agle, P. Philips, M. E. Egger, C. R. Scoggins, K. M. McMasters and R. C. Martin University Of Louisville, Louisville, KY Background and Objectives: The multimodality management of resectable hepatic malignancies commonly utilizes chemotherapy either in a pre-operative/post-operative or post-operative delivery. The benefit of post-operative chemotherapy occurs with early initiation, but is commonly delayed due to procedural related complications. Minimally invasive surgery in the obese population is known to significantly reduce complications, length of stay (LOS) and earlier to return to baseline. The aim of this study was to compare open versus laparoscopic surgical approaches to major liver resection in obese patients. Methods: We performed a single center retrospective review of 118 consecutive patients with a BMI ≥25, a diagnosis of cancer, underwent a major hepatectomy (>3 segments) from 2007 through 2012 with planned post-operative chemotherapy. Results: A matched group of 42 laparoscopic were compared to 76 open cases having equivalent resections based on tumor location and diagnosis. There were no differences in the past medical/surgical history, surgical procedure, and 30-day mortality. The open cohort had more blood loss (475 vs 250 ml p = <0.001), longer LOS (Median 6.5 vs 4.0 days, p = <0.001) and a prolonged time to post-operative chemotherapy initiation (37 vs 24 days, p = <0.001). On multivariate analysis laparoscopic resection was found to be an © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

independent predictor of a shorter LOS and earlier initiation of postoperative chemotherapy. Conclusion: The benefits of laparoscopic major hepatectomies in the obese population included a shorter length of stay and earlier initiation of chemotherapy compared to the open group. These benefits were seen without jeopardizing surgical margins or extent of resection with associated improved outcomes.

LO-A.02 SURVIVAL ANALYSIS OF PATIENTS WITH STAGE I AND II HEPATOCELLULAR CARCINOMA FOLLOWING LIVER TRANSPLANTATION AND LIVER RESECTION R. M. Seshadri, D. J. Niemeyer, R. Z. Swan, D. Sindram, J. B. Martinie, M. Russo and D. A. Iannitti Carolinas Medical Center, Charlotte, NORTH CAROLINA Introduction: Liver transplantation (LT) is a treatment option in select patients with hepatocellular carcinoma (HCC). We compared survival in patients with HCC who had a transplant versus liver resection based on AJCC clinical stage. Method: The study is a retrospective analysis from the Commission on Cancer’s National Cancer Data Base (NCDB) between the years 1998–2011. 148,882 patients with liver cancer were identified, of which 126,858 had HCC. 64,227 patients from 1998–2006 had 5-year survival data. Patients were stratified by AJCC clinical stage I and II, and resection was compared to LT. The study was limited to stage I and II patients as most would fit Milan Criteria for transplantation. Kaplan-Meier curves and log rank tests were used for statistical analysis. Results: 3,194 HCC patients met the analysis criteria. Mean age in the LT group and resection group was 56 and 63 respectively. Among stage I HCC, 839 had LT and 823 had a liver resection. Among stage II HCC, 806 had LT and 726 had liver resection. In the stage I group, median survival for LT and resection was 127.9 and 54.1 months respectively (p < 0.0001), and in the stage II group the median survival was 111.8 and 41.8 months (p < 0.0001). Resection patients with stage I HCC had a longer median survival than stage II patients (p = 0.0002), however there was no significant difference in the LT group. Conclusion: Liver transplantation offers a survival advantage compared to resection among patients with AJCC clinical stage I and II HCC.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

LO-A.03 VENOUS THROMBOEMBOLIC PROPHYLAXIS FOLLOWING HEPATIC RESECTION: PATTERNS OF CARE AMONG LIVER SURGEONS M. Weiss, A. Ejaz, G. Spolverato, Y. Kim, K. Hirose, C. Wolfgang, M. Choti and T. M. Pawlik Johns Hopkins Hospital, Baltimore, MD Introduction: Consensus does not exist around the utilization of venous thromboembolic(VTE)-prophylaxis for patients undergoing hepatic resection. We sought to define what clinical factors impact choice of VTE-prophylaxis among hepato-pancreato-biliary (HPB) surgeons. Method: Surgeons who perform liver resections completed a web-based survey regarding VTE-prophylaxis. Impact of

HPB 2014, 16 (Suppl. 1), 1–99

5

physician and clinical factors on VTE-prophylaxis was analyzed. Results: 204 responses were received (response rate-67%). Most respondents practiced at an academic center (87%); surgical training varied: HPB (24%), transplantation (25%), surgical oncology (33%), combined HPB/transplantation (13%), or no specialty training (5%). Most respondents (57%) estimated the risk of VTE at 4–9%. Although 98% of respondents used VTE-prophylaxis, there was considerable variability: sequential-compression devices (89%), unfractionated heparin Q12hr (31%), unfractionated heparin Q8hr (30%), and low-molecular weight heparin (37%). While 88% of respondents noted their use of VTEprophylaxis was not impacted by operative indication (benign vs. malignant), 16% stated that major resection reduced their likelihood of administering prophylaxis (Figure). Other factors associated with decreased likelihood of administering pharmacologic-prophylaxis included: elevated INR (75%), thrombocytopenia (63%), liver insufficiency (58%), and large blood loss (45%). 47% of respondents routinely waited until ≥POD1 before initiating pharmacologic-prophylaxis and 35% hold VTEprophylaxis until there are no signs of coagulopathy.Aminority of respondents (14%) said they routinely send patients home on prophylaxis. Surgical oncology fellowship trained providers were more likely to wait until ≥POD1 (OR 2.51) as well as hold prophylaxis until no signs of coagulopathy (OR 3.92) (both P < 0.05). Conclusion: There is considerable practice variation regarding VTE prophylaxis among liver surgeons. While the majority of surgeons routinely employ VTE prophylaxis, the methods, timing and purported contraindications to VTE prophylaxis differ significantly among liver surgeons.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

6

Abstracts

LO-A.04 HYPOPHOSPHATEMIA AFTER MAJOR HEPATECTOMY AND THE RISK OF POSTOPERATIVE HEPATIC INSUFFICIENCY AND MORTALITY: AN ANALYSIS OF 719 PATIENTS M. H. Squires1, G. C. Dann1, N. L. Lad1, S. B. Fisher1, B. M. Martin1, D. A. Kooby1, J. M. Sarmiento2, M. C. Russell1, K. Cardona1, C. A. Staley1 and S. K. Maithel1 1 Department Of Surgery, Division Of Surgical Oncology, Winship Cancer Institute, Emory University, Atlanta, GA; 2 Department Of General And Gastrointestinal Surgery, Emory University, Atlanta, GA Background: Hypophosphatemia after hepatectomy correlates with extent of liver resected and suggests hepatic regeneration. We hypothesized that following major hepatectomy, absence of hypophosphatemia is associated with poor regeneration and increased postoperative hepatic insufficiency (PHI) and complications. Methods: Patients who underwent major hepatectomy (≥3 segments), from 2000–2012 at a single institution were identified. Postoperative serum phosphorus levels were assessed. Primary outcomes were PHI (peak bilirubin >7 mg/dL), major (Clavien Grade III-V) complications, and 30- and 90-day mortality. Univariate and multivariate regression analyses were performed for each endpoint. Results: Of 749 major hepatectomy cases, 719 patients had postoperative phosphorus levels available for analysis. PHI and major complications occurred in 63 (8.8%) and 169 (23.5%) patients, respectively. 30- and 90-day mortality rates were 4.0% and 5.4%, respectively. Average phosphorus level on postoperative day (POD) 2 was 2.4 mg/dL; 201 patients (28.0%) had phosphorus >2.4 on POD2. Patients with POD2 phosphorus >2.4 had a significantly higher incidence of PHI, major complications, and mortality (Figure). On multivariate analysis, POD 2 phosphorus >2.4 remained a significant risk factor for PHI (HR:1.78, 95% CI: 1.02–3.17, p = 0.048), major complications (HR: 1.57, 95% CI: 1.02–2.47, p = 0.049), 30-day mortality (HR: 2.70, 95% CI: 1.08–6.76, p = 0.034), and 90-day mortality (HR: 2.51, 95% CI: 1.03–6.15, p = 0.044). Similarly, patients whose phosphorus level reached nadir after POD3 had a significantly higher incidence of PHI, major complications, and mortality (Table). Conclusion: Persistently elevated phosphorus levels >2.4 mg/dL on POD2 and a delayed nadir in phosphorus beyond POD3 are significantly associated with increased postoperative hepatic insufficiency, major complications, and early mortality. Failure to develop hypophosphatemia within 72 hours after major hepatectomy may reflect insufficient liver remnant regeneration.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

LO-A.05 BORDERLINE OPERABILITY IN HEPATECTOMY PATIENTS IS ASSOCIATED WITH HIGHER RATES OF FAILURE TO RESCUE AFTER SEVERE COMPLICATIONS C. D. Tzeng1, A. B. Cooper2, S. A. Curley2, J. N. Vauthey2 and T. A. Aloia2 1 University Of Kentucky, Lexington, KY; 2UT MD Anderson Cancer Center, Houston, TX Introduction: The influence of age and comorbidities on post-hepatectomy complications is not clear. To address this knowledge gap, this study analyzed the incidence and risk factors for post-hepatectomy morbidity/mortality in patients with “borderline” operability (BL). Methods: 7,184 elective hepatectomies were identified in the 2005–10 ACS-NSQIP database. Perioperative variables were compared for 30-day morbidity/mortality in BL patients (defined as age ≥ 75, poor performance status, lung disease, ascites/varices, recent myocardial infarction, stroke, steroids, weight loss > 10%, and/or sepsis). Results: 1,720 (23.9%) patients met the BL definition. Despite non-BL/BL patients undergoing hepatectomies of similar extent (p > 0.2), BL patients had higher rates of severe complications (SC, 25.2% vs. 16.2%) and overall mortality (4.2% vs. 1.4%, p < 0.001). BL patients with any SC experienced a 16.9% mortality rate (vs. 8.5% in non-BL, p < 0.001). They frequently experienced multiple SC (54.5% ≥ 2 SC; 34.9% ≥ 3 SC), and the mortality disparity was more pronounced with ≥2 and ≥3 SC (26.7% vs. 17.2%; 39.1% vs. 28.8%, p < 0.001). Multivariate analysis identified the following risk factors for SC in BL patients: anesthesia score >3 (odds ratio, OR-2.02), diabetes (OR-1.58), smoking (OR-1.64), albumin < 3.5 g/dL (OR-1.50), intraoperative transfusion (OR-2.36), operative time >240 min (OR-1.77), and concurrent procedure (OR-2.81, all p < 0.01). Independent predictors of mortality included albumin < 3.5 g/dL (OR2.92), thrombocytopenia (OR-2.97), and extended/right hepatectomy (OR-2.16, all p < 0.01). Conclusion: BL patients are more likely to experience posthepatectomy SC. With less physiologic reserve, they are less able to be clinically rescued. To improve outcomes, surgeons should preoperatively optimize comorbidities and limit procedural magnitude, specifically limiting longer, more complex resections with higher transfusion risk. HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

7

LO-A.07 ADJUVANT STEREOTACTIC BODY RADIATION FOLLOWING CHEMOEMBOLIZATION IMPROVES SURVIVAL IN PATIENTS WITH NON-RESECTABLE HEPATOCELLULAR CARCINOMA OVER 3 CENTIMETERS R. Jacob, K. Keene, F. Turley, D. T. Redden, M. K. Bryant, D. P. Dorn, L. Dover, J. White, S. H. Gray, D. E. Eckhoff and D. A. DuBay University of Alabama at Birmingham, Birmingham, AL

LO-A.06 RENAL FUNCTION AFTER LOW CENTRAL VENOUS PRESSURE ASSISTED LIVER RESECTION. ASSESSMENT OF 2116 CASES C. Correa-Gallego, A. Berman, S. C. Denis, V. Arslan-Carlon, T. Kingham, M. I. D’Angelica, P. J. Allen, Y. Fong, R. P. DeMatteo, W. R. Jarnagin, J. Melendez and M. Fischer Memorial Sloan-Kettering Cancer Center, New York, NY Background: Low central-venous pressure (LCVP) assisted hepatectomy is associated with decreased blood-loss and lower transfusion rates. Concern about its impact on renal function have prevented wide-spread application. We reviewed the dynamics of renal function after LCVP-assisted hepatectomy. Methods: Retrospective analysis of prospective surgical database. GFR was calculated using the Cockcroft-Gault equation and its variation recorded over time. RIFLE criteria were used to define biochemical postoperative acute renal failure (ARF). (RIFLE: Risk: 25–50% GFR decrease, Injury: 51–75%, Failure: >75%, Loss/End-stage: irreversible failure >4 weeks). Clinically-relevant ARF was queried from our postoperative database. Results: Between 2003–2012, 2116 LCVP assisted hepatectomies were identified. Median age was 59 yrs (IQR: 50–69), 50% were male, and 53% had ASA ≥3. Median number of resected segments was 2 (1–4) and median estimated blood-loss was 300 ml (200–600). Morbidity and 90-day mortality were 21% and 2%. Low baseline GFR (<90 ml/min) was seen in 64% (median 67 ml/min; IQR: 55–79). Biochemically, 16% (329/2116) had evidence of postoperative ARF (85% Risk, 14% Injury, and 1% Failure), of which 198 normalized by the time of discharge. Patients presenting with low baseline GFR were less likely to develop biochemical ARF: 11% vs 24% (P < .001). Only nine patients (<1%) had documented clinical evidence of isolated renal-dysfunction, none requiring dialysis. Conclusion: Biochemical alterations in GFR are transient in the vast majority of patients after LCVP hepatectomy, and its clinical impact is limited. Our data suggest that clinically relevant renal dysfunction (exclusive of multiple organ system failure) is an uncommon event in patients undergoing LCVP liver resection.

HPB 2014, 16 (Suppl. 1), 1–99

Background: The optimal locoregional treatment for nonresectable hepatocellular carcinoma (HCC) >3 cm is unclear. Chemoembolization (TACE) is the most common initial intervention performed, but rarely completely eradicates the HCC. The purpose of this study is to measure survival in HCC patients treated with adjuvant stereotactic body radiation (SBRT) following TACE. Methods: Patients with HCC >3 cm treated with TACE alone (n = 124) were compared with those treated with TACE/SBRT (n = 37). Standard survival analysis was performed. Results: There were no significant differences between the 2 groups with regard to age, gender, race, or liver disease etiology. The MELD scores were not different between groups (TACE: 11.2 ± 4.2 vs. TACE/SBRT: 10.2 ± 3.6, p = 0.21) nor were measures of sarcopenia (TACE: 10.6 ± 3.4 vs. TACE/SBRT: 11.2 ± 3.8, p = 0.25). The preTACE mean number of tumors (TACE: 2.1 ± 1.6 vs. TACE/ SBRT: 1.8 ± 1.1, p = 0.57), largest tumor size (TACE: 5.8 cm ± 3.0 vs. TACE/SBRT: 6.1 cm ± 2.4, p = 0.09), total tumor diameter (TACE: 7.7 cm ± 4.9 vs. TACE/SBRT: 7.8 ± 3.3, p = 0.21) and median alpha feto-protein level (TACE: 19.8 ± 192.1 vs. TACE/SBRT: 32.7 ± 456.0, p = 0.53) were not significantly different between groups. Modified response evaluation criteria in solid tumors (mRECIST), measured after TACE (but before SBRT administration), were not different between groups (p = 0.69). After censoring for liver transplantation, the overall survival was significantly increased in the TACE/SBRT group compared to the TACE only group (33 vs. 20 months, p = 0.02, Figure). Conclusion: Adjuvant SBRT following TACE improves survival in patients with HCC greater than 3 cm.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

8

Abstracts

LO-A.08 SURGICAL MANAGEMENT OF HEPATIC HEMANGIOMAS: AN INTERNATIONAL MULTI-INSTITUTIONAL EXPERIENCE

LO-A.09 PREDICTING TRANSPLANT-ELIGIBLE RECURRENCE AFTER INITIAL RESECTION OF HEPATOCELLULAR CANCER

J. T. Miura1, A. Li2, R. Schmocker4, S. Nichols5, D. Sukato7, E. Winslow4, G. Spolverato6, A. Ejaz6, M. H. Squires3, D. Kooby3, S. Maithel3, J. M. Sarmiento3, M. Bloomston5, K. K. Christians1, F. M. Johnston1, S. Tsai1, K. Turaga1, A. Tsung7, T. Pawlik6, M. Wu2 and T. C. Gamblin1 1 Department Of Surgery, Medical College Of Wisconsin, Milwaukee, WI; 2Eastern Hepatobiliary Surgery Hospital, The Second Military Medical University, Shanghai, SHANGHAI; 3Department Of Surgery, Emory University School Of Medicine, Atlanta, GA; 4Department Of Surgery, University Of Wisconsin School Of Medicine And Public Health, Madison, WI; 5Department Of Surgery, The Ohio State University, Columbus, OH; 6Department Of Surgery, Johns Hopkins University School Of Medicine, Baltimore, MD; 7Department Of Surgery, University Of Pittsburgh School Of Medicine, Pittsburgh, PA

S. Lee, I. Konstantinidis, T. P. Kingham, M. I. DAngelica, P. J. Allen, Y. Fong, R. P. DeMatteo and W. R. Jarnagin Memorial Sloan-Kettering Cancer Center, New York, NY

Introduction: Management of hepatic hemangiomas remains ill defined. We sought to investigate the indications, surgical management, and outcomes of patients with hemangiomas. Methods: A retrospective review from 7 major liver centers identifying patients who underwent surgery for hepatic hemangiomas between 1994 and 2013 was performed. Clinico-pathologic, treatment and perioperative data were evaluated. Results: Of 1,899 patients identified, the median age was 45 years (IQR: 39–52) with patients being predominantly female (69.9%). Majority of hemangiomas were solitary lesions (54.8%) with a median size of 8 cm (IQR: 6–10). Surgery was performed for abdominal symptoms (19%), diagnostic uncertainty (6%), increasing hemangioma size (31%), and patient preference (44%). Patients undergoing resection at western centers (W) were more likely to have symptomatic hemangiomas as compared to the eastern center (E) (66.4 vs 9.7%, p < 0.01). Surgical resection requiring a hemihepatectomy or greater was performed on 535 patients (28.1%). Median length of stay was 13 days (W: 5 vs E: 14 days, p < 0.01). Clavien Grade 3 or higher complications occurred in 43 patients (2.3%). The 30 day mortality was 0.2% (n = 4) with no additional mortalities at 90 days. Final pathology revealed 13 patients (1%) harbored a concomitant malignancy within the resected hemangioma specimen; eight of the 13 patients underwent surgery secondary to radiographic uncertainty. Of patients with abdominal symptoms at the time of surgery, 75% reported improvement of symptoms postoperatively. Conclusion: Surgery for hemangiomas can be safely performed at high volume institutions. Variable indications for liver surgery demonstrate a regional difference in approach to management.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

Aim: Initial(IT) and salvage transplantation (ST) are effective for hepatocellular carcinoma (HCC) within the Milan criteria (MC). Reliable prediction of post-resection recurrence patterns might allow better prioritization of patients for transplantation. This study analyzes rates of transplantable recurrence after complete resection of HCC, stratified by preoperative MC status and histopathologic factors. Methods: From January 1992 to January 2012, 320 patients underwent partial hepatectomy with negative margins at Memorial Sloan-Kettering Cancer Center. Patients were categorized preoperatively as within- or beyond-MC, as were recurrences. Clinicopathological and perioperative parameters were analyzed to identify prognostic factors to predict recurrence pattern. Results: A total of 304 patients were analyzed. After median follow-up of 37 months, 180 patients (59%) developed recurrence. Patients with tumors within-MC had better survival outcomes than those beyond-MC (p < 0.001). In patients who developed recurrence, 71(39%) recurred within-MC and 109 (61%) recurred beyond-MC (Table 1). Of those beyond-MC before resection, a significant proportion (21%) recurred within-MC. Conversely, 19% of patients initially within-Milan recurred beyond-Milan (Figure 1). Age, gender, previous major/prior liver resection, tumor differentiation and cirrhosis were not significant predictors if recurrence was within- or beyond-MC. Significant factors that determined recurrence within MC included blood loss during liver resection (p = 0.007), alpha-fetoprotein (<200 ng/ml) (p = 0.009), tumor size(p = 0.0004), lymphovascular invasion (p = 0.02), microsatellites (p = 0.003) and pre-operative MC status (p = 0.003) (Table 1). Conclusion: After curative hepatectomy, a significant proportion of patients initially beyond-MC experience recurrence eligible for ST, while a similar proportion of patients initially within MC are transplant ineligible after recurrence. Incorporation of histopathologic parameters to the MC status may allow better identification of patients for IT and ST.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

9

Conclusion: Patients with sCRLM managed with either a staged or simultaneous approach had comparable perioperative and long-term outcomes. However, patients treated with a simultaneous approach spent significantly fewer days in the hospital and had a median cost savings of $30,420. When appropriate, the simultaneous approach for sCRLM should be preferred.

FRIDAY, FEBRUARY 21, 2014, 2:55PM–4:55PM PARALLEL LONG ORAL ABSTRACT B PANCREAS LO-B.01 PATIENT SELECTION AND THE VOLUME EFFECT IN PANCREATIC SURGERY: UNEQUAL BENEFITS? LO-A.10 SYNCHRONOUS PRIMARY COLORECTAL CANCER AND LIVER METASTASIS: IMPACT OF OPERATIVE APPROACH ON CLINICAL OUTCOMES AND HOSPITAL COSTS A. Ejaz, G. Spolverato, E. Semenov, D. Tanner, J. Hundt and T. M. Pawlik Johns Hopkins Hopsital, Baltimore, MD Introduction: Management of patients with colorectal cancer (CRC) and synchronous colorectal liver metastasis (sCRLM) remains controversial. We examined the clinical and economic impact of managing sCRLM with a staged versus simultaneous approach. Method: We identified 229 patients treated for sCRLM between 1990–2012. Clinicopathologic features, perioperative outcomes, and total hospital costs(inflationadjusted) were collected and analyzed. Results: 118 (52%) patients underwent a staged approach while 111 (48%) had a simultaneous CRC and liver operation. Overall intra-hepatic tumor burden was similar among patients treated with staged versus simultaneous approach (largest metastasis: 3.2 cm vs. 1.9 cm; number of metastasis: 2 vs. 2; both P > 0.05). Liver-directed therapy included hepatectomy (73%), ablation alone (3%) or combined resection + ablation (24%). A major hepatectomy (>3 segments) was more common with a staged approach (54% vs. 31%;P < 0.001); 43% of patients underwent an extended hepatic resection(staged, n = 64 vs. simultaneous, n = 35). There were 68 (30%) patients who had a post-operative complication with no difference in the staged (38%) vs. simultaneous (27%) groups (P > 0.05). 90-day mortality was 1.7% (staged, 1.7% vs. simultaneous, 1.8%; P = 0.94). Median total length of hospitalization was longer in the staged (12.5 days) versus simultaneous(7.0 days) group (P < 0.001). Median total hospital costs were also higher among patients undergoing a staged approach ($57,821) compared with a simultaneous operation ($27,401) (P < 0.01). Long-term outcomes were similar(5-year disease-free survival: staged, 39.8% vs. simultaneous, 48.6%; P = 0.61). HPB 2014, 16 (Suppl. 1), 1–99

L. A. Bliss1,3, C. J. Yang1, K. Lagisetty1, Z. Chau2, S. Ng1, T. S. Kent1, A. J. Moser1, D. W. McFadden3, M. P. Callery1 and J. F. Tseng1 1 Beth Israel Deaconess Medical Center, Boston, MA; 2 University Of Massachusetts Medical School, Worcester, MA; 3University Of Connecticut Health Center, Farmington, CT Background: The volume effect in pancreatic surgery is well established. Regionalization to high-volume centers has been proposed. The effect of volume-based proposals on practice patterns is unknown. Methods: Retrospective review of patients undergoing pancreatectomy in Nationwide Inpatient Sample (NIS) 2004– 2011. Inpatient mortality and complication rates calculated. Patients stratified by annual center volume (low <5, medium 5–18, high >18) based on prior publication. Multivariable regression model evaluated predictors of morbidity, mortality and resection at a high-volume center. Results: 129,609 patients underwent pancreatectomy. Crude inpatient mortality rate was 4.3%, decreasing over 2004–2011 (5.2% to 3.1%, p < 0.0001). 36.0% experienced complications. 65.2% underwent resection at high-volume centers. In 2004, low-, medium-, and high-volume centers resected 16.3%, 26.8%, and 56.9% of patients, compared with 7.65%, 20.7%, and 71.7% in 2011. High-volume centers had lower mortality (3.2% versus 6.5%, p < 0.0001), fewer complications (33.0% versus 41.6%, p < 0.0001) and shorter median length of stay (9 versus 11 days, p < 0.0001). However, patients at medium- or low-volume centers had more co-morbidities (40.9% versus 37.2% with Elixhauser >2, p = 0.0020), lower rates of private insurance (40.0% versus 43.8%, p = 0.0030) and more non-elective admissions (30.6% versus 17.8%, p < 0.0001). Conclusion: Overall, mortality after pancreatectomy continues to improve. Despite a shift to high-volume hospitals, a substantial cohort still receives care outside of these centers. Patients receiving non-high-volume care demonstrate less favorable comorbidities, insurance, and urgency of operation. The implications are twofold: already disadvantaged patients may not benefit from a high-volume effect; and patients predisposed to do well may contribute to observed superior outcomes at high-volume centers. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

10

Abstracts

LO-B.02 REDUCING READMISSIONS FOLLOWING PANCREATECTOMY: COORDINATION OF THE CARE CONTINUUM E. P. Ceppa1, H. A. Pitt2, A. Nakeeb1, C. M. Schmidt1, N. J. Zyromski1, M. G. House1, B. Brand1, D. Weidert1 and K. D. Lillemoe3 1 Indiana University School Of Medicine, Indianapolis, IN; 2 Temple University Health System, Philadelphia, PA; 3 Massachusetts General Hospital, Bostom, MA Background: Recent analyses of major gastrointestinal operations document that postoperative complications are a key driver of readmissions. However, successful programs to reduce readmissions after major operations have not been reported. This analysis documents a series of steps that were implemented to reduce complications and readmissions following pancreatectomy. Methods: From July 2007 through June 2012, the 30-day readmission rate for patients undergoing proximal, distal and total pancreatectomy was measured. Length of stay and mortality indices were risk-adjusted by the University Health Consortium. In 2008, a multifactorial effort was undertaken which resulted in fewer surgical site infections. In 2010, a Readmissions Team of surgeons, nurses, advanced providers, pharmacists, social workers and care managers was formed. Discharges with home care were increased from 20 to 50%, and relationships with post-acute care facilities were © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

strengthened. In 2011, “Project RED” (ReEngineering Discharges) was adopted and included a “discharge coach” to assure that patients were ready for discharge and had follow-up care. Statistical analysis included control charts and Fischer’s Exact tests. Results: Over five years, 1,147 patients underwent proximal (69%), distal (26%) or total pancreatectomy (5%). The mortality index (observed/expected rate) was 0.73 and did not vary significantly over time. The length of stay (LOS) index (observed/expected days) and 30-day all-cause readmission rates are presented in the table. Conclusion: All-cause readmissions following pancreatectomy can be reduced without altering length of stay. Efforts by surgeons to reduce complications as well as care coordination by nurses, advanced providers, pharmacists, social workers, care managers and a “discharge coach” are required to reduce readmissions.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

LO-B.03 PATENCY RATES OF PORTAL VEIN/SUPERIOR MESENTERIC VEIN RECONSTRUCTION AFTER PANCREATECTOMY FOR PANCREATIC ADENOCARCINOMA A. N. Krepline, K. Duelge, K. K. Christians, B. George, P. Ritch, B. Erickson, A. Mahmoud, E. J. Quebbeman, K. K. Turaga, F. M. Johnston, T. C. Gambliin, D. B. Evans and S. Tsai Medical College Of Wisconsin, Milwaukee, WI Background: The frequency of venous reconstruction (VR) to facilitate a complete resection of pancreatic adenocarcinoma (PC) has increased. Few studies have examined the long-term patency of the SMV-PV after reconstruction. Methods: From 2007–2013, patients (pts) with resectable and borderline resectable PC who underwent resection with VR were identified and classified by type of reconstruction. Postoperative imaging, performed at 3-month intervals, was reviewed to assess patency of VR from the time of surgery to the date of last follow-up. Demographic and clinicopathologic data were abstracted. Results: VR was performed in 47 pts and the median follow-up was 22 months. Techniques for VR and repair included: tangential resection with primary repair (9 [19.1%]) or saphenous vein patch (SVP; 11 [23.4%]), segmental resection with primary anastomosis (15 [31.9%]), or internal jugular vein interposition (IJV;12 [25.5%]). Following hospital discharge, 46 (97.7%) pts received prophylactic anticoagulation or anti-platelet therapy including lowmolecular-weight heparin (n = 4, 8.7%), warfarin (n = 3, 6.5%), 81 mg aspirin (n = 8, 17.4%), or 325 mg aspirin (n = 31, 61.7%). Of the 47 pts, 4 (8.5%) developed thrombosis of the SMV-PV; 2 pts required segmental resection with primary anastomosis (1) or interposition grafting (1) and received prophylactic LMWH and 2 pts had SVP and received 325 mg aspirin. Time to development of thrombosis ranged from 16–238 days after surgery with 2 pts developing thrombosis 2 and 3 months after surgery. Conclusion: The majority of venous reconstructions remain patent at a median of 2 year follow-up. The optimal pharmacologic therapy to prevent venous thrombosis requires further study.

11

pancreatic injury were analyzed. Descriptive tables compared blunt and penetrating traumas. After excluding emergency department deaths, mortality was compared by management strategy. Results: The NTDB contained 2.5 million patients meeting inclusion criteria, with 79,154 (3.1%) hepatic injuries and 8,073 (0.3%) pancreatic injuries. Penetrating injury rates were 22.8% and 38.1% for hepatic and pancreatic traumas, respectively. A third of penetrating hepatic trauma required repair. More than 25% of penetrating pancreatic trauma required resection, while repair and drainage were performed in 16.1% and 6.5%, respectively. Nearly 40%-50% of hepatopancreatic traumas suffered complications (Table). Mortality rates for hepatic and pancreatic traumas were 14.9% and 20.3%, respectively. Hepatic trauma patients managed with non-operative or drainage therapy had improved mortality (5.8%-6.7%) compared to those undergoing resection or repair (mortality of 17.1%-23.9%). Mortality rates were 7.4% for pancreatic injuries managed with drainage versus >15% in those requiring repair or resection (Figure). Conclusion: Despite advances in management of hepatopancreatic trauma, mortality remains elevated in a large, contemporary, trauma-specific dataset. Unadjusted analysis suggests less aggressive surgical management, when possible, may provide improved outcomes.

LO-B.04 PANCREATIC TRAUMA IN THE UNITED STATES: A CONTEMPORARY APPRAISAL B. R. Englum, A. M. Ganapathi, P. J. Speicher and J. E. Scarborough Duke University School Of Medicine, Dept. Of Surgery, Durham, NC Purpose: Management and outcomes in hepatopancreatic trauma have been examined in small, single-institution studies and older, multi-institutional databases; however, large, contemporary studies examining national outcomes with trauma-specific clinical details are lacking. Methods: The National Trauma Data Bank (NTDB) comprises trauma data from >900 hospitals in the US. We identified adult patients from 2007–2011 with blunt or penetrating injury. Patients with ICD-9 codes for hepatic or HPB 2014, 16 (Suppl. 1), 1–99

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

12

Abstracts Modified Accordion Severity Grading System and ACSNSQIP, is a quantitative measure of postoperative morbidity. This study utilizes PMI to establish the unadjusted complication burden for distal pancreatectomy (DP). Methods: Nine centers contributed ACS-NSQIP complication data for 655 DPs from 2005–2011. Each complication was assigned an Accordion severity weight ranging between 0.11 for grade 1 to 1.00 for grade 6 (death). The PMI is the sum of complication severity weights divided by the total number of patients and can range from 0 (no complication in any patient) to 1.000 (all patients died). Contribution of each complication grade to burden is displayed as a severity “spectrogram.” Results: ACS-NSQIP complications occurred in 177 patients (27.0%). The non risk-adjusted PMI for DP is 0.087, and for complication-bearing cases was 0.322. Bleeding/ Transfusion and Organ Space Infection were the most common complications, both occurring in 7.2% of cases. Frequency and burden differed across Accordion grades (Figure). Grade 2 complications were highest in both frequency and burden. While major complications (grades 4/5/6) represented only 14.9% of complication occurrences, they accounted for 31.0% of the burden. Subgroup analysis reveals how PMI can assess technical aspects of DP (Table). On multivariate analysis, independent predictors for PMI include: functional status, BMI, alcohol intake, neurologic disease, and preop-SGOT. Conclusion: PMI analysis establishes the morbidity burden of DP, allowing for quantitative comparisons between pancreatic resections, other operations, institutions, and surgeons.

LO-B.05 DEFINING THE POSTOPERATIVE MORBIDITY INDEX FOR DISTAL PANCREATECTOMY M. K. Lee1, R. S. Lewis2, S. Strasberg1, B. L. Hall1, P. Study Group2 and C. M. Vollmer2 1 Washington University In St. Louis School Of Medicine, St. Louis, MISSOURI; 2Perelman School Of Medicine, Philadelphia, PA Objective: Accurate assessment of complications is critical in reporting and analyzing surgical outcomes effectively. The Postoperative Morbidity Index (PMI), derived from the © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

13

LO-B.06 THE ROLE OF HEPATIC ARTERY LYMPH NODE METASTASIS IN PANCREATIC CANCER: PROGNOSTIC FACTOR OR SURGICAL SELECTION CRITERIA P. Philips, S. C. Agle, C. R. Scoggins and R. C. Martin University Of Louisville, Louisville, KY Background: Presence of lymphatic metastasis is an independent adverse prognostic factor in pancreatic cancer (PDAC). Hepatic artery lymph node (HALN) metastasis reportedly should be a contraindication to resection since it confers a worse overall survival. Methods: We identified consecutive patients between 2002–2012 who underwent PD for PDAC from our prospective-hepato-pancreatico-biliary database. Overall (OS) and disease-free survival (DFS) were estimated by Kaplan–Meier test. Log-rank test and multivariate Coxproportional-hazards-regression model were used to investigate clinio-pathological variables’ association with survival. Results: A total of 420 patients underwent PD during this period, of which 241 had HALN evaluation. They were subcategorized into LN negative(LN-, n = 50, 20.7%), peripancreatic LN (PPLN) positive,HALN negative (HALN-/PPLN+, n = 150, 62, 2%) and PPLN+/HALN+ (n = 41, 17%). HALN was the only site of LN metastasis in only 3 patients (1.2%). Median Follow-up was 13.5 months (IQR 4.1–28.2). Median OS and DFS were 21.6 months (95% CI: 17–25) and 12.4 months (95% CI: 9–14). On univariate analysis, positive LN status, higher T-stage, higher histological grade and perineural invasion (PNI) were associated with significantly worseOS. On multivariate analysis, hazard ratio(HR) of death was highest among HALN+/ PPLN+, then HALN-/PPLN+ (p = 0.04) and PNI patients (p = 0.02). Similarly, HALN-/PPLN+ and HALN+/PPLN+ (p = 0.01) were the only predictors of worse DFS on Multivariate analysis (See table-1). Among LN positive patients, there was no significant difference in medianOS for HALN positive patients (18.4 months, 95% CI: 12.3–24) compared to HALN negative group(19.7 months, 95% CI: 16.7–22.6, p = 0.7). Conclusion: Lymph node positive disease is an adverse independent prognostic factor in pancreatic cancer, regardless of HALN involvement. Among lymph node positive PDAC after PD, survival is comparable regardless of HALN status. Therefore HALN positive disease should not preclude the performance of PD

LO-B.07 PROPHYLACTIC OCTREOTIDE FOR PANCREATODUODENECTOMY: MORE HARM THAN GOOD? M. T. McMillan1, J. D. Christein2, M. P. Callery3, S. W. Behrman4, J. A. Drebin1, T. S. Kent3, B. C. Miller1, R. S. Lewis1 and C. M. Vollmer1 1 University Of Pennsylvania, Philadelphia, PA; 2University Of Alabama At Birmingham Medical Center, Birmingham, AL; 3Beth Israel Deaconess Medical Center, Boston, MA; 4 University Of Tennessee Health Science Center, Memphis, TN Introduction: The accrued evidence regarding prophylactic octreotide for pancreatoduodenectomy (PD) predates the advent of the ISGPF classification system for postoperative pancreatic fistula (POPF), and its efficacy in the setting of high POPF risk is unknown. The Fistula Risk Score (FRS) predicts the risk and impact of clinically-relevant fistulas (CR-POPF) and can be useful in assessing the impact of octreotide in scenarios of risk.

HPB 2014, 16 (Suppl. 1), 1–99

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

14

Abstracts

Methods: From 2001–2013, 1018 PDs were performed at four institutions, with prophylactic octreotide administered at the surgeon’s discretion. The FRS and the Postoperative Morbidity Index (PMI) were used to analyze occurrence and burden of POPF across various risk scenarios. Results: Overall, 391 patients (38.4%) received octreotide. Without risk adjustment, POPFs occurred more frequently in those who received octreotide (28.1% vs. 21.5%; P = 0.017). Octreotide administration correlated with higher PMI and increased hospital Duration of Stay (13.4 vs. 10.9 d; p < 0.001). Octreotide significantly diminished the incidence of Grade A (biochemical) POPFs (7.16% vs. 14.51%; Figure). Conversely, CR-POPF actually occurred more often when octreotide was used (21.0% vs.7.0%; p < 0.001), especially when there was advanced FRS risk (Table). Regression analysis, controlling for FRS risk, demonstrates that CR-POPF rates correspond with how often individual surgeons used octreotide. Conclusion: This multi-institutional study, using ISGPF criteria, evaluates POPF development across the entire risk spectrum. While octreotide improves biochemical fistula rates, it appears to offer no benefit in preventing CR-POPF. Furthermore, it may potentiate CR-POPF development in the presence of risk factors. This comprehensive analysis suggests octreotide should not be utilized as a POPF mitigation strategy.

LO-B.08 INTERNATIONAL CONSENSUS GUIDELINE PARAMETERS CARRY UNEQUAL WEIGHT AND ARE NOT CUMULATIVE IN INTRADUCTAL PAPILLARY MUCINOUS NEOPLASM MALIGNANT RISK PREDICTION A. M. Roch1, E. P. Ceppa1, J. M. DeWitt2, M. A. Al-Haddad2 and C. M. Schmidt1 1 Department of Surgery, Indiana University School Of Medicine, Indianapolis, IN; 2Department Of Medicine, Division Of Gastroenterology, IU Health University Hospital, Indianapolis, IN Introduction: The International Consensus Guidelines (ICG) stratify malignant risk in patients with Intraductal Papillary Mucinous Neoplasm (IPMN) into 3 progressive categories: “no worrisome features”(NO), “worrisome features”(WF) and “high risk stigmata”(HR). Hypothesis: Type (clinical versus radiologic) and quantity of ICG WF and HR carry unequal weight and are not cumulative in malignant/invasive IPMN risk prediction. Methods: A retrospective review of a prospectively maintained database of patients who underwent surgical resection for IPMN at a single, university-based, medical center (1992–2012) was performed. P < 0.05 was significant. Results: Of 362 patients, 340 were eligible and categorized as 70 NO, 185 WF, 85 HR. WF had more malignant/invasive IPMN versus NO (27% vs. 4%, p = 0.001 and 16% vs. 4%, p = 0.02). HR had more malignant/invasive IPMN versus WF (57% vs. 27%, p = 0.0001 and 42% vs. 16%, p = 0.0001). When radiological parameters only were considered for WF versus HR, no difference was found in malignant/invasive IPMN rate. Contrarily, when clinical parameters only were considered, HR had more malignant/invasive IPMN (67% vs. 8%, p = 0.04 and 67% vs. 3%, p = 0.01). Malignant but not invasive IPMN increased with the quantity of WF (p = 0.001). Quantity of HR had no effect on malignancy/ invasive rate. Conclusion: Type and quantity of ICG WF and HR carry unequal weight and are not cumulative in malignant/invasive IPMN risk prediction. Clinical parameters, in general, are more predictive and should carry more weight in surgical decision making. The quantity of WF is cumulative for malignant IPMN risk prediction, but there is no cumulative impact of the quantity of HR.

LO-B.09 UPREGULATION OF S1P-RECEPTOR-1 IN PANCREATIC DUCTAL ADENOCARCINOMA A. Cardenas1, D. Cucher1, T. Jie2, R. Heimark1,3 and E. Ong2 1 University Of Arizona College Of Medicine – Department Of Surgery, Tucson, AZ; 2University Of Arizona College Of Medicine – Department Of Surgery – Hepatobiliary Surgery, Tucson, AZ; 3University Of Arizona Cancer Center, Tucson, AZ Background: Sphingosine-1-phosphate (S1P) is a lipid signaling molecule that promotes growth and migration and inhibits apoptosis through activation of trans-membrane receptors. Expression of the receptor S1PR1 is increased in © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts cancer. S1P is generated from sphingosine by two kinases, Sphk1 and Sphk2. Altered cytoplasmic/nuclear localization of these kinases in malignant cells is associated with tumor aggressiveness and worse prognosis. We hypothesized that S1PR1 expression would be increased in pancreatic ductal adenocarcinoma (PDAC) in association with altered cytoplasmic/nuclear localization of sphingosine kinase 1 and 2. Methods: Western blot analysis was used to determine S1PR1, Sphk1 and Sphk2 expression in PDAC cell lines. Cytoplasmic and nuclear localization was characterized by biochemical isolation and blotting. Expression of these molecules in 21 FFPE samples of human PDAC was determined using immunohistochemistry techniques. Results: S1PR1 was strongly expressed in PDAC cell lines, but not in non-malignant pancreatic ductal epithelium cells (Fig. 1A). Altered localization of both Sphk1 and Sphk2 was observed in PDAC cell lines (Fig. 1B). S1PR1 was expressed in 76% of human PDAC samples. Nuclear localization of Sphk1 and Sphk2 was detected in 22.2% and 47.6% of the samples. Conclusion: S1PR1 is highly expressed in vitro and in vivo in PDAC, but not in normal pancreatic ductal epithelium. Sphingosine kinase subcellular localization is similarly altered in PDAC. This implies that S1P signaling is altered in cancer, and further study will elucidate the downstream pathway. Inhibition of S1P receptors and/or kinases may be a viable treatment strategy for metastatic PDAC in conjunction with chemotherapy.

HPB 2014, 16 (Suppl. 1), 1–99

15

LO-B.10 A STANDARDIZED CARE PLAN IS ASSOCIATED WITH SHORTER HOSPITAL LENGTH OF STAY IN PATIENTS UNDERGOING PANCREATICODUODENECTOMY D. P. Nussbaum1, K. Penne1, P. J. Speicher1, S. S. Stinnett2, A. Cocieru1, D. G. Blazer1, S. Zani1, B. M. Clary1, D. S. Tyler1 and R. R. White1 1 Duke University Department Of Surgery, Durham, NC; 2 Duke University Department Of Biostatistics & Bioinformatics, Durham, NC Introduction: Standardized Care Plans (SCPs) have gained popularity in pancreatic surgery, yet outcomes data are limited. Our institution introduced a SCP in 1/2011 which included a fast-track recovery pathway (FTRP) and the selective placement of feeding jejunostomy tubes (FJTs). Methods: A retrospective review of 242 consecutive patients undergoing pancreaticoduodenectomy between 2008–2012 was completed. Patients were stratified by SCP implementation. Univariate comparison followed by multivariable linear regression were performed to identify predictors of hospital length of stay (HLOS). Results: SCP patients (n = 100) were slightly older but otherwise similar to pre-SCP patients (n = 142). FJT placement occurred less frequently in SCP patients (38 vs. 94%, p < 0.001). All SCP patients were initiated on the FTRP. Complication rates (Clavien-Dindo grade ≥ 3) were similar (41 vs. 43%, p = 0.79). Among SCP patients, an oral diet was introduced earlier (5 vs. 8.5 days, p < 0.001) and HLOS was shorter (11 vs. 13 days, p = 0.01). There were no differences in readmission, reoperation, or mortality rates. Following adjustment with multivariable linear regression, we confirmed SCP status as an independent predictor of HLOS (-0.9 days, p = 0.047). Vascular reconstruction was the only other independent predictor of HLOS (+1.3 days, p = 0.002). In exploratory analysis to assess SCP components, HLOS was further evaluated based on FTRP status and FJT placement. Although both factors were highly associated with HLOS, neither was independently predictive in multivariable analysis (Table 1). Conclusion: Implementation of a SCP resulted in shorter HLOS without an increase in readmissions. Future studies are necessary to identify the individual components of SCPs that most influence outcomes.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

16

Abstracts

FRIDAY, FEBRUARY 21, 2014, 2:55PM–4:55PM PARALLEL LONG ORAL ABSTRACT C TRANSPLANT BILIARY LO-C.01 ABSORBABLE BIOPROSTHESIS FOR THE TREATMENT OF BILE DUCT INJURY IN AN EXPERIMENTAL MODEL E. E. Montalvo Javé1,5, A. R. Jaime Limon1, M. J. Kim Koh1, M. A. García Pineda1,5, B. León Mancilla1, F. Rojas Mendoza 2, J. Tapia Jurado1, C. Montalvo Arenas3 and J. A. Saavedra 4 1 Department of Surgery, School of Medicine, Universidad Nacional Autónoma de México (UNAM), MEXICO CITY, MEXICO; 2Department Of Gastroenterology, Medica Sur Clinic Foundation, MEXICO CITY, MEXICO; 3Department Of Cellular And Molecular Biology, School Of Medicine, UNAM, MEXICO CITY, MEXICO; 4Department Of Pathology, Medica Sur Clinic Foundation, MEXICO CITY, MEXICO; 5Department Of Surgery, Hospital General De México, MEXICO CITY, MEXICO Biodegradable stent implants are an alternative replacement of the bile duct due to benign stenosis, bile duct injury or cancer. We employed an acellular bone based bioprosthesis (10% water, 65% collagen, 25% hydroxyapatite) as an alternate treatment for biliary injury. Eight Landrace pigs underwent an initial laparotomy in which 3 cm of the common bile duct was excised. An end to end anastomosis (bile ductprosthesis-bile duct) with the liquid collagen coated and hydrogen peroxide gas plasma sterilized bioprosthesis was performed. Every pair of pigs underwent a 1, 6, 12, 18-month follow-up, with liver function test, cholangiography, cholangioresonance, cholangioscopic procedure with Spyglass, Cytokeratin (CK-7, CK-19) immunohistochemistry, and H-E staining for microscopic evaluation. No mortality or infectious complications were observed. All gained weight. None showed acolia or jaundice, nor increased hepatobiliary enzymes preoperatively or at 7, 30, 60, 90 or 180 days. Cholangioresonance showed adequate permeability and no stenosis, moderate dilatation of extra-intrahepatic ducts were found in an 18-month follow-up model. Cholangioscopy with Spyglass showed the bioprosthesis insertion in place, without stenosis. No necrosis, stricture or bile leakage was found on the exploratory laparotomy. Biopsies were obtained from the bifurcation of the bile duct to the descending duodenum, including the graft site. Microscopy at 1-month postimplantation showed a rich deposit of collagen fibers with inflammatory cellular infiltrate. At the 3-month the graft site developed a columnar epithelium similar to the original, CK-7 and CK-19 positive. This collagen based bioprosthesis has potential for its application as a novel treatment for biliary injury and stenosis.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

LO-C.02 NATIONAL VIEW ON BILIARY CANCER: RISING RESECTIONS AND CONSIDERABLE COMPLICATIONS C. J. Yang1, L. A. Bliss1, W. B. Al-Refaie2, A. R. Evenson1, K. Khwaja1, S. C. Ng1 and J. F. Tseng1 1 Beth Israel Deaconess Medical Center, Boston, MA; 2 MedStar Georgetown University Hospital, Washington, DC Background: Resection remains the mainstay of curative treatment for biliary tract neoplasms. Literature suggests a minority of eligible patients undergo resection. Aim of this study was to describe current trends in hepatopancreaticobiliary resections and to delineate complication rates. Methods: Retrospective review of U.S. Nationwide Inpatient Sample 2004–2011 for biliary tract cancer. Procedures: Hepatectomy, lobectomy, cholecystectomy, ablation of liver lesion, bile duct resection, biliary-enteric anastomosis, pancreaticoduodenectomy. Categorical variables analyzed by χ2, trend analyses by Cochran-Armitage, adjusted predictors of outcomes by multivariable logistic regression. Results: 238,968 patient-discharges occurred. 40,414 (16.9%) underwent resection. Majority of resections were performed at teaching hospitals (87.3%) and on elective admissions (82.8%). Proportion of patients resected and proportion at teaching hospitals increased over time (p < 0.0001). Mortality remained stable over time 4.2% (2.9– 4.9%). There was a statistically significant increase in complications over time (p < 0.0001). Complication rates were not affected by teaching status or by annual HPB casevolume of hospital after adjusting for confounding. Risk factors for complication: emergency admission (OR 1.75, 95% CI 1.49–2.05), Medicare (vs private insurance; OR 1.18, 95% CI 1.02–1.38), male (OR 1.42, 95% CI 1.25– 1.61), increasing comorbidities, and age ≥50 (OR 1.54, 95% CI 1.28–1.85). Conclusion: An increasing proportion, yet still a minority, of biliary tract cancers are being resected in the US. There has been a shift towards teaching hospitals, high volume centers, and elective admissions. Despite this favorable trend, complication rates are substantial. To maximize the benefit of hepatopancreaticobiliary resections to this small cohort of patients, future investigations should target complication reduction interventions.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

LO-C.03 IS LOCAL RESECTION ADEQUATE FOR T1 STAGE AMPULLARY CANCER? A. Amini, J. Miura, T. Jayakrishnan, F. Johnston, S. Tsai, K. Christians, C. Gamblin and K. Turaga Medical College Of Wisconsin, Milwaukee, WISCONSIN Introduction: Concerns for morbidity of pancreaticoduodenectomy (PD) has led to practitioners adopting endoscopic resection or ampullectomy without lymphadenectomy in the treatment of T1 ampullary cancer. We hypothesized that survival for patients undergoing local resection of ampullary cancer was inferior to those undergoing PD. Methods: All patients with ampullary cancer reported in the Surveillance, Epidemiology, and End Results database between 2004 and 2010 were collected. Demographic features, adjuvant radiation and 5-year survival rates according to nodal disease and histologic type were compared. Results: There were 1,916 cases of ampullary cancer and 421 (22%) had T1 disease. Among those with T1 disease, 217 (51%) received endoscopic treatment only, 21 (5%) underwent ampullectomy alone, 20 (5%) underwent ampullectomy with regional lymphadenectomy, and 163 (39%) underwent PD. For patients with complete nodal staging (PD), 23% had metastatic disease in the nodes. Grade was significantly associated with node positivity with 10% positivity in well-differentiated (n = 61), 12% in moderatelydifferentiated (n = 163), and 27% in poorly-differentiated tumors (n = 85, P = 0.007). After adjusting for demographic characteristics, node positivity was significantly associated with a worse survival (HR 2.2 (95% CI 1.3–3.8, p = 0.003)) in multivariate models. Survival was improved with either an ampullectomy with regional lymphadenectomy (HR 0.16 (95% CI 0.08-0.03, p < 0.001)) or a PD (HR 0.26 (95% CI 0.15–0.44, p < 0.001)). Conclusion: Patients with early stage ampullary cancer have a high risk for nodal metastases especially if they are higher grade lesions. Nodal clearance with a regional lymphadenectomy or a PD is essential to ensure long term survival for these patients.

LO-C.04 MODEL TO PREDICT SURVIVAL AFTER SURGICAL RESECTION OF INTRAHEPATIC CHOLANGIOCARCINOMA: THE MAYO EXPERIENCE S. M. Ali1, C. J. Clark1, T. Mounajjed3, T. Wu3, W. Harmsen2, K. M. Reid-Lombardo1, M. L. Kendrick1, M. B. Farnell1, D. M. Nagorney1 and F. Que1 1 Division Of Gastroenterologic And General Surgery, Rochester, MN; 2Division Of Biostatistics, Rochester, MN; 3 Division Of Anatomic Pathology, Rochester, MN Background: The AJCC 7th edition has recently been validated and shown to accurately predict survival of patients with intrahepatic cholangiocarcinoma (ICC). We attempted to investigate the validity of recent findings. Methods: A single center retrospective cohort study; with histo-pathological restaging of all consecutive ICC patients who underwent primary surgical resection between 1997 and 2011. Patients with metastatic disease, palliative operations HPB 2014, 16 (Suppl. 1), 1–99

17

or evidence of hepatocellular/cholangiocarcinoma variant on microscopic examination were excluded. Overall survival was compared using Kaplan-Meier estimates and log-rank tests. Results: A total of 151 patients (54% females) with median follow up of 4.5 years met our inclusion criteria. The mean age and mean BMI of the cohort was 58.7 ± 12.1 years and 28.3 ± 5.6 kg/cm2 respectively. AJCC was [stage I, 44(29%), Stage II, 44(29%), Stage III, and Stage III/IVa 63(42%)]. Type of tumor growth was periductal in 36 (25%) patients and mass forming in 105 (75%) patients. Nine(6%) patients had cirrhosis, 19(12%) had steatohepatitis and 68 (45%) had necrosis. Tumor grade was low in 75 (49%) and high in 76(51%) patients. Resection margin was <1 mm in 35 (25%), 2–9 mm in 63(45%) and >10 mm in 41(30%) patients. Major liver resection (>3 segments) was performed in 64 (42%) patients. Survival analysis showed statistically significant results for cirrhosis (p < 0.01), resection margin (p < 0.01), grade (p < 0.01) and AJCC staging (p = 0.04). Conclusions: Our analysis shows that in addition to AJCC criteria; inclusion of grade, tumor growth type, extent of resection margin and cirrhosis offers a more accurate model for survival in our cohort.

LO-C.05 ECONOMIC ANALYSIS OF NEAR INFRARED NAVIGATING SURGERY IN CHOLECYSTECTOMIES IN CLEVELAND CLINIC FLORIDA F. Dip, M. Roy, A. Rosales-Velderrain, C. Simpfendorfer, E. Lo-Menzo, R. Rosenthal and S. Szomstein Cleveland Clinic Florida, Weston, FLORIDA Introduction: Fluorescent cholangiography (FC) is a novel intraoperative procedure involving infrared visualization of the biliary structures. We evaluated the costs and effectiveness of routinely implemented FC and intraoperative cholangiogram (IOC) during laparoscopic cholecystectomy (LC). Material and Methods: All patients enrolled in an IRB approved prospective study comparing LC with FC and IOC between February and June 2013 were reviewed. The procedure time, procedure cost and effectiveness of the two methods were analyzed and compared. Results: A total of 43 patients (21 males and 22 females) were analyzed. The mean age was 49.53 ± 14.35 years and mean body mass index (BMI) was 28.35 ± 8 kg/m2. The overall mean operative time was 64.95 ± 17.43 minutes. FC was faster than IOC (0.71 ± 0.26 mins vs. 7.15 ± 3.76 mins; p < 0.0001). FC was successfully performed in 43 out of 43 cases (100%) and IOC in 40 out of 43 cases (93.02%). FC was less expensive than IOC (14.10 ± 4.31 vs. 778.43 ± 0.40 dollars; p < 0.0001). All surgeons found the routine use of FC useful. Conclusion: In our study FC was effective in delineating important anatomic structures during LC. It was cheaper than IOC.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

18

Abstracts

LO-C.06 COMPARISON OF LONG TERM SURVIVAL BETWEEN LIVER TRANSPLANTATION AND RESECTION IN PATIENTS WITH STAGE I AND II CHOLANGIOCARCINOMA – A NATIONAL RETROSPECTIVE ANALYSIS R. M. Seshadri, D. J. Niemeyer, R. Z. Swan, D. Sindram, J. B. Martinie, M. Russo and D. A. Iannitti Carolinas Medical Center, Charlotte, NC Introduction: Optimal surgical and oncologic treatment for patients with early stage hepatocellular carcinoma is considered to be liver transplantation. It is unknown if a similar survival benefit exists for cholangiocarcinoma (CCA). We aimed to investigate the comparative survival of early stage cholangiocarcinoma treated with resection, locoregional therapy or transplantation. Methods: Stage I and II CCA patients (from 1998–2006) were identified from the National Cancer Database (a joint venture of the Commission on Cancer and the American Cancer Society) and survival analysis was calculated using Kaplan Meier Curves and Log Rank tests. Results: 837(452 M /385 F) patients with stage I or II CCA (out of 7581 CCA pts) met criteria for the study. 81 patients (66 M, 15 F) had LT, 376 patients (164 M, 212 F) had LR and 38 had LRTX (22 M, 17 F). The mean age of LR, LT, LRTX groups was 54, 64 and 65 respectively. Median survival for patients who underwent LT was 89.9 months, LR was 49 months and LRTX was 19.6 months. Conclusion: There was no statistical difference between patients undergoing a liver transplant or a resection (p = 0.07). The survival curves suggest that liver resection offers a survival benefit after 3 years compared to transplant. The locoregional therapy group had the lowest survival but this may be due to patient or tumor characteristics which could not be validated from the database.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

LO-C.07 HEPATIC ARTERIAL FLOW PREDICTS BILIARY COMPLICATIONS AND SURVIVAL IN LIVER TRANSPLANT RECIPIENTS P. T. Kim, G. Saracino, L. Jennings, M. Ramsay, G. J. McKenna, G. Testa, N. Onaca, T. L. Anthony, R. Ruiz, M. Levy, R. Goldstein and G. Klintmalm Baylor University Medical Center, Dallas, TX Introduction: Adequate hepatic arterial (HA) flow to the bile duct is essential for its viability in liver transplantation. This study was conducted to determine the relationship between directly measured hepatic arterial (HA) flow and biliary complications and survival after deceased donor liver transplantation. Methods: A retrospective review of HA and portal vein flows from 2684 liver transplant recipients over a 25-year period was performed from a prospectively maintained database. Rates of biliary complications (biliary leaks, anastomotic and non anastomotic strictures) and survival were compared between two groups: patients with HA flow per body weight ratio of <5 ml/min/kg (N = 884), (Group 1) and >5 ml/min/kg (N = 1800), (Group 2). Results: Patients in Group 1 had higher body weight (93 vs. 77 kg, P < 0.0001), higher MELD scores (17 vs. 16, P = 0.0002), lower HA flow (352 vs. 617 ml/min, P < 0.0001), and higher portal venous (PV) flow 2206 vs. 2143 ml/min, P = 0.0452). There was no difference in cumulative hepatic arterial complications between the groups (13.9 vs. 12.2%, P = 0.2191). Pattients in Group 1 had higher rates of biliary complications at 2 months, 6 months and 12 months (19.8%, 28.2% and 31.9% vs. 14.8%, 22.4% and 25.8%, respectively, P < 0.001) and lower graft and overall survival at 1 year, 2 years and 5 years (graft: 79.4%, 73.4%, and 63.5% vs. 85.8%, 81.1% and 71.5%, P = 0.001; overall 84.3%, 79%, 68.7% vs. 89%, 84.3%, and 75.2%, P < 0.001). Conclusion: Hepatic arterial flows per weight ratio of <5 ml/min/kg predicts biliary complications and is associated with lower graft and patient survival.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

LO-C.08 LIVER TRANSPLANTATION IS A HIGHLY EFFECTIVE TREATMENT FOR MANAGEMENT OF HEPATOCELLULAR CARCINOMA, EVEN FOR HIGHLY SELECTED ADVANCED STAGE PATIENTS M. Doyle, N. Vachharajani, K. Fowler, K. Korenblat, J. Lowell, S. Shenoy, J. Crippin, Y. Lin, J. Wellen, N. Saad and W. Chapman Washington University School Of Medicine, St. Louis, MO Introduction: Optimal management of HCC remains controversial, with a high recurrence rate in patients undergoing hepatic resection. While results for orthotopic liver transplantation (OLT) have been favorable within Milan, concerns about recurrence have limited this therapy in patients beyond Milan. Since 2000, we have used a strategy of downstaging with a 4–6 month period of observation prior to OLT. Method: We reviewed our experience from 1/1/2000–12/ 31/2012 using prospective database, during which time 937 cadaveric adult OLT were performed. Patient demographics, recurrence, and survival were reviewed and Kaplan-Meier plots generated with statistical comparison made using LogRank test. Results: 32% (300/937) patients had HCC. 243 (81%) of HCC cases were diagnosed in the pre-transplant work-up and 210 underwent liver directed therapy prior to OLT. 45 (22.3%) of patients were outside of Milan before downstaging (ALTSG stage III – 27; stage IVA1 – 13; Stage IVA- 5). No patient beyond Milan who failed to downstage underwent OLT. One-, 3-, and 5-year patient survival in 45 down-staged cases was 91%, 82.4%, 78.9%, comparable to within-Milan TACE’d cases (92.9%, 80.8% and 69.3%). One-, 3-, 5-year recurrence-free survival in 45 down-staged cases was 86.5%, 80.4%, 72.6%, also comparable to withinMilan (88.4%, 77.7% and 64.2%). Recurrence rates were 19/249 (7.6%) for within Milan and 3/45 (6.7%) initially beyond Milan but downstaged. Conclusion: Liver transplantation remains a highly effective therapy for patients with HCC, with excellent long-term survival and low recurrence risk. Patients beyond Milan should be required to undergo tumor downstaging with a period of observation prior to undergoing OLT.

HPB 2014, 16 (Suppl. 1), 1–99

19

LO-C.09 IS LIVER TRANSPLANTATION SAFE AND EFFECTIVE FOR GERIATRIC (>70 YEARS) RECIPIENTS? A CASE CONTROLLED ANALYSIS G. C. Wilson, R. C. Quillin, K. Wima, J. M. Sutton, D. J. Hanseman, I. M. Paquette, F. Paterno, E. S. Woodle, D. E. Abbott and S. A. Shah University Of Cincinnati Department Of Surgery, Cincinnati, OH Introduction: As the US population continues to age, patients aged ≥70 years are being evaluated for liver transplantation (LT) with increasing frequency but their outcomes after LT are unknown. Methods: Using a novel linkage of the University HealthSystem Consortium (UHC) and Scientific Registry of Transplant Recipients (SRTR) databases, we identified 13,819 patients who underwent LT from 2007–2012. Patients were sorted into two cohorts: recipient aged ≥70 years at LT (n = 318) vs. recipients <70 years (n = 12,073). A 2 : 1 case control analysis was performed with matched cohorts based on propensity scores. Results: Older recipients had lower MELD scores at LT (median 15 vs. 19, p < 0.0001), were transplanted at high volume centers more (47.3% vs. 34.4%, p < 0.0001) and received more grafts from donors aged >60 years (24.2% vs. 15.5%, p < 0.0001), but had similar donor race, cause of donor death, DRI, and cold ischemia time (p > 0.05 for each). The two cohorts had similar hospital length of stay, in hospital mortality, hospital cost, and 30-day readmission rates. Graft survival was similar between the two cohorts (p = 0.09) but recipients ≥70 years had worse overall longterm survival (p = 0.008; Figure 1). A case controlled analysis confirmed that although hospital LOS was longer in the older cohort, short-term results were similar between the two groups. Conclusions: Selected geriatric recipients have similar perioperative outcomes and graft survival compared to younger recipients; however, they have worse overall survival following LT suggesting that medical care and health maintenance may be important in long-term patient survival.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

20

Abstracts

LO-C.10 RECIPIENT OBESITY ADVERSELY AFFECTS LONG-TERM GRAFT AND OVERALL SURVIVAL IN A SINGLE INSTITUTION ANALYSIS K. D. Conzen1, N. Vachharajani1, K. M. Collins1, C. D. Anderson2, Y. Lin1, J. R. Wellen1, S. Shenoy1, J. A. Lowell1, M. Doyle1 and W. Chapman1 1 Washington University In St. Louis, Department Of Surgery, St. Louis, MO; 2University Of Mississippi Medical Center, Department Of Surgery, Jackson, MS Introduction: The effect of obesity in liver transplantation remains controversial. Earlier data from our institution demonstrated no significant difference in post-operative complications or one-year mortality. We hypothesized that obesity alone has minimal effect on long-term graft and overall survival. Methods: A retrospective study of all primary orthotopic liver transplants (OLT) at our center (1/1/2002–12/31/2012) was conducted. OLT recipients were divided into six groups based on pre-transplant BMI: Group 1 (BMI < 18), Group 2 (18–24.9), Group 3 (25–29.9), Group 4 (30–35), Group 5 (35.1–40) and Group 6 (>40). Pre- and post-transplant parameters were compared between groups. Differences in outcomes were determined using the log-rank test, Chisquare test, and student’s t-test; p-value < 0.05 was considered significant. Results: Seven hundred eighty-five patients met inclusion criteria. Numbers of recipients in each group were: (group 1)9, (2)210, (3)294, (4)169, (5)77, (6)26. No differences were observed in race, age, massive ascites, or diabetes. Recipients with BMI >35 were more likely to have NASH cirrhosis (p < 0.0001) and wait longer for transplant (p = 0.002). There were no differences in operative time, ICU LOS, hospital LOS, vascular thrombosis, bleeding, biliary complications, or infection. Seven year graft survival (36.7% vs 62–79%) and overall survival (38% vs 62–81%) were significantly decreased for BMI >40 (p < .02). Conclusion: Obesity increasingly impacts the field of liver transplantation. Although limited by a single institution, our data suggests that morbid obesity adversely affects long-term outcomes despite similar short-term results. Further analysis is indicated to identify risk factors for poor outcomes in morbidly obese patients.

FRIDAY, FEBRUARY 21, 2014, 5:00PM–6:30PM ORAL POSTER I (LIVER I, II; PANCREAS I, II; BILIARY I; TRANSPLANT I; OTHER) OP-I.01 HIDA SCAN OF THE FLR IS SUPERIOR TO VOLUMETRY TO ASSESS LIVER FUNCTION IN ALPPS E. Schadde1, M. Lesurtel1, N. Schäfer2 and P. Clavien1 Dept. Of Visceral And Transplant Surgery, University Of Zurich Hospital, Zurich, ZURICH; 2Dept. Of Nuclear Medicine, University Of Zurich Hospital, Zurich, ZURICH

1

Introduction: ALPPS (Associating Liver Partition with Portal vein ligation for Staged hepatecomy) induces rapid hypertrophy, but the functionality of the tissue has been questioned even when volume cut-offs of the future liver remnants of 30% prior to resection are respected. 99mTechnetium-mebrofenin clearance capacity of the future liver remnant (FLR) of less than 1.8 %/min/BSA using HIDA scintigraphy has been postulated to increase the risk of postoperative liver failure after resection. We assessed the value of HIDA in a pilot series of patients undergoing ALPPS and Portal Vein Embolization (PVE). Methods: Four patients undergoing ALPPS were compared with 4 patients undergoing PVE before stage 1 and stage 2. Scintigraphy was performed by intraveneous injection 80 MBq99m Technetium-mebrofenin. Dual head dynamic acquisitions were used to calculate FLR function. MRI volumetry was used to assess volume changes. Postoperative liver failure was assessed using 50/50 criteria. Results: One patient in the ALPPS group and one patient in the PVE group developed fatal postoperative liver failure. These two patients had a FLR function after hypertrophy of only 2.14 (ALLPS) and 1.28 (PVE) %/min/BSA, despite having reached a FLR volume of greater than 30%. In contrast the FLR function of patients with an uncomplicated postoperative course was 3.6, 5.5 and 4.6 %/min/BSA in the ALPPS group and 2.3, 3.7 and 5.4 %/min/BSA in the PVE group. Conclusion: In ALPPS and also PVE, assessment of FLR function using 99m Technetium-mebrofenin clearance capacity of the future liver remnant is superior to volumetry to assess resectibility.

OP-I.02 MORBIDITY AND MORTALITY IN 1174 PATIENTS UNDERGOING HEPATIC PARENCHYMAL TRANSECTION USING A STAPLER DEVICE M. Raoof2, T. A. Aloia1, J. Vauthey1 and S. A. Curley1 University Of Texas, MD Anderson Cancer Center, Houston, TX; 2University Of Arizona, Tucson, AZ

1

Background: Transection of liver parenchyma using staplers is now commonly performed. Large studies are needed to assess the usefulness of the technique as well as perioperative outcomes. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts Methods: This is a retrospective study of a prospectively maintained database. A total of 1174 patients undergoing liver resections in routine surgical practice, using a stapler device at our Cancer Center between January 1, 1994 and November 10, 2011 were evaluated. Results: There were 900 (77%) major resections (3 segments or more) and 274 (23%) minor resections (<3 segments or wedge resections). A vast majority, 1133 (96.5%) were indicated for an underlying malignancy (24% primary liver or gall bladder and 72.5% metastatic) compared to benign disease 41 (3.5%), with the most common indication being metastatic colorectal cancer 584 (49.7%). Of the total 1174 patients 128(10.9%) had a prior liver resection. Median OR time and blood loss was 206 min and 300 ml, respectively, with 11% of patients requiring transfusion in the perior post-operative period. Overall-morbidity and mortality rate was 14% and 3.2%, respectively, with a median hospital stay of 7 days (IQR 4). Multivariate logistic regression demonstrated blood loss and extent of liver resection to be independent predictors of adverse outcome. Thirteen instances (1.1%) of misfired staplers were noted and were associated with higher blood loss (p < 0.001) and mortality (15% vs. 3.1%, p = 0.013). Conclusion: Use of stapler device for hepatic resection is safe and effective but rare instances of a misfired stapler device are associated with an adverse outcome.

OP-I.03 MRI WITH GADOXETATE DISODIUM FOR COLORECTAL LIVER METASTASIS: IS IT THE NEW GOLD STANDARD? S. Patel, S. Cheek, H. Osman, T. Winston and D. Jeyarajah Methodist Dallas Medical , Dallas, TX Background: Approximately 15–45% of patients develop hepatic metastasis from colorectal cancers (CRLM) within 3 years of diagnosis. Accurate detection of CRLM is paramount in the role of hepatectomy. This study aims to compare Magnetic Resonance Imaging (MRI) with gadoxetate disodium (Eovist) to enhanced Computed Tomography (CT) in detecting CRLM. Methods: A retrospective chart analysis of 30 patients from 2011–2013 with CRLM was performed. Patients with more than 6 weeks or 2 cycles of chemotherapy between the two imaging modalities were excluded. The number of lesions identified on enhanced CT vs. MRI with Eovist was compared. Results: A total of 30 patients underwent enhanced CT scans initially then subsequently evaluated with MRI with Eovist. 12 (40%) patients had more lesions identified on MRI with Eovist. 18(60%) had no change in the number of lesions identified. All lesions identified on CT were evident on MRI. When MRI with Eovist detected more lesions, the mean number of additional lesions detected was 1.5 (1–4). Conclusion: MRI with gadoxetate disodium (Eovist) is superior to enhanced CT in identifying CRLM. The increased number of lesion identified on MRI with Eovist can profoundly change the surgeon’s management. It should be considered the “gold standard” in pre-operative imaging for liver metastasis in these patients.

HPB 2014, 16 (Suppl. 1), 1–99

21

OP-I.04 RACIAL DISPARITY OUTCOMES IN PATIENTS UNDERGOING HEPATECTOMY: IS RENAL FUNCTION THE DRIVING FORCE? J. E. Dowden, D. J. Taber, J. W. McGillicuddy, C. Bratton, P. Baliga and K. D. Chavin Medical University Of South Carolina, Division Of Transplant Surgery, Charleston, SC Objective(s): The aim of this study was to determine if African-Americans (AA) undergoing hepatic resection experience differences in surgical outcomes and conduct modeling to develop potential mechanisms that help to explain this disparity. Methods: This was an observational longitudinal cohort study. Comprehensive baseline and clinical data was collected from medical records in patients undergoing partial hepatectomy from 2004–2012. Both open and laparoscopic procedures were included; pediatric patients were excluded. Analysis included univariate and multivariate statistics. Results: 166 patients (59 AA, 107 non-AA) underwent partial hepatectomy and were included. AA patients were more likely to be female, have a higher BMI, have resections due to hemangiomas or adenomas, and have steatosis on explant analysis. Intra-operatively, AA patients had longer surgeries, more blood loss and higher blood product utilization. Significant post-operative complications were significantly more common in AA patients. Multivariable modeling demonstrated that race, history of Hepatitis C and EBL were the only variables that were independently associated with a major post-operative complication; however, baseline serum creatinine was the only variable that significantly modified the effect of race on complications. AAs with normal SrCr levels had similar rates of complications, but as the baseline SrCr level increased, the odds-ratio of developing a complication for AA patients dramatically increased (Figure 1). Conclusions: AA patients have longer and more complex hepatectomy surgeries and experience a higher rate of major post-operative complications, which is predominantly driven by a subset of highly vulnerable AA patients that have preexisting renal insufficiency.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

22

Abstracts rate of margin negative and preserve the hepatic parenchyma in hepatectomy. In this study we focused on the effect of NAC on parenchymal sparing in hepatectomy for CRLM with analyzing the change of tumor size and relationship with the intrahepatic vessels. Methods: Forty tumors in 20 patients who received NAC before hepatectomy for CRLM were evaluated in this study. The change of maximal tumor size and distance from major intrahepatic vessels were measured before and after NAC. Results: The maximal tumor size before and after NAC was 44.8 +/− 41.9 mm and 34.3 +/− 34.3 mm and the tumor size decreased 20.6 +/− 29.8%. The distance between the center of the tumors to the vessels before and after NAC were 23.8 +/− 20.3 mm and 16.9 +/− 14.5 mm and the distance decreased 28.7 +/− 13.9%. However, the space between the edge of the tumor and the vessels did not change (before and after NAC were 0.6 +/− 1.3 mm and 1.1 +/− 2.6 mm). After NAC, operation mode was changed to spare the hepatic parenchyma in 7 patients. Summary: NAC induced the tumor downsizing; however, the space between the tumor and the vessel did not change. These results indicated that the adjacent vessels to the tumor should be resected to achieve R0 resection and the parenchymal sparing was not achieved by preserving intrahepatic vessels but only by tumor downsizing after NAC for CRLM.

OP-I.06 FALLING BARRIERS TO LAPAROSCOPIC LIVER RESECTION A. Hauch, S. Lee, A. Golden, E. Kane, E. Kandil and J. Buell Tulane Transplant Institute, New Orleans, LA

OP-I.05 IMPACT OF NEO-ADJUVANT CHEMOTHERAPY ON PARENCHYMAL SPARING IN HEPATECTOMY FOR COLORECTAL LIVER METASTASIS S. Kaihara, K. Okada, K. Inoguchi, S. Yo, T. Yamamoto, S. Yagi, K. Uryuhara and R. Hosotani Kobe City Medical Center General Hospital, Kobe, HYOGO Object: One of the aims of neo-adjuvant Chemotherapy (NAC) for colorectal liver metastasis (CRLM) is to increase © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

Introduction: Laparoscopic liver resection is an accepted modality for the management of a variety of benign and malignant tumors. Initial efforts were focused on peripheral lesions. We now examine selected patient demographics and extent of operative interventions that might have previously precluded a laparoscopic approach. Methods: Retrospective review of 615 patients managed by a single surgeon from 2000 to 2013. Analysis was performed to compare and contrast demographics and operative characteristics between both open and laparoscopic hepatic resections. Results: 753 hepatic resections in 615 patients were reviewed. Demographics (age, BMI, incidence of cancer and cirrhosis) were compared between open and laparoscopic approaches and found to be equivalent. In contrast ASA was higher in the open group (3.0 vs. 3.1:p < 0.05) as was the extent of resection as indicated by number of segments resected (3.0 vs. 2.4:p < 0.001) and percentage of major resections (formal lobe) (23.3 vs. 47.0:p < 0.001). Conclusion: This analysis identified few patient demographics as continued barriers to a laparoscopic approach. Initial analysis does confirm open surgical techniques are reserved for patients deemed more critically ill, using ASA as a surrogate marker. The incidence of formal lobar resection was also significantly different indicating large atypical resections are most often reserved for an open approach. This study demonstrates that some barriers to a laparoscopic approach for liver resection are diminishing. Future studies are necessary to delineate the persistent barriers for continued adaptation of the laparoscopic approach. HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

23

OP-I.08 RISK FACTORS FOR REPEAT CHEMOEMBOLIZATION FOR HEPATOCELLULAR CARCINOMA J. White, D. T. Redden, M. K. Bryant, D. P. Dorn, S. Saddekni, A. K. Aal, S. H. Gray, D. E. Eckhoff and D. A. DuBay University of Alabama at Birmingham, Birmingham, AL

OP-I.07 THE EFFECT OF ALCOHOL ON SIRT1 EXPRESSION AND FUNCTION IN ANIMAL AND HUMAN MODELS OF HEPATOCELLULAR CARCINOMA K. Thompson1, J. Humphries2, D. J. Niemeyer1, D. A. Iannitti1, D. Sindram1 and I. McKillop1 1 Carolinas Medical Center, HPB Surgery, Charlotte, NC; 2 University Of North Carolina At Charlotte, Department Of Biology, Charlotte, NC Risk factors for HCC include chronic ethanol (EtOH) consumption. Sirt1 is a deacetylase implicated in alcoholinduced liver injury and overexpressed in HCC. The aims of this study were to investigate Sirt1 expression in models of EtOH feeding and HCC, and in human HCC cells with differing alcohol metabolizing profiles. Juvenile C57BL/6 and B6C3 mice were injected with DEN and randomized to receive water or 10%-EtOH for 8 Wks when 36 Wks old. Livers were analyzed for HCC incidence, size and Sirt1 expression. In parallel, human HepG2 HCC cells or HepG2 cells transfected with ADH and CYP2E1 were treated with alcohol (0–50 mM) and/or 10 μM CAY10591.B6C3 mice exhibited significantly elevated Sirt-1 expression vs. C57BL/6 mice and Sirt-1 expression was elevated in HCC vs.non-tumor liver. However, EtOH-feeding did not further alter Sirt1 expression in mice of either background despite EtOH increasing size and incidence of HCC in B6C3 mice. In vitro, VL-17a-cells exhibited significantly greater alcohol metabolism vs.HepG2-cells and EtOH treatment significantly decreased Sirt1 expression in VL-17a-cells, an effect not observed in HepG2. EtOH-treatment stimulated VL-17acell growth but not HepG2 cells and EtOH-dependent changes in VL-17A-cell growth were abrogated by CAY10591. Sirt1 expression correlates with HCC susceptibility in different mouse strains, but not further affected by alcohol. Conversely, Sirt1 expression and function is directly influenced by alcohol metabolism in HCC cells. These discrepancies in Sirt1 expression-function may reflect differences in enzyme expression vs. activity or changes in genes that are targeted for deacetylation during tumor progression in the setting of alcohol feeding-treatment.

HPB 2014, 16 (Suppl. 1), 1–99

Background: Repeat chemoembolization (TACE) is a common intervention performed for hepatocellular carcinoma (HCC). The purpose of this study is to measure risk factors associated with repeat TACE. Methods: Between 2007 and 2012, patient and tumor variables were collected on 262 patients treated with a first TACE procedure as an initial oncologic intervention for HCC at a single University Hospital. Results: Repeat TACE was performed in 66/262 (25%) patients: 2 TACE in 53 and 3 TACE in 11 and 4 TACE in 2. HCC necrosis, measured by mRECIST criteria after first TACE, was lower in patients who subsequently received a repeat TACE (p < 0.05). The following factors were predictive of repeat TACE on univariate analysis: more than 1 tumor (p = 0.03), size (p = 0.01), total tumor diameter (p = 0.001), and age (p = 0.008). On multivariable analysis, only total tumor diameter (OR 2.76/cm, 95% CI 1.45–5.25, p = 0.002) and increasing age (OR 1.04/year, 95% CI 1.00– 1.07, p = 0.03) were predictive of repeat TACE. Interestingly, measures of liver function, TACE vehicle (lipiodol vs. drug eluting beads), and TACE approach (selective vs. nonselective) were not predictive of repeat TACE. After censoring for liver transplantation, median survival was not statistically different among patients that received repeat TACE procedures (1 TACE 26.1 months vs. >1 TACE 21.1 months, p = 0.70). Liver Transplantation was less common in patients with repeat TACE (7.5% vs. 20.6%, p = 0.01). Conclusion: Receiving a repeat TACE procedure is associated increased HCC tumor burden and/or subtotal TACEinduced HCC necrosis. Importantly, repeat TACE was not associated with lower overall survival.

OP-I.09 STARTING A NEW LAPAROSCOPIC LIVER SURGERY PROGRAM: INITIAL EXPERIENCE AND IMPROVED EFFICIENCY S. Liang1 and S. Jayaraman1,2 Division Of General Surgery, University Of Toronto, Toronto, ON; 2HPB Service, St. Josephs Health Centre Toronto, Toronto, ON

1

This study evaluates the safety and outcomes of introducing a laparoscopic liver program to a hospital. This is a retrospective review of a prospectively acquired database of all laparoscopic liver resections between August 2010 and July 2013 at St. Joseph’s Health Centre, Toronto. The primary outcome was safety defined as mortality, major morbidity, and negative margins. Secondary outcomes included other perioperative outcomes. Non-parametric tests were used to compare the first (Group A) and second (Group B) 18-months of the experience. There were 19 patients in Group A, and 25 in Group B with 8 major resections in Group A, and 5 in Group B. Group A had the only mortality due to portal and hepatic vein thrombosis resulting in liver © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

24

Abstracts

failure. There was no statistical difference between the groups in major complications (10.6% vs. 16%, p = 1), or length of stay (3 vs. 4 days, p = 0.59). In each group, there was one case that required intra-operative margin revision. One case in Group B had a positive parenchymal margin on final pathology. There was a significant decrease in operative time (210 vs. 164 min, p = 0.013) between the groups, explained by improvement in the operative times of major resections (296 vs. 193 min, p = 0.048). Furthermore, there was a significant increase in procedures performed for malignancy (34% vs 84%, p = 0.002). This study demonstrates that laparoscopic liver resection can be safely introduced into a hospital. The decrease in operative time despite increasingly complex cases is likely a reflection of increased efficiency with major resections.

OP-I.10 LAPAROSCOPIC CAUDATE LIVER RESECTION S. Lee, A. Hauch, E. Kane, E. Kandil and J. Buell Tulane Transplant Institute, Tulane University School Of Medicine, New Orleans, LA Background: Laparoscopic caudate lobe resection is a feasible, yet still infrequently performed procedure. As proficiency and acceptance of laparoscopic techniques for hepatectomy progress among liver surgeons, data is needed to guide whether patients with anatomically challenging locations can be better treated with a laparoscopic approach. To date, perioperative outcomes following laparoscopic and open approaches have not been compared. Methods: Cohorts of 22 patients undergoing laparoscopic and 20 patients undergoing open caudate resection were matched based on preoperative criteria and compared. Caudate resections with and without additional segmentectomy were included. Outcomes included survival, operating time, surgical margin, cancer recurrence, transfusion requirements, ICU utilization, and length of stay. Results: Open resection had a longer operating time (5 vs. 3 hrs, p = 0.004), increased blood loss (425 vs. 150 mL, p = 0.008), and was more likely to require post-operative critical care (60.0 vs. 13.6%, p = 0.002). Length of stay was not significantly different between groups (7.0 vs. 5.5days, p = 0.111). Morbidity was 13.6% in the laparoscopic group compared to 40.0% in the open group (p = 0.052). R0 margins were achieved in 71.4% of laparoscopic versus 78.6% of open resections (p = 1.0). Rates of recurrence of all cancers appeared to be higher in the laparoscopic group, but this was not significantly different when compared to open resection. Conclusion: The laparoscopic approach is a viable option for caudate lobe resection in select patients. Compared to an open approach, laparoscopic resection may benefit patients in terms of shorter operations, less operative blood loss, decreased ICU utilization, and fewer post-operative complications.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

OP-I.11 LIVER RESECTION IN THE ELDERLY FOR COLORECTAL LIVER METASTASIS: OUTCOMES OF OCTOGENARIANS E. Simoneau, M. Shaheen, A. Salman, P. Chaudhury, M. Kyres and P. Metrakos Division Of Hepatopancreatobiliary Surgery, McGill University Health Center, Montreal, QC Introduction: Liver resections in elderly patients(>70 year old) have previously been shown to be safe, tolerable and with similar outcomes. However these reports made comparisons to young patients and a combination of different diagnoses, which all have inherently different outcomes. The aims of our study were to evaluate the perioperative course of very elderly patients(over 80 years old) with colorectal liver metastasis(CRCLM). Methods: A retrospective study was conducted identifying elderly patients undergoing liver resections for CRCLM between 1996–2013. Baseline characteristics, complications and perioperative mortality were obtained. Results: 65 patients were included, 17 patients over 80 years old (80+) and 48 patients in a less elderly group (75–79 years-old). Resectability was 88.2% and 91.7% in the 80+ and 75–79 groups respectively.There was no difference in baseline characteristics and types of procedure done (60.0% vs 43.2% major resections in 80+ vs 75–79 groups respectively, p = 0595). Post-operatively, the octogenarian group had more intra-abdominal collections (20.0% vs 2.3%, p = 0470) and length of stay was longer (13(7–17) days vs 9(6–11) days, p = 0.0456). Perioperative mortality was the same, with 2 deaths in the 75–79 group compared to 1 death in the 80+ group. Conclusion: This is the largest series reported to date of octogenarians with CRCLM undergoing liver resection. In this patient population, liver resection may be performed with similar perioperative mortality to a younger patient cohort. There appears to be a greater risk of intra-abdominal collection and longer length of stay in the octogenarian group. Selected patients over 80 may benefit from resection of CRCLM.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

OP-I.13 SURGERY AND STEREOTACTIC BODY RADIOTHERAPY (SBRT): A COMBINED APPROACH FOR LOCAL CONTROL OF LIVER CANCER N. L. Thai, K. Tom, A. Khan, P. Abrams, J. Oliva, A. Kirichenko and M. Szramowski Allegheny Health Network, Pittsburgh, PA Combined surgery and radiation therapy can allow patients with larger hepatic tumor burdens to receive aggressive local therapy. Recent advances in SBRT have made radiation more focused, allowing for precise eradication of liver tumors. At our facility, the local control rate is 100% for tumors <=4 cm at 2 years. We initiated a prospective study to evaluate the combination of radiation and surgery in patients that otherwise may not have qualified for aggressive local treatment. The 26 patients had hepatic masses of varying origins. Eight patients underwent surgical resection of some lesions and SBRT to others based on inaccessibility of the tumors to just one modality. Two patients received up-front radiation to downsize the tumor prior to resection. Ten patients underwent SBRT as a bridge to liver transplantation. Six patients were originally outside of Milan criteria for transplant. There have been no recurrences at SBRT sites or after transplant at up to 41 months. Two resection patients have developed extra-hepatic metastases, and 5 have liver lesions at new locations. One patient had 7 colorectal metastases resected, followed by SBRT to an 8th lesion. He has no known disease at 41 months. Ten patients received SBRT as a bridge to transplant; 8 of these have been transplanted without tumor recurrence at up to 35 months. Six patients were successfully downsized to within Milan criteria using radiation. This proof-of-concept study demonstrates that radiation therapy can be used in conjunction with a surgical approach to expand the indications for local control of liver cancers.

OP-I.14 NOVEL EVALUATION FOR HEPATIC FUNCTIONAL RESERVE AFTER LIVER RESECTION IN COMBINATION WITH 3D-CT AND 99MGSA SCINTIGRAPHY M. TANIGUCHI1, A. Okizaki2, K. Watanabe1, K. Uchida1, K. Imai1, N. Shuke2 and H. Furukawa1 1 Division Of Gastroenterologic And General Surgery, Asahikawa Medical University, Asahikawa, HOKKAIDO; 2 Department Of Radiology, Asahikawa Medical University, Asahikawa, HOKKAIDO Background and Aim: It is important to develop accurate diagnostic tools that can predict the risk of poor hepatic functional reserve (PHFR) after liver resections. In this study, the role of preoperative liver scintigraphy and CT volumetric measurement of the liver in the preoperative risk assessment for PHFR was evaluated. Patients and method: Sixty-four patients who were scheduled for liver resection more than sectionectomy because of liver tumors were enrolled in this study. Indocyanine green clearance, technetium -99m -diethylenetriaminepentaacetic acid-galactosyl human serum albumin scintigraphy (99mTcGSA), remnant liver volume calculated by CT were estiHPB 2014, 16 (Suppl. 1), 1–99

25

mated preoperatively. 99mTc-GSA, and conventional liver function tests on 14 and 28 days after liver resection were estimated postoperatively. The correlations between pre and postoperative parameters were evaluated. Multivariate analysis and receiver operating characteristics analysis was performed to assess the risk of PHFR. Result: Significant correlations were observed between postoperative 99mTc-GSA studies and conventional liver function tests. Preoperative 99mTc-GSA clearance times remnant liver volume (GSA × RLV) showed better correlation with postoperative 99mTc-GSA studies and conventional liver function tests than preoperative 99mTc-GSA clearance alone, or remnant liver volume. According to multivariate analysis, GSA × RLV was the only significant independent predictor of PHFR. No patients with GSA × RLV above 200 manifested poor hepatic functional reserve. Using a GSA × RLV of 180, it was possible to predict postoperative PHFR at a sensitivity of 92%, specificity of 83%. Conclusion: Preoperative measurement of 99mTc-GSA clearance and the remnant liver volume on CT proved valuable in assessing the risk of PHFR after liver resection.

OP-I.15 TUMOR SIZE INDEPENDENTLY IMPACTS SURVIVAL IN PATIENTS UNDERGOING SURGERY FOR HEPATOCELLULAR CARCINOMA O. C. Kutlu, M. Wachtel and S. Dissanaike Texas Tech University Health Sciences Center, Lubbock, TEXAS Background: Recently the validity of tumor size as a predictor of survival has been questioned. Posited was the notion that analysis of a large database might help assess this question. Methods: A SEER case listing session acquired patients with hepatocellular carcinoma between 2004–6, with known surgical, radiotherapy, and race status. All were M0 and N0 or NX at diagnosis. Variables and data stratifications are provided below. Important proportional hazards assumption violations were detected and 101 of the 4,863 (2.1%) patients were lost to follow-up, binomial logit regression estimated odds ratios (OR), with 1.25th and 98.75th percentiles of 10,000 bootstrap replicates being 97.5% confidence intervals. Two multivariate regression analyses were performed, one 2 y, one 4 y post diagnosis; null hypotheses were rejected when P < 0.05/2 = 0.025. Results: Recorded were 2,683 (55.2%) deaths at 2 y and 3,324 (68.4%) at 4 y. Compared with tumors 1–50 mm, 51–100 mm were associated with a worse outcome, as were 100+ mm. Compared with patients who underwent resection, those who underwent local extirpation or did not undergo surgery fared worse while those who underwent transplantation did better. Although 40–54 y and 55–69 y were not shown to differ from 18–39 y patients, 70+y patients fared worse. Asians/Native Americans/Pacific Islanders survived more often than Whites. Gender or radiotherapy didn’t increase odds of death. (Wald’s χ2 test, P > 0.025). Conclusions: Tumor size proved an important predictor of survival. Although 70+y patients had increased risks of death, this was not shown to be the case for 40–54 y or 55–69 y patients. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

26

Abstracts

OP-I.16 BLOCKING CD47 PROMOTES PHAGOCYTOSIS OF HUMAN HEPATOCELLULAR CARCINOMA CELLS Z. Xiao1, K. Ott1, Y. Lin1, C. Chen1, L. Lin1, G. Upadhya1, P. Manning3, W. Frazier 2, T. Mohanakumar 1 and W. Chapman1 1 Department Of Surgery,Washington University School Of Medicine, Saint Louis, MISSOURI; 2Department Of Biochemistry & Molecular Biophysics ,Washington University School Of Medicine, Saint Louis, MISSOURI; 3 Vasculox, Inc. St Louis, Saint Louis, MISSOURI CD47 allows cancer cells to evade immunologic surveillance via a ‘don’t eat me’ signal. No data exists supporting the role of CD47 in hepatocellular carcinoma (HCC). We aimed to determine CD47 expression in HCC and to evaluate the ability of blocking monoclonal antibody in promoting phagocytosis in vitro. Methods: We determined the CD47 expression in HCC samples and normal livers by immunohistochemistry, and in human hepatocytes and HepG2 cells by flow-cytometry. Toxicity of CD47mAb to HepG2 cells was determined by MTT. To assess phagocytosis, macrophages from NOD/SCID/r mice were co-cultured with HepG2 cells under three conditions: IgG control, non-blocking mAb 2D3,and blocking CD47mAb400. A phagocytic index was calculated by counting cells under confocal microscopy. Results: Immunofluorescence staining revealed that CD47 expression is higher in HCC (9/10) than in normal liver (0/6) (p < 0.001). Immunofluorescence intensities of HCC, adjacent tissues and normal liver were 88.8 ± 2.50, 23.9 ± 1.45 and 18.0 ± 0.50 respectively. CD47 expression was upregulated in HepG2 but absent in human hepatocytes (p < 0.001). HepG2 cells incubated overnight with CD47 mAb did not demonstrate any cell death. Phagocytosis of HCC cells was significantly increased by the blocking CD47mAb400 compared to control groups (both p < 0.001). The phagocytic index of the anti-CD47mAb400, IgG and 2D3 were 71.1 ± 1.08, 11.5 ± 0.36 and 9.6 ± 0.61 respectively. Conclusions: CD47 is more highly expressed in HCC and on HepG2 cells than in normal tissue or hepatocytes.CD47 blockade increased macrophage phagocytosis of HCC cells suggesting a role for CD47 blockade in managing HCC.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

OP-I.18 TACHOSIL VERSUS SURGICEL ORIGINAL FOR THE SECONDARY TREATMENT OF LOCAL BLEEDING IN ADULT PATIENTS UNDERGOING HEPATIC RESECTION Y. Genyk1, T. Kato2, J. J. Pomposelli3, K. W. Lophaven4 and W. C. Chapman5 1 University Of Southern California, Los Angeles, CA; 2 Columbia University Medical Center, New York, NY; 3 Lahey Clinic Medical Center, Burlington, MA; 4Takeda Development Centre Europe, Langebjerg 1, ROSKILDE; 5 Washington University In St Louis, St. Louis, MO Background: TachoSil® is an absorbable fibrin sealant collagen patch that is currently licensed in the US as an adjunct to hemostasis in cardiovascular surgery. This randomized, prospective, multicenter study compared the efficacy and safety of TachoSil® versus Surgicel® Original as standard adjunctive treatment for hemostasis in patients undergoing open or laparoscopic hepatic resection. Methods: A total of 224 hepatic resection patients with minor to moderate bleeding from the resection area after primary control of arterial bleeding or major venous hemorrhage were randomized to treatment with TachoSil® (n = 114) or Surgicel® Original (n = 110). Primary outcome was the proportion of patients with intraoperative hemostasis at target bleeding site within 3 minutes of application of randomized treatment. Results: Hemostasis within 3 minutes was achieved by 92 (80.7%) patients in the TachoSil® group and 55 (50.0%) in the Surgicel® Original group (OR 4.87; 95% CI: 2.55, 9.29; P < 0.001). The proportion of patients with hemostasis at 5 minutes was also significantly higher in the TachoSil® group than the Surgicel® Original group (94.7 % vs 76.4%, P < 0.001), while the time to hemostasis was significantly shorter (P < 0.001). Most adverse events (AEs) were mild in severity and considered to be unrelated to trial treatment. AEs considered to be treatment-related were reported for 5 patients in the TachoSil® group and 4 patients in the Surgicel® Original group. Conclusions: TachoSil® was superior to Surgicel® Original for achieving hemostasis in adult patients undergoing hepatic resection. TachoSil® was well tolerated, with no evidence of increased frequency of AEs compared with Surgicel® Original.

OP-I.19 HEPATIC STELLATE CELL-DERIVED TENASCIN-C IS ASSOCIATED WITH OBESITY-INDUCED HEPATOCELLULAR CARCINOMA T. Li, J. H. Benbow, T. L. Walling, A. C. Smith, I. H. McKillop, D. Sindram, K. J. Thompson and L. W. Schrum Carolinas Medical Center, Charlotte, NC Obesity is an independent risk factor for developing hepatocellular carcinoma (HCC), and nonalcoholic fatty liver disease (NAFLD) is a consequence of obesity. Hepatic stellate cell (HSC) activation is essential for the development of NAFLD-induced fibrosis/cirrhosis and HCC. Activated HSCs secrete tenascin-C (TnC), an extracellular matrix glyHPB 2014, 16 (Suppl. 1), 1–99

27

coprotein, which can be modulated by Rev-erbα, a nuclear repressor that plays a key role in HSC activation and the fibrogenic response. The aim of this study was to determine if increased secretion of TnC from activated HSCs is associated with obesity-induced HCC. Methods: Primary rat HSCs and hepatocytes were isolated from male Sprague-Dawley rats. TnC secretion, detected by immunoblots, was assessed in activated HSCs with or without Rev-erb agonists (10 μM). Hepatocytes were treated with TnC (50 nM) and/or LPS (1 μg/ml) and inflammatory gene expression was measured by RealTime PCR. Tissue from mice fed a high fat diet (HFD) or control diet (CD) +/− diethylnitrosamine (DEN, liver tumor promoter) was assessed for inflammatory gene expression and HSC activation. Results: TnC secretion increased during HSC activation which was blocked by Rev-erb agonists. TnC synergized with LPS to increase inflammatory gene expression (IL-6, IL-8, TNFα) in hepatocytes. Similar results were observed in livers of HFD compared to CD. Increased TnC expression was associated with activated HSCs in HFD + DEN animals compared to CD + DEN. Conclusion: HSC-derived TnC may be a key regulator of obesity-induced HCC and suggests TnC and upstream regulator Rev-erb as potential therapeutic targets.

OP-I.20 PREOPERATIVE ASSESSMENT OF REMNANT LIVER FUNCTION FOLLOWING HEPATECTOMY BY HEPATOBILIARY SCINTIGRAPHY AND INDOCIANINE GREEN CLEARANCE K. Okada, H. Kinoshita, Y. Sakamoto, S. Yao, T. Yamamoto, K. Inoguchi, R. Hosotani and S. Kaihara Kobe City Medical Center General Hospital, Kobe, HYOGO Introduction: It is extremely important to assess the limit of hepatectomy before operation to prevent fatal postoperative liver failure. In our department, we determine the resection area of hepatectomy using 99mTc-GSA with single-photon emission computed tomography (GSA-SPECT) and indocianine green clearance (KICG) to estimate postoperative hepatic function. We report the efficiency of our assessment methods. Method: Using simulated surgery on fusion images of SPECT and CT, we calculate residual ratio of the functional volume after hepatectomy (RFV). We multiply the value of K-ICG by the RFV and calculate the predicted K-ICG of remnant liver (Rem-KICG). We define it operable when Rem-KICG is over 0.05. In our hospital between January 2006 and November 2012, there are 43 cases which satisfy that conditions. The relationships between Rem-K-ICG and postoperative hepatic function parameters are analyzed. We use Speaman’s rank correlation coefficient to analyze correlations. Result: In analysis of the 43 cases, there are significant relationships between Rem-KICG and postoperative peak serum total bilirubin (T-Bil), peak prothrombine time(PTINR) (T-Bil ρ = −0.3567 p = 0.0189, PT-INR ρ = −0.4260 p = 0.0044). There are only 3 cases have failed in Grade B post hepatectomy liver failure (PHLF in ISGLS criteria). The © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

28

Abstracts

3 cases have refractory ascites, however, they don’t fail in critical liver failure. Duration of hospital stays is 6–72 days (median 12days). Conclusion: Using Rem-KICG, probably we can assess postoperative hepatic function. We conducted hepatectomy safely by using the operable criteria(Rem-KICG > 0.05). The criteria may be useful to indicate the acceptable limit of hepatectomy before operation.

OP-I.21 CHARACTERIZATION OF PANCREATICO-ENTERIC ANASTOMOTIC HEALING IN A PORCINE SURVIVAL MODEL Y. S. Khajanchee1, W. Johnston1, C. B. Bifulco2, M. A. Cassera1, C. W. Hammill1 and P. D. Hansen1 1 Providence Portland Cancer Center, Hepatobiliary And Pancreatic Surgery Program, Portland, OREGON; 2 Providence Portland Cancer Center, Department Of Pathology, Portland, OREGON Objective: Leakage from the pancreatico-enteric anastomosis after pancreatico-duodenectomy is the single most important cause of postoperative morbidity and mortality. Histological studies of ileal and colonic anastomoses have provided valuable insights regarding causes of anastomotic failure. However, this crucial information is lacking for pancreatico-enteric anastomoses. The aim of the current study is to characterize the histology, assess collagen deposition and bursting strength of pancreatico-enteric anastomoses in a porcine survival model during the early postoperative period when dehiscence is most likely to occur. Methods: Under general-anesthesia, pancreaticoduodenectomy was performed followed by two layered endto-side pancreatico-enteral anastomosis. Animals were survived for five-days. Drain amylase was measured on postoperative day (POD) 3 and 5. Animals were euthanized on POD-5 and the pancreatico-enteral anastomosis specimen was resected en-block. Bursting pressure testing was performed using Gastrograffin solution under direct fluoroscopic guidance. Histological sections of the anastomoses were prepared with trichrome staining and were examined under high power light microscopy. Results: Postoperative drain amylase were 10 times greater than serum amylase levels in both animals indicating clinical fistula. However, on bursting pressure testing, no leakage of contrast was noted at pressures >300 mmHg. Gross examination and microscopic examination of the anastomotic site revealed excellent healing with significantly higher quantities of collagen deposition at the pancreatico-enteral anastomotic site as compared to entero-enteral anastomosis. Conclusion: Our initial results indicate that healing of pancreatico-enteral anastomosis is much better than enteroenteral anastomosis during the early post-operative period. High drain amylase level does not necessarily indicate a pancreatic fistula.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

OP-I.22 PREOPERATIVE NEUTROPHIL-TO-LYMPHOCYTE RATIO (NLR) AND SERUM CA19-9 AS SIGNIFICANT PREDICTORS FOR PANCREATIC HEAD CANCER T. Asaoka, A. Miyamoto, K. Yamamoto, N. Haraguchi, M. Miyake, K. Nishikawa, M. Hirao, M. Ikeda, M. Sekimoto and S. Nakamori Department Of Surgery, Osaka National Hospital, Osaka, OSAKA Background: Recently, several preoperative proinflammatory markers and nutritional factors such as neutrophil-tolymphocyte ratio (NLR) and prognostic nutrition index (PNI) have been reported as significant predictor for poor prognosis of various malignant tumors. In this study, we evaluated the prognostic values of these preoperative parameters in patients with resectable pancreatic head cancer. Methods: We retrospectively reviewed consecutive patients who underwent pancreatoduodenectomy for pancreatic head cancer between 2007 and 2012. A total of 46 patients were enrolled in this analysis. Preoperative parameters such as CRP, CA19-9, NLR and PNI at the time of presentation were recorded as well as overall survival. Results: In multivariable analysis preoperative high NLR (>2.7) and high CA19-9 (>232) were independent prognostic factors for poor survival (P value: 0.03 and 0.025 respectively). Kaplan-Meier survival analysis demonstrated the overall 2-year survival rate in patients with high NLR or high CA19-9 were 37.5% compared with 89.9% in patients with low NLR and low CA19-9. Conclusion: Preoperative NLR and serum CA19-9 offer significant prognostic information associated with overall survival following pancreatoduodenectomy in the patients with pancreatic head cancer.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

29

OP-I.24 USE OF GLAND AND TUMOR CHARACTERISTICS AND PREOPERATIVE CT SCANS TO PREDICT PANCREATIC FISTULA AFTER DISTAL PANCREATECTOMY C. Johnston1, L. Harmon2, L. Haykin1, J. Shen1, H. Hahn2, D. Coy2, M. Lasarev3, C. Mathews1, M. Cassera1, M. Brown1, R. Wolf1, P. Hansen1, C. Hammill1, A. Alseidi2 and P. Newell1 1 Providence Portland Cancer Center, Portland, OR; 2 Virginia Mason Medical Center, Seattle, WA; 3Oregon Health And Science University, Portland, OR Background: Postoperative pancreatic fistula (POPF) remains the dominant cause of morbidity following distal pancreatectomy. Preoperative risk stratification of patients undergoing distal pancreatectomy would be useful for treatment algorithms and clinical trials. Methods: Risk factors for POPF in 199 consecutive patients undergoing distal pancreatectomy at two major institutions were recorded retrospectively. Gland density was measured on noncontrast CT scans (n = 101), and histological scoring of fat infiltration and fibrosis was performed by a pathologist based on H&E staining (n = 120). Recursive partitioning was used to construct a decision tree to predict POPF based on eleven variables. Results: 42 patients (21%) developed a clinically significant POPF (ISGPF Grade B or C) within 90 days. Univariate analysis of age, operative approach, method of gland transection, splenectomy, staple line reinforcement, and pathologic diagnosis in 199 patients showed none were predictive of POPF. Fat infiltration was significantly associated with gland density (p = .0013), but density did not predict POPF (p = 0.5). ROC curve analysis did not identify values for gland thickness at the margin that predicted POPF (area under curve = 0.635). Recursive partitioning resulted in a decision tree that predicted POPF in this cohort with a misclassification rate of <15% using gland fibrosis (histology), density (HU), margin thickness (cm), and pathologic diagnosis. Discussion: This multicenter study shows that no single perioperative factor reliably predicts POPF after distal pancreatectomy. However, a novel decision tree was constructed using recursive partitioning that predicts POPF with >85% accuracy. Further work is required to validate the applicability of this model to clinical practice.

HPB 2014, 16 (Suppl. 1), 1–99

OP-I.25 HIGHER LYMPH NODE RATIO PREDICTS EARLY RECURRENCE AFTER CURATIVE RESECTION FOR PANCREATIC ADENOCARCINOMA E. W. Gilbert, N. W. Bronson, T. Kaur, K. Billingsley and B. C. Sheppard Oregon Health & Science University, Portland, OR Introduction: The majority of patients who undergo resection for pancreatic adenocarcinoma (PDAC) are diagnosed with recurrence. In this study we aim to characterize the pattern of recurrence and the predictors of early recurrence following curative resection. Methods: Patients with PDAC surgically resected between 1996 and 2010 were retrospectively reviewed and categorized into early (<12 months) or late (>/=12 months) recurrence. Multivariate Cox proportional hazard models were used to estimate the effects of variables on pattern of recurrence. Survival estimates were calculated using Kaplan Meier analysis. Results: From 1996–2010, 256 patients underwent PDAC resection at a single-center. Overall median survival was 17 months (95% CI = 15–22) and disease free survival was 15 months (95% CI = 13–21). Recurrence was documented in 107 patients at a median of 315 days (IQR = 154–471). Thirty-two (30%) recurred regionally, 69 (64%) had distant metastasis and 6 (6%) had both. In multivariate analysis, R1 resection, lymph node positivity and a higher lymph node ratio carried an increased risk of recurrence (p < 0.05 for all). Patients with early (64, 60%) compared to late recurrence (43, 40%) were significantly more likely to have a distant recurrence (72% vs. 53%, p = 0.037), a trend toward lymph node positivity (70% vs. 53%, p = 0.076) and to have a higher median lymph node ratio (0.167 vs. 0.050, p = 0.04; Table 1). Conclusion: A higher ratio of positive over total number of lymph nodes predicts early recurrence following surgical resection for pancreatic adenocarcinoma. Patients who recur early are more likely to be diagnosed with a distant metastasis. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

30

Abstracts

OP-I.26 SINGLE-INSTITUTION EXPERIENCE WITH RESECTION OF SIDE-BRANCH IPMN: A REVIEW OF PRE-OPERATIVE CHARACTERISTICS AND FINAL PATHOLOGIC DIAGNOSIS J. D. Dortch, J. A. Stauffer and H. J. Asbun Mayo Clinic-Florida , Jacksonville, FL Background: Given the malignant potential of main duct IPMN (M-IPMN) surgical resection is generally indicated. With regard to side branch IPMN (SB-IPMN), resection versus observation is a topic of debate. We present a single institution experience with surgical resection of SB-IPMN. Methods: A retrospective review of all patients undergoing resection for IPMN from 2002–2013 was performed. Our specific focus was on isolated SB-IPMN with non-malignant FNA cytology. Preoperative clinical characteristics and diagnostic selection criteria were reviewed to determine association with pathologic diagnosis and clinical outcomes. Results: A total of 147 patients undergoing resection for IPMN were identified. Of these, 80 patients (54%) had a component of M-IPMN or invasive disease on FNA, leaving 67 (46%) patients with SB-IPMN and non-malignant cytology. The results of the latter group are summarized in the table. Ductal adenocarcinoma within SB-IPMN was found in 9 patients (13%). The mean diameter (cm) of benign SB-IPMN was 2.0 (Range: 0.3–5.7) (SD ± 1.1) vs. that of malignant © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

SB-IPMN which was 3.1 (Range: 1.5–6.0) (SD ± 1.3) (p = 0.014). Of the 9 patients found to have malignant SB-IPMN, symptoms, mural nodules, and septations were found in 6 (67%), 2 (22%), and 5(56%), respectively. Tumor staging was as follows: IA (2) IB (2) 2A (4) 2B (1). 1/9 (11%) patients with malignant SB-IPMN died during the study period. Conclusion: With proper selection criteria, SB-IPMN is associated with a low rate of invasive pancreatic ductal adenocarcinoma at the time of resection. Nevertheless, given the demonstrated incidence of malignancy, appropriate operative candidates should undergo resection.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

OP-I.27 PANCREATIC CANCER METASTATIC TO A LIMITED NUMBER OF LYMPH NODES DOES NOT ADVERSELY AFFECT OUTCOME M. Kukar, S. W. Badwin, K. Attwood, N. Wilkinson, S. N. Hochwald and B. Kuvshinoff II Roswell Park Cancer Institute, Buffalo, NY Metastatic nodal disease in pancreatic cancer is generally associated with poor prognosis; however the current AJCC staging does not take into account the number of positive lymph nodes or their ratio. This study aims to relate the extent of metastatic lymph node involvement with survival in pancreatic cancer. A prospective database identified patients undergoing resection for pancreatic adenocarcinoma from 1999–2009. Patient characteristics were compared using Wilcoxon or Kruskal Wallis test for continuous variables, while categorical variables were compared using Fisher’s exact or chi-square test. The survival outcomes

OP-I.28 VARIATIONS IN MEDICAL ONCOLOGY UTILIZATION PRACTICES BY PANCREATIC CANCER PATIENTS IN A SINGLE PAYER HEALTH CARE SYSTEM S. Hurton1,2, G. Porter1,2, M. Walsh1 and M. Molinari1,2 Department Of Surgery, Dalhousie University, Halifax, NS; 2Department Of Community Health & Epidemiology, Dalhousie University, Halifax, NS; 3Cancer Outcomes Research Program, Cancer Care Nova Scotia, Halifax, NS

1

Population-based studies in the USA have shown that pancreatic cancer (PC) has the lowest rates of cancer-directed therapy among solid organ malignancies. Socio-economic factors, referral patterns and unequal access to health-care might be responsible. Few studies examine PC patients living in regions where health care is uniformly available. Main aim of this study was to assess PC-directed therapy in a cohort of patients diagnosed with PC in Nova Scotia (Canada). A cohort of 1129 patients with PC was identified using the Provincial Cancer Registry over a 10-year period (2001– 2010). Demographic, clinical and socio-economic data were HPB 2014, 16 (Suppl. 1), 1–99

31

were summarized using Kaplan-Meier methods. Patients (N = 165) were divided into 3 groups based on the number of positive lymph nodes {0 (Group A), 1–2 (Group B), >3 (Group C)}. Group A (n = 55) and B (n = 55) were similar, however patients in Group C (n = 55) were more likely to have a higher T stage, be poorly differentiated, have lymphovascular invasion, higher mean intraoperative blood loss, have positive margins, in the uncinate process, and more likely to have a Whipple. Median overall survival for patients in each group was 25.5 months (A), 21 months (B) and 12.3 months (C) (p < 0.001, Figure 1a), however no difference was noted for survival between A and B (p = 0.86). The ratio of involved lymph nodes less than 0.2 was also predictive of improved survival (p < 0.001, Figure 1b). Resected pancreatic cancer patients with 1–2 positive lymph nodes or less than 20% involvement have a similar prognosis to patients without nodal disease. Accurate staging should consider stratification of the extent of nodal involvement.

extracted through the linkage of multiple administrative databases. Predictors of medical oncology health-care utilization in PC patients were explored by multivariate analyses. Curative resection was performed in 163 patients (14.4%) and palliation in 996 (85.6%). Of all resected patients, 45% were referred to medical oncology for adjuvant therapy and 34.4% of patients had a visit within 10 weeks. Among unresected patients, 31% were referred for palliative chemotherapy. There was greater attrition between referral to medical oncology and a visit to medical oncology in resected patients vs. unresected patients (14.1% vs. 4.6%). When adjusted for socioeconomic status, distance from medical center and rurality, only advanced age (P = 0.0092) and presence of comorbidities (P = 0.0029) were significant predictors for failure of PC-directed therapy (Table 1). The majority of patients with PC did not receive cancer directed therapy. Socioeconomic characteristics and distance to medical centers did not have a significant effect in utilization of adjuvant or palliative oncological treatments.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

32

Abstracts

OP-I.29 LONG TERM OUTCOMES FAVOR PYLORUS PRESERVING PANCREATICODUODENECTOMY OVER DUODENUM PRESERVING PANCREATIC HEAD RESECTION FOR CHRONIC PANCREATITIS: A META-ANALYSIS AND SYSTEMATIC REVIEW P. B. Sukharamwala1, S. B. Ross1, S. Parikh1, C. E. Ryan1, T. W. Wood1 and A. S. Rosemurgy1 1 Florida Hospital Tampa, Tampa, FL; 2American University Of Antigua, New York, NY Introduction: Many consider the pancreatic head the epicenter of chronic pancreatitis. Pylorus preserving pancreaticoduodenectomy (PPPD) and duodenum preserving pancreatic resection (DPPHR) aimed at resecting pancreatic head are important treatment options. The aim of this

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

Meta analysis is to compare the long term outcomes of PPPD versus DPHR in patients with chronic pancreatitis. Methods: A systematic literature search was conducted using Embase, MEDLINE, Cochrane and PubMed databases on all studies published between January 1991 and January 2013 reporting intermediate and long term outcomes after PPPD and DPHR for chronic pancreatitis. Outcomes of interest were complete pain relief, quality of life, professional rehabilitation, exocrine insufficiency and endocrine insufficiency. Results: Eleven studies (seven RCT) were included comprising 692 patients. There was no significant difference in complete pain relief between PPPD and DPPHR (odds ratio (OR) 95% CI 0.82 [0.52, 1.30]; P = 0.43). However quality of life (WMD 95% CI -12.25 [-15.20, -9.30] ; P < 0.00001) and professional rehabilitation (OR 95% CI, 0.47 [0.28, 0.79]; P = 0.004) were significantly better for patients undergoing DPPHR compared to PPPD. This can be, in part, attributed to significantly reduced exocrine insufficiency (OR 95% CI, 2.25 [1.43, 3.55]; P = 0.004) and endocrine insufficiency (OR 95% CI, 1.74 [1.10, 2.77]; P = 0.02) in patient who underwent DPPHR. Conclusions: There is no significant difference in postoperative pain relief in patients undergoing PPPD versus DPPHR. Improved long term outcomes including quality of life, professional rehabilitation and preservation of exocrine and endocrine function, make DPPHR procedure more attractive than PPPD for patients with chronic pancreatitis.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

33

OP-I.30 NEOADJUVANT CHEMORADIATION DECREASES LYMPH NODE RATIO IN BORDERLINE RESECTABLE PANCREATIC DUCTAL ADENOCARCINOMA

OP-I.31 DELAYED GASTRIC EMPTYING FOLLOWING PANCREATICODUODENECTOMY: INCIDENCE, RISK FACTORS, AND HEALTHCARE UTILIZATION

J. S. Peng, N. Ali, G. Falk, D. Joyce, C. O’Rourke, J. McMichael, J. Wey, S. Chalikonda, K. M. El-Hayek and R. M. Walsh Cleveland Clinic Foundation, Cleveland, OH

S. Mohammed1,2, G. Van Buren1,2, A. McElhany1,2, M. Issazadeh1,2, E. J. Silberfein1, D. A. Anaya1, S. E. Hodges1,2 and W. E. Fisher1,2 1 Baylor College Of Medicine, Houston, TX; 2Elkins Pancreas Center, Houston, TX

Background: In patients undergoing pancreaticoduodenectomy (PD) for pancreatic ductal adenocarcinoma (PDAC), lymph node metastasis and higher lymph node ratio adversely affect prognosis. This study aims to examine the impact of neoadjuvant therapy on lymph node and margin status in borderline resectable PDAC. Methods: A neoadjuvant protocol was established at our tertiary care institution in 2012 for all patients with borderline resectable PDAC. These patients underwent diagnostic laparoscopy and eligible patients received neoadjuvant chemoradiation with a gemcitabine or 5-fluorouracil based protocol. Patients were subsequently re-staged and those without progression of disease underwent curative resection. Fourteen patients who underwent PD after neoadjuvant chemoradiation were compared to a control group of 28 patients who underwent PD without neoadjuvant therapy, matched for age, sex, and tumor size. Baseline demographic, operative and histopathologic characteristics were analyzed. Results: The neoadjuvant and control groups were comparable with respect to the number of lymph nodes harvested (17.3 vs 15.5, p = 0.45), operative time (509 vs 440 minutes, p = 0.11), blood loss (664 vs 476 mLs, p = 0.13), and postoperative length of stay (11.6 vs 10.9 days, p = 0.76). In the neoadjuvant group, there were more R0 resections (86 vs 61%; P = 0.043), more N0 stage (71 vs 21%; P = 0.002), and a lower lymph node ratio (mean 0.045 vs 0.19; P = 0.03). Conclusion: This single institution study demonstrates that neoadjuvant chemoradiation in borderline resectable PDAC shows favorable histopathologic outcomes, including increased R0 resection, N0 stage, and lower lymph node ratio. These results suggest a more favorable long-term prognosis with neoadjuvant therapy.

Background: Delayed gastric emptying (DGE) is a frequent complication following pancreaticoduodenectomy (PD). Further examination of DGE and understanding its implications on healthcare utilization is important. This study characterizes incidence and risk factors for DGE following PD and examines healthcare utilization. Methods: A prospectively-maintained database was retrospectively reviewed. DGE was classified using the International Study Group of Pancreatic Surgery consensus definition. Subjects who developed DGE and those who did not were analyzed using univariate and multivariate models. Results: 276 patients underwent PD (>80% pyloruspreserving with antecolic reconstruction and no enteric feeding tubes). DGE developed in 49 patients (17.8%); 9.1% grade A, 5.1% grade B, and 3.6% grade C. Demographic, clinical, and operative variables were similar between patients with DGE and those without. DGE patients were more likely to experience multiple complications (32.6% vs 4.4% with ≥3 complications, p < 0.001), including pancreatic fistula (PF) (42.9% vs 18.9%, p = 0.001) and intraabdominal abscess (IAA) (16.3% vs 4.0%, p = 0.012). Only 9 (18.4%) patients experienced isolated DGE. Patients with DGE had a longer hospital stay (median 12 vs 7 days, p < 0.001) and were more likely to be discharged to transitional care facilities (24.5% vs 6.6%, p < 0.001). On multivariate analysis, presence of pancreatic fistula (OR 5.02, p < 0.001) was a significant predictor for DGE development. Conclusion: Although DGE occurs in <20% of PD patients, it is associated with significant increase in healthcare utilization postoperatively. Patients with PF are at highest risk for developing DGE and should be the target of future interventions to improve management of DGE.

OP-I.32 PORTAL VEIN RESECTION IS ASSOCIATED WITH IMPROVED SURVIVAL AFTER PANCREATICODUODENECTOMY FOR PANCREATIC CANCER P. B. Sukharamwala, K. D. Patel, R. L. Klein, T. W. Wood, A. F. Teta and A. S. Rosemurgy Florida Hospital Tampa, Tampa, FL Introduction: Pancreatic cancer with close proximity to the portal or superior mesenteric veins is no longer considered a contraindication to pancreaticoduodenectomy. With improvements in staging, perioperative care, and surgical experience, portal vein resection is undertaken more frequently. This study was undertaken to compare long-term survival after pancreaticoduodenectomy with or without portal vein-superior mesentric vein resection. HPB 2014, 16 (Suppl. 1), 1–99

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

34

Abstracts

Methods: An Embase, MEDLINE, Cochrane, and PubMed search was undertaken to identify published studies that report long-term outcomes after portal vein-superior mesenteric vein resection during pancreaticoduodenectomy for pancreatic carcinoma from 1997 through 2013. Survival rates between patients with portal vein resection and patients without portal vein resection were compared. Data were analyzed using hazard ratios and standard error. The hazard ratios for all trials were pooled using an inverse variance method. Results: 21 studies compared survival with versus without portal vein-superior mesenteric vein resection during pancreaticoduodenectomy for pancreatic carcinoma. Patients

who underwent pancreaticoduodenectomy with portal vein resection had better survival compared to patients who underwent pancreaticoduodenectomy without portal vein resection (hazard ratio (OR) 95% CI 0.89 [0.80, 0.98], p = 0.02). Conclusion: Long-term survival after pancreaticoduodenectomy with portal vein-superior mesenteric vein resection can be achieved. Given the improved survival rates associated with pancreaticoduodenectomy and portal veinsuperior mesenteric vein resection, aggressive surgical resection and reconstruction should be undertaken for suspected portal vein-superior mesenteric vein invasion during pancreaticoduodenectomy for pancreatic carcinoma.

OP-I.34 ROLE OF SURGICAL PALLIATION IN PATIENTS WITH PANCREATIC DUCTAL ADENOCARCINOMA (PDAC)

underwent SP, most often (N = 96, 78%) when initial intent was curative resection. Of these, 75 (78%) patients had locally advanced disease at operation, and 21 (22%) had metastasis. 27 (of 123) patients underwent SP because endoscopic palliation was not possible, and/or for tissue diagnosis. The majority of SP patients underwent biliary and/or gastric bypass (N = 93, 75%). SP and diagnostic-only patients were equally likely to start chemotherapy (84v81%, p = 0.71), but time to initiation was longer (median 44v31d, p = 0.013). Complications occurred (Table 1) but most were minor (Clavien-Dindo 1–2, 78%). 30 d readmission rate was 11.3% (14 patients). Median survival was slightly longer for SP patients, but was not statistically significant (11.2vs9.5 m, p = 0.239). Conclusion: Most complications following SP tended to be minor, although major morbidity/mortality did occur. Overall, SP patients were equally likely to enter palliative chemo/radiotherapy and survival was slightly, but not significantly, longer. Further study is warranted of the durability of SP, particularly bypass, in the context of contemporary ERCP capabilities, chemotherapy regimens, and survival.

A. Abbasi1, R. Lo1, A. J. Moser1, J. F. Tseng1, C. M. Vollmer2, M. P. Callery1 and T. S. Kent1 1 Beth Israel Deaconess Medical Center Department Of Surgery, Boston, MA; 2University Of Pennsylvania Department Of Surgery, Philadelphia, PA Introduction: With widespread availability of endoscopic palliation for unresectable PDAC, fewer patients undergo surgical palliation (SP). Contemporary data on this subgroup of patients are limited. This study investigated outcomes for surgically palliated patients, with a goal of determining its role amongst palliative options for these patients. Methods: Retrospective chart review was conducted in a high-volume pancreatic surgery institution. Patients included those with PDAC (2003–11) who underwent diagnostic or palliative operations only. Results: Of 193 patients, 70 underwent diagnostic procedures only, mainly staging laparoscopy. 123 (64%) patients © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

35

OP-I.36 REGIONALIZATION OF HEPATO-PANCREATO-BILIARY CANCER TREATMENT WITHIN THE STATE OF FLORIDA ACCORDING TO THE AMERICAN COLLEGE OF SURGEONS COMMISSION ON CANCER ACCREDITATION

OP-I.35 PROPHYLACTIC ANTICOAGULATION FOR SUPERIOR MESENTERIC VEIN (SMV) OR PORTAL VEIN (PV) NARROWING/OCCLUSION IN BORDERLINE RESECTABLE (BLR) PANCREATIC ADENOCARCINOMA (PC) PATIENTS A. N. Krepline, K. Duelge, B. George, P. Ritch, B. A. Erickson, A. Mahmoud, E. J. Quebbeman, K. K. Turaga, F. M. Johnston, T. C. Gamblin, K. K. Christians, D. B. Evans and S. Tsai Medical College Of Wisconsin, Milwaukee, WI Background: BLR PC patients (pts) with significant SMV/PV narrowing are at risk for developing mesenteric venous thrombosis (MVT) which may prevent eventual pancreatectomy. Prophylactic anticoagulation in this population has unknown benefit. Methods: From 2007–2013, BLR PC pts receiving neoadjuvant therapy were classified by >50% narrowing of the SMV/PV, SMA abutment, marginal performance status, Ca19–9 >2000 U/mL, or having lesions indeterminate for metastases. Anticoagulation status, radiographic measurement of venous diameter, and clinicopathologic data were abstracted. Results: Eighty-four BLR PC pts were identified; 42 (50%) had SMV/PV narrowing, 8 (10%) had short segment venous occlusion, 24 (28%) had SMA abutment alone, and 10 (12%) were BLR by other criteria. Of the 42 pts with SMV/PV narrowing, the mean vessel diameter was 5.6 mm (standard deviation 2.7 mm). Of the 50 pts with either SMV/PV narrowing (42) or occlusion (8), 35 (70%) received either prophylactic or therapeutic anticoagulation during their neoadjuvant treatment. MVT developed in 3 (6%) pts. Two developed MVT on prophylactic anticoagulation; one underwent pancreaticoduodenectomy with vein resection and one had metastatic disease at laparoscopy. The third pt received therapeutic anticoagulation for short segment venous occlusion but propagated the thrombus and was unresectable. Thirty-nine (78%) of the 50 pts with SMV/PV narrowing/ occlusion completed all neoadjuvant therapy including resection, compared to the 21 (62%) of 34 pts with arterial only or other BLR reason (p = 0.11). Conclusion: MVT may develop from existing narrowing/ occlusion of SMV/PV despite prophylactic anticoagulation. The utility of therapeutic anticoagulation requires further assessment given the increasing application of vascular resection/reconstruction for BLR pts.

HPB 2014, 16 (Suppl. 1), 1–99

N. Agee1, P. Veldhuis 1, S. Eubanks1, P. Arnoletti1,2 and S. G. De La Fuente1,2 1 Florida Hospital Orlando, Orlando, FL; 2University Of Central Florida, Orlando, FL Background: Evidence demonstrates improved outcomes among hepato-pancreato-biliary (HPB) cancer patients treated at high-volume institutions. We analyzed the degree of HPB cancer treatment regionalization within the state of Florida using the Florida Agency for Health Care Administration (AHCA) database. Methods: HPB cancer patients treated at an inpatient facility were identified between April 2011 through March 2012. Analyzed records included all patients diagnosed with HPB malignancy within the study period (MS-DRG codes 435, 436 and 437) and associated total hospital charges. Facilities were subsequently stratified by the American College of Surgeons (ACS) Commission on Cancer Accreditation (CoC), which is based on the organization type, services provided, and cases accessioned. Results: 283 hospitals were identified from the database. Treatment of HPB cancer patients was centralized to 60 institutions distributed throughout the state. Forty percent of these facilities had an ACS Comprehensive Community Cancer Program (CCCP) designation, while 12% were Academic Comprehensive Cancer Program (ACAD) designated, 10% were designated as an Integrated Network Cancer Program (INCP), 3% as a National Cancer Institute Designated Comprehensive Cancer Program (NCIP), and 23% (n = 14) had no CoC accreditation. The majority of patients were treated at a CCCP designated program with average total charge over two million dollars and a mean length of stay of 6.0 days. Conclusion: In Florida, most patients are treated at Comprehensive Community Cancer Programs that provide full range of diagnostic and treatment services and participate in cancer-related research. Based on these findings, Florida displays adequate regionalization of HPB cancer care to high volume centers.

OP-I.37 OBESITY IS NOT ASSOCIATED WITH PANCREATIC CARCINOMA OR INPATIENT MORTALITY AFTER PANCREATIC RESECTION E. S. Glazer and E. S. Ong The University Of Arizona, Tucson, AZ Introduction: The interaction between obesity and pancreatic carcinoma (PC) is incompletely understood. The purpose of this study was to investigate the relationship between obesity and PC using the Nationwide Inpatient Sample (NIS), a representative sample of patient-discharge data. Methods: We reviewed the NIS database for years 2009– 2010 using ICD-9 diagnostic codes for PC & obesity, © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

36

Abstracts

inpatient mortality, demographic data, and procedure codes. Student’s t-test determined statistical significance (P < 0.05). Results: 7.8% of all patients had a diagnosis of obesity (N = 1,230,496); 0.16% had a diagnosis of PC (N = 25,585). The average age was 63 years with 59% female. Obese patients were more likely to be treated at teaching hospitals than non-obese patients (48% vs. 44%, P < 0.001). Nonobese patients had a higher rate of PC than obese patients (0.2% vs. 0.1%, P < 0.0005). 8% of patient-discharges with PC underwent pancreatic resection (PR). For those who died after PR for PC, obese patients were younger than non-obese patients (62 years vs. 71 years, P = 0.05). The mortality rate of obese PC patients not undergoing PR was 44% lower than for non-obese patients (P = 0.003). For PC patients undergoing PR, obese patients had similar mortality rates to nonobese patients (6% vs. 4%, P = 0.4). Conclusion: Despite limitations in the NIS, obese patients are diagnosed with PC at a lower rate than non-obese patients and have lower mortality rates. There was no statistically significant difference in inpatient mortality rates after PR for PC based on obesity.

OP-I.38 PANCREATODUODENECTOMIES IN COLOMBIA: A QUALITATIVE STUDY OF THE PANCREATIC SURGERY FROM THE SURGEONS PERSPECTIVE J. Sabogal Olarte, C. G. Ball, F. Sutherland, O. Bathe and E. Dixon Hepatobiliary And Pancreatic Department. Foothills Medical Centre – University Of Calgary, Calgary, AB Introduction: The Whipple procedure has been historically performed by general surgeons in Colombia, Subspecialists, have become important participants in this field over the past 10 years however. Objectives: This qualitative study examines surgeons’ attitudes towards the Whipple operation and the possibility of regionalized care models with regards to pancreaticoduodenectomy in Colombia. Methodology: An on-line survey was created with 10 questions addressed to surgeons who performed pancreatic surgery. Data was obtained from the Colombian Association of Surgery. Individual surgeons in different cities and hospitals around the country were contacted via e-mail. Results: 132 e-mails were delivered with 52 surveys being completed (40%) by General surgeons, surgical gastroenterologists, HPB surgeons and surgical oncologists. Whipples performed per surgeon per year were less than 12 cases in 90% of participants. Self reported mortality was more 10% in 59% of participants and 30-day morbidity was more 20% in 63% of participants. Pancreatic surgery is performed most commonly by general surgeons (59.52%). The majority of participants (84.31%) believe these patients are ideally treated in referral centres for pancreatic surgery and the majority (75%) believe pancreatic surgery should be performed by hepatobiliary surgeon. Conclusion: Pancreatic surgery in Colombia is increasingly common. Care of these patients in regional high volume centers is beginning to occur. There is still significant room for improvement with regards to peri-operative morbidity and mortality. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

OP-I.40 PANCREAS TRANSPLANTATION OUTCOMES DIFFER BASED ON AGE GROUPINGS: AN ANALYSIS OF THE UNOS DATABASE E. Siskind1, C. Maloney1, M. Akerman1, A. Shen1, A. Basu1, E. Molmenti1 and J. Ortiz2 1 North Shore Long Island Jewish Health System-Department Of Transplantation, Manhasset, NY; 2 Albert Einstein Medical Center, Philadelphia, PA The analysis of pancreas transplantation outcomes based on age groupings has largely been based on single center reports. In our study, the UNOS database on all adult pancreas and kidney-pancreas transplant patients between 1996 and 2012 were analyzed (n = 20,854). Patients divided into groups based on age categories: 1.18-29 (n = 1823) 2.30-39 (n = 7624) 3.40-49 (n = 7967) 4.50-59 (n = 3160) 5. ≥ 60 (n = 280) Kaplan-Meier analysis demonstrated a difference in patient survival (p < 0.0001) and death-censored graft survival (p < 0.0001) among the age category groups. The following tables include 1-, 3-, 5-, 10-, and 15-year unadjusted patient survival rates and graft survival rates for each age category group. Patient Survival 18–29 30–39 40–49 50–59 ≥ 60 1-year 95.4% 96.0% 94.9% 93.3% 91.0% 3-year 91.3% 91.9% 90.5% 87.7% 80.2% 5-year 86.3% 87.8% 85.7% 81.6% 71.4% 10-year 73.5% 76.8% 71.8% 61.5% 42.5% 15-year 65.3% 65.6% 56.1% 41.8% 0.0% Graft Survival 18–29 30–39 40–49 50–59 ≥ 60 1-year 80.7% 82.7% 82.9% 83.8% 82.3% 3-year 67.2% 73.3% 75.3% 74.8% 69.8% 5-year 56.8% 65.9% 67.8% 67.4% 59.9% 10-year 39.6% 48.2% 52.2% 47.5% 30.2% 15-year 26.6% 33.4% 37.2% 29.1% 0.0% Graft survival was consistently the greatest in the 40–49 age group. Graft survival was least in the 18–29 age group at 1, 3, and 5 year intervals. At ten year and fifteen year intervals graft survival was least in the age >60 group. Patient survival and age were inversely proportional; as age increased, patient survival decreased. This analysis differs from previous studies which reported equivalent graft and patient survival amongst age groups.

OP-I.42 INTERVENTIONAL RADIOLOGY STENTING AND TPA FOR HEPATIC ARTERY THROMBOSIS IN LIVER TRANSPLANTATION REVISITED: A FALSE HOPE WITH FREQUENT NEED FOR RETRANSPLANTATION C. W. Michalski, J. Pasko, K. Roayaie, D. L. Scott and S. L. Orloff Oregon Health And Science University, Portland, OR Objective: Hepatic artery thrombosis (HAT) remains one of the most devastating complications following orthotopic liver transplantation (OLT). Endovascular techniques addressing post operative HAT have recently become more prevalent, however, there have been few reports of successful outcomes. Methods: We retrospectively reviewed a prospective database for patients with HAT after OLT. All patients transplanted between 1992 and 2012 were analyzed. Recipients of Donation after Cardiac Death (DCD) organs were excluded. HPB 2014, 16 (Suppl. 1), 1–99

Abstracts Results: Within the study period, 1065 OLTs were performed at our institution. Twenty patients developed HAT (1.88%). Etiology of liver disease was PSC (5 pts), HBV (1), HCV (5, (2-HCV/ETOH)), HBV&HCV (2), ETOH (2), and other (5). Median time from OLT to HAT diagnosis (angiography or CT angio) was 30 days. Six patients underwent direct re-transplantation, 9 patients were treated with TPA+stenting and 2 patients required no treatment (collaterals had formed). TPA+stenting was not technically feasible in 3 patients. Re-OLT was performed in 7/9 HAT patients where TPA+stenting was the initial therapy. Thus, in total 13/20 (65%) patients were re-transplanted. Median time to re-transplantation was 11 months. Biliary interventions were required in 10 patients: 7 PTC, 2 ERCP, 1 surgical revision of choledochojejunostomy. Median survival following OLT was 94 months. Conclusion: Failure rates for TPA+stenting for HAT are high and eventual re-transplantation was required in the majority of patients. These data raise the question of the value, both in monetary cost and in morbidity to the patient, of an interventional radiology approach in the treatment of HAT after OLT.

OP-I.43 HPB SURGERY FROM SCRATCH: A SINGLE SURGEON’S ONE YEAR EXPERIENCE IN ESTABLISHING A HIGH VOLUME HPB CENTER J. S. Cardinal and S. Cho West Virginia University, Morgantown, WV Introduction: Previous reports suggest current levels of fellowship training will result in an excess of hepato-pancreatobiliary (HPB) trained surgeons by the year 2020. “High volume” HPB centers are defined as those performing over 40 cases per year. This report describes a single surgeon’s first year post fellowship experience in establishing a “high volume” HPB center. Methods: Case logs and clinic schedules were reviewed. HPB cases were defined as those involving liver, pancreas, gallbladder, bile duct or duodenum. Cholecystectomy was excluded. Operative outcomes were collected and referral patterns were tracked. Results: From August 2012 – August 2013, 37 HPB cases were performed (15 pancreatic, 14 hepatic, 8 other). There were 169 new patient encounters relating to HPB problems. For operative outcomes, average OR time was 380 min. Average EBL was 460 mL. Eight patients received blood transfusion(s) (range 1–8 units PRBCs). Average length of stay was 13 days (range 2–74). 51% of patients experienced complications, including 2 deaths which involved cases of necrotizing pancreatitis and penetrating trauma to the porta hepatis, respectively. Ninety-seven of the 169 (57%) patients seen in clinic for HPB related problems were referred by physicians from inside the health system. Conclusion: Establishing a “high volume” HPB practice directly out of fellowship is an achievable goal. Careful patient selection so as to ensure excellent surgical outcomes is the most critical aspect to practice building. Other important tenants include capturing the majority of the in house HPB related cases as well as robust communication with outside referring doctors. HPB 2014, 16 (Suppl. 1), 1–99

37

OP-I.44 SURGERY TO THE EXTREME: THE CONCEPT OF TRANSPLANT ONCOLOGY T. Hibi, Y. Abe, O. Itano, H. Obara, M. Shinoda, M. Kitago, H. Yagi and Y. Kitagawa Department Of Surgery, Keio University School Of Medicine, Tokyo, TOKYO Introduction: Extreme cases of abdominal pathology are a significant challenge both from the technical and tumor biological perspective. We describe 3 patients who underwent extended resections under the principle of “no-touch isolation” and by using techniques derived from liver/ multivisceral transplant and organ procurement. Methods: Case 1: A 69-yo woman presented with massive retroperitoneal recurrence of a liposarcoma. After creating an axillo-femoral bypass, en-bloc resection with a Whipple procedure, concurrent right hemicolectomy and nephrectomy, and excision of the infrarenal aorta and inferior vena cava was performed. Case 2: A 36-yo man developed tumor thrombi in the main portal trunk and superior mesenteric vein after a Whipple procedure for a neuroendocrine tumor. We established a portal bypass between the tributary of the superior mesenteric vein and right gonadal vein. The portomesenteric vein with tumor thrombi was excised and reconstructed with an autologous right femoral venous graft. Case 3: A 38-yo woman suffered ruptured renal cell carcinoma of the left kidney with a tumor thrombus obstructing the left renal vein. The spleen, pancreas, and left kidney were mobilized en-bloc to avoid tumor manipulation. Extended left nephrectomy with distal pancreatectomy and splenectomy was performed without exposing the tumor capsule. Results: Major complications included postoperative bleeding and chylous ascites but no in-hospital mortalities occurred. All patients resumed normal daily life with no signs of recurrence during a follow-up period of 3–6 months. Conclusion: The integration of surgical oncology and organ transplant, “Transplant Oncology,” has opened up new horizons in the field of surgery.

OP-I.45 AMPULLARY AND EXTRA-AMPULLARY DUODENAL ADENOCARCINOMAS HAVE COMPARABLE SURVIVAL AFTER POTENTIALLY CURATIVE RESECTION K. H. Dinh, V. Bathini, B. Switzer, V. Kasturi, W. Wassef, J. Zivny, M. Sullivan, G. Whalen and J. LaFemina University Of Massachusetts Medical School, Worcester, MA Background: There is little data comparing outcomes of ampullary and extra-ampullary duodenal adenocarcinomas. The aim of this study is to evaluate survival and prognostic factors of ampullary and extra-ampullary duodenal adenocarcinomas. Methods: Adults who underwent surgical intervention for a duodenal adenocarcinoma from 2002–2012 were identified from a prospectively-maintained database. Results: Twenty-seven patients (ampullary n = 22, extraampullary n = 5) had an operation for duodenal adenocarcinoma. Twenty-six (96%) underwent potentially curative © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

38

Abstracts

resection, including 21 (81%) for ampullary and 5 (19%) for extra-ampullary. There was no significant difference in 5-year overall survival (OS), which was 70% in the entire cohort, 72% in the ampullary group and 67% in the extraampullary group (Fig. 1). Similarly, 5-year recurrence-free survival (RFS) was not significantly different between groups: 56%, 52% and 75% respectively. There was no difference in gender (p = 1.00), age at diagnosis (p = 0.62), number of lymph nodes harvested (p = 0.54), or margin status (R0 resection, 100%). Tumor size (OS: p = 0.60; RFS: p = 0.64), stage (p = 0.16; p = 0.91), and presence of positive nodes (p = 0.67; p = 0.53) were not associated with OS or RFS. Chemotherapy (p = 0.62) was not associated with OS. Conclusion: Ampullary and extra-ampullary duodenal adenocarcinomas have similar overall survival and recurrence-free survival following potentially curative surgical resection.

ruptured paraumbilical hernias over a 6-year period. Appropriate patient resuscitation and optimisation with intravenous fluids, prophylactic antibiotics and local measures, such as non occlusive dressings, was instituted. One failed attempt for conservative management was followed by a successful primary repair. In all cases a primary repair with nonabsorbable Nylon, interrupted sutures was performed. No mesh was used. The perioperative complication rate was 25% and the recurrence rate 8.3%. No mortality was recorded. Mean length of hospital stay was 15.5 days (range: 4–31 days). Conclusion: Based on our experience, the management of ruptured paraumbilical hernias in patients with advanced hepatic cirrhosis and refractory ascites is feasible without the use of TIPS routinely in the preoperative period, provided that meticulous patient optimisation is performed.

OP-I.47 A MULTI-DETECTOR CT ANALYSIS OF THE ABDOMINAL WALL STRUCTURE AROUND THE UMBILICUS WITH SPECIAL REFERENCE TO THE DEHISCENCE OF THE LINEA ALBA: COULD IT BE AN UNDER-RECOGNIZED PREDISPOSING FACTOR FOR ADULT UMBILICAL HERNIA? A. Shimizu1,2, U. Mitsuhashi1, M. Tanaka1 and Y. Yasuda2 Koganei Central Hospital, Shimotsuke, TOCHIGI; 2Jichi Medical University, Shimotsuke, TOCHIGI

1

OP-I.46 TIPS IS NOT NECESSARY FOR THE SUCCESSFUL MANAGEMENT OF RUPTURED PARAUMBILICAL HERNIAS IN PATIENTS WITH ADVANCED HEPATIC CIRRHOSIS AND REFRACTORY ASCITES N. A. Chatzizacharias, J. A. Bradley, E. Huguet, A. Jah, N. V. Jamieson, R. K. Praseedom and P. Gibbs Department Of HPB And Transplant Surgery, Addenbrooke’s University Hospital, Cambridge, UK, Cambridge, CAMBRIDGESHIRE Introduction: Acute umbilical hernia rupture in patients with hepatic cirrhosis and ascites is an unusual, but potentially life-threatening complication. The postoperative morbidity has been evaluated as high as 71%, with mortality rates varying between 60–80% after supportive care and 6–20% after urgent surgical repair. Management options include primary surgical repair with or without concomitant portal venous system decompression. Recently, the routine use of transjugular intrahepatic portosystemic shunting (TIPS) in the preoperative setting resulted in improved perioperative and longer-term results. We present our centre’s experience in this area. Results: Eleven patients (mean age: 52,5 years, range: 36–63 years) with advanced hepatic cirrhosis (Childs B and C) and refractory ascites were treated in our institution for © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

Introduction: Most adult umbilical hernias that hepatobiliary and general surgeons treat are attributed to ascites, obesity and multi-parity. Patients without these factors, however, occasionally present with this condition. Therefore other cause must be sought. To date little is investigated about the radiological anatomy of the umbilical structures. Purpose: The aim of this study is to address the incidence of the linea alba dehiscence at the umbilicus using multi-detector computed tomography (CT) and explore its relation to the patients’ habitus and body fat composite. Methods: One hundred and forty eight abdominal CT scans performed in our institution between June and August 2013 from patients without history of previous surgery around the umbilicus were analyzed. Multi-planar reconstruction was performed to evaluate the abdominal wall. CT scan cuts at the level of the umbilicus were used to calculate the total fat area (TFA), visceral fat area (VFA), subcutaneous fat area (SFA) by fat measurement soft ware (Toshiba). Results: Opening of the fascial structure at the umbilicus was noted in 25% (37 out of 148) of the patients although none of them demonstrated umbilical hernia clinically. Open group are younger, showed statistically larger BMI, abdominal girth, TFA, and VFA than closed group (p < 0.05). (table) Conclusion: Occult dehiscence of the linea alba is often radiologically appreciated in adult population but the opening per se does not lead to the development of the umbilical hernia. Visceral fat accumulation may play a role in the increase in intra-abdominal pressure and diastasis of the fascia.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

39

in recent years in the diagnosis and treatment of this disease the video show a left hepatectomy with caudate lobe resection did it in a healthy 46 years old woman a left hanging maneuver was used and make transection more easy and faster combined liver and bile duct resection can be performed for hiliar cholangiocarcinoma with acceptable mortality hanging manuver it is useful for right and for left liver resection.

FRIDAY, FEBRUARY 21, 2014, 6:30PM–7:15PM RECEPTION VIDEO PRESENTATIONS VR.01 LAPAROSCOPIC REPAIR OF ERCP RELATED DUODENAL PERFORATION R. M. Seshadri, D. J. Niemeyer, R. Z. Swan, J. J. Lewis, D. Sindram, J. B. Martinie and D. A. Iannitti Carolinas Medical Center, Charlotte, NC ERCP is a common procedure with a 0.3–2.0% risk of duodenal perforation. Free perforation into the abdominal cavity (Type I perforation) has a higher mortality risk and requires surgical intervention. Traditionally this is performed by laparotomy. This video demonstrates a laparoscopic repair of a Type I ERCP-related duodenal perforation. The patient is an 84 year old female with nausea and elevated liver enzymes who was found on MRCP to have a dilated bile duct and on EUS to have sludge in the distal bile without associated mass. On ERCP she was found to have ampullary stenosis and sludge in the bile duct. An endoscopic sphincterotomy was performed. Immediately post-procedure she developed abdominal distention and pain. Free air was confirmed on abdominal X-ray. She was taken to the OR for laparoscopic repair with intra-operative EGD. Bile was aspirated from the upper abdomen and the abdomen was irrigated. A 2 cm posterior perforation was identified on the second portion of the duodenum by endoscopic insufflation of the duodenum under saline submersion. The duodenum was mobilized and the perforation was closed with interrupted PDS sutures. A round ligament flap and fibrin glue was used to reinforce the closure. The patient did well and was discharged home on post-operative day three. This video demonstrates that laparoscopic repair of an ERCP-related type I duodenal perforation is an effective alternative to open laparotomy.

VR.02 KLATSKIN TUMOR: LEFT HEPATECTOMY WITH CAUDATE LOBE RESECTION USING LEFT HANGING MANEUVER R. Sanchez Clariá, N. Resio, O. Mazza, F. Alvarez, M. de Santibañes and S. Bilbao Hospital Italiano Buenos Aires, Capital Federal, BUENOS AIRES carcinoma ohf the hepatic duct confluence is the most common site of bile duct malignancies significant progress has been made HPB 2014, 16 (Suppl. 1), 1–99

VR.03 ROBOTIC ASSISTED CYST GASTROSTOMY AND EXCISION OF A PANCREATIC PSEUDOCYST EXTENDING TO THE PELVIS T. K. Nguyen and A. H. Zureikat University Of Pittsburgh Department Of Surgery, Pittsburgh, PA Introduction: Pancreatic pseudocysts can be a debilitating sequelae of pancreatitis. Many pseudocysts can be managed conservatively with observation or with percutaneous or endoscopic drainage. Rarely, do pseudocysts enlarge despite conservative measures to extend to the pelvis. In this video, we present a minimally invasive surgery approach for management of an extraordinarily large pseudocyst in a patient with chronic pancreatitis secondary to systemic lupus erythematosus. Methods: A cyst gastrostomy, pancreatic debridement, and excision of a pelvic pseudocyst were performed using a robotic assisted and laparoscopic technique. Results/Conclusion: A robotic assisted approach for the surgical management of a symptomatic pseudocyst is still feasible even with extension to the pelvis.

SATURDAY, FEBRUARY 22, 2014, 7:00AM–8:00AM PARALLEL LONG ORAL ABSTRACT D TRICKS OF TRADE LO-D.01 SPLENOPNEUMOPEXY: DECOMPRESSION OF PORTAL HYPERTENSION IN THE SETTING OF PORTAL VENOUS OCCLUSIVE DISEASE J. M. Sutton, S. L. Starnes and S. A. Shah Department Of Surgery, University Of Cincinnati Medical Center, Cincinnati, OH Background: Portal venous hypertension due to complete occlusion of the portal, splenic, and superior mesenteric veins may result in significant life-threatening bleeding from ruptured esophagogastric varices. Decompression options, such as transjugular intrahepatic portosystemic shunt (TIPS), surgical shunts, or the Sigiura procedure, are limited due to lack of venous flow. Splenopneumopexy provides an alternative decompressive procedure resulting in decompression through the splenic pulp into the pulmonary venous system and then the right atrium. Methods: We describe two patients who have recently undergone splenopneumopexy (video presentation). The day prior to surgery, splenic artery embolization is performed by © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

40

Abstracts

interventional radiology. With the patient in the right lateral decubitus position, the thoracic cavity is entered through the left 7th intercostal space. The diaphragm is opened, the upper pole of the spleen is brought through the diaphragm into the chest, and it is fixed in place with a circumferential running 3-0 PDS, adhering the diaphragm to the spleen. The spleen is then amputated to expose a 5 cm bare area, and this is lined up with the diaphragmatic surface of the left lower lobe. The visceral pleura corresponding to this area is incised, stripped, and removed. The denuded lung and spleen are then anastomosed using 3-0 circumferential running PDS. Chest tubes are placed, and the thoracic cavity is closed. Results: With the development of collateral circulation within weeks postoperatively, esophagogastric varices will decompress and resolve. Historically, reports of rebleeding or hepatic dysfunction are rare, and angiographic confirmation of portopulmonary shunting still exists months to years postoperatively.

LO-D.02 EARLY VEIN RECONSTRUCTION AND RIGHT-TO-LEFT PANCREATIC DISSECTION FOR LEFT SIDED PANCREATIC TUMORS WITH PV/SMV OCCLUSION J. M. Cloyd and B. C. Visser Department Of Surgery, Stanford University, Stanford, CA Large left sided pancreatic tumors with complete portal vein and/or superior mesenteric vein (PV/SMV) occlusion (Fig A) present a unique surgical challenge. Division of venous collaterals, as is done in a traditional left-to-right fashion, leads to progressive portal hypertension and increased risk of hemorrhage. Our approach has been to perform early PV resection and reconstruction in order to restore mesenteric-portal flow and decompress downstream varices before proceeding with a right-to-left dissection. An internal jugular (IJ) vein graft is harvested at the beginning of the operation. The pancreas is exposed in the standard fashion. The splenic artery is ligated or divided, as early as feasible, in order to reduce splenic inflow. A Kocher maneuver is performed and the porta hepatis is dissected. The neck of the pancreas is divided in order to expose the PV/SMV at which time proximal and distal control is obtained. The involved segment of PV is resected and the IJ graft is sewn in an end-to-end fashion (Fig B). Immediately, peri-pancreatic varices are decompressed and pancreatic dissection proceeds in a right-to-left fashion. After the pancreatectomy is completed, the vascular anastomosis is inspected once again. Occasionally, the graft may need to be revised if excessive slack is found. If large tumor size or dense collaterals make adequate PV exposure prohibitive, the IJ graft may be used temporarily as a mesocaval shunt, then reconstructed at the conclusion of the case. Postoperatively, the patient is maintained on antiplatelet therapy.

LO-D.03 LAPAROSCOPIC EXPOSURE AND LIGATION OF THE RIGHT BRANCH OF THE PORTAL VEIN S. Jayaraman University Of Toronto, Toronto, ON Laparoscopic right hepatectomy remains a challenging technical exercise. To facilitate safe resection with reduced intraoperative blood loss, control of the portal venous and hepatic arterial inflow may be beneficial. Due to the anatomic location of the right branch of the portal vein it can be very difficult to expose and control with the restricted dexterity of conventional laparoscopy. Once exposed, it can be challenging to introduce an endoscopic stapler around the right portal vein without excessive torque and tension. The right portal vein can be reproducibly and reliably exposed and ligated laparoscopically using 5 steps: dissection of the triangle of Calot; keeping the cystic duct attached to the gallbladder and using it as a retractor to expose the right hepatic inflow vessels; isolation and ligation of the right hepatic artery; dissection of the right portal vein; and division of the parenchyma of the caudate process posterior to the right portal vein. The final move allows easy and direct insertion of the laparoscopic stapler around the right portal vein through an epigastric 12 mm port unimpeded. The right portal vein can then be stapled safely and smoothly with no torque or tension. This “Tips and Tricks” video demonstrates this reproducible and reliable surgical approach to isolate, ligate, and divide the right branch of the portal vein during a laparoscopic right hepatectomy.

LO-D.04 LAPAROSCOPIC LIVER MOBILIZATION TRICKS OF THE TRADE TO AVOID COMPLICATIONS N. Ikoma1,2, G. Oshima2 and O. Itano2 University Of Texas Health Science Center At Houston, Houston, TX; 2Keio University School Of Medicine, Tokyo, TOKYO 1

Laparoscopic liver resection is gaining popularity because of the availability of new laparoscopic instruments and advanced techniques. This technique involves two steps: mobilization and parenchyma transection. Laparoscopic liver mobilization is necessary not only for pure laparoscopic liver resection but also for laparoscopy-assisted liver resection © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

41

with small laparotomies for liver transection and specimen extraction, which is a good learning transition step before further advanced laparoscopic liver resection. Moreover, sufficient liver mobilization expands indication for laparoscopic liver resection by providing an adequate margin between tumors and large vessels. Although experienced surgeons prefer laparoscopic liver mobilization because of the clear magnified vision, the procedures are still technically

demanding. It is very important to understand the pitfalls of this procedure to prevent injury to large vessels such as the IVC or hepatic veins. In this presentation, the details of the operative technique of laparoscopic liver mobilization are reported along with videos. This should aid the wide acceptance of laparoscopic liver resection as a standard operation. Here, the technique of liver mobilization is summarized, along with the short-term outcomes at the author’s facility.

LO-D.05 TRANSGASTRIC PANCREATIC NECROSECTOMY – TRICKS OF THE TRADE

SATURDAY, FEBRUARY 22, 2014, 7:00AM–8:00AM PARALLEL LONG ORAL ABSTRACT E VIDEO

N. J. Zyromski Indiana University Department Of Surgery, Indianapolis, IN Contemporary management of necrotizing pancreatitis is evolving rapidly, and may include percutaneous, endoscopic, or surgical approaches. Importantly, individual patient characteristics should dictate specific interventional approach; decisions regarding intervention are ideally planned in the context of a multidisciplinary group. Select patients with necrotic collections confined to the lesser sac may be approached in a transgastric fashion. Surgical approach (either laparoscopic or open) offers the advantage of “one stop shopping” – allowing complete debridement as well as cholecystectomy (for patients with biliary pancreatitis) in one setting. The goals of this “Tricks of the Trade” presentation are: to review optimal patient selection; to highlight which patients are suitable for laparoscopic versus open approach; to review our institutional experience with transgastric debridement; to highlight specific operative “pearls” (avoiding and managing hemorrhage, utility and techniques of intraoperative ultrasound, determining gastrotomy siting, port placement); and to review long term outcomes and complications following transgastric pancreatic debridement.

HPB 2014, 16 (Suppl. 1), 1–99

LO-E.01 SEGMENT VIII RESECTION FOR COLON CANCER METASTASIS USING THE IN SITU COLD PERFUSION TECHNIQUE WITH RIGHT AND MIDDLE HEPATIC VEIN RECONSTRUCTION M. Segedi, I. McGilvray, P. Greig, P. Kelly, A. Fung and G. Tait University Of Toronto, Toronto, ONTARIO Colorectal metastases that occur in anatomic location precluding classic resection can be removed with a novel technique using in situ cold perfusion. A video presented here incorporates 3-D animations based on the patient’s radiology, with overlays and animated technical summaries to enhance trainee’s understanding of this complex resection. Here we present a segment 8 resection in a 41 year old female found to have recurrent solitary metastatic colon cancer. Previous wedge resections left her with a relatively small left lobe liver. Tumor was found lodged between right and middle hepatic veins, located within the segment VIII. The video begins with an animated sequence summarizing preoperative planning. Oncologic and anatomic considerations are presented with three-dimensional renderings of patient anatomy. The steps include mobilization of liver and exposure of hepatic veins proximally to the tumor and at their insertion into the inferior vena cava (IVC). The caudate lobe is resected to facilitate further mobilization and vascular clamping. After establishing inflow and outflow control with appropriate clamps, the cold perfusion is set up and the right © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

42

Abstracts

and middle veins are divided at their origins within the liver. After segment VIII resection, the right and middle hepatic veins are reconstructed to their origins on IVC. The patient tolerated the procedure well and was discharged home seven days after the surgery. This video of a complex case illustrates the blend of transplantation and surgical oncology techniques, and it also highlights how animation combined with real-time operative video can convey subtle and complex teaching points.

LO-E.02 TOTALLY LAPAROSCOPIC CAUDATE LOBECTOMY WITH PARTIAL CAVAL RESECTION C. U. Corvera, C. A. Galanopolus and V. Kuo University Of California San Francisco, San Francisco, CA The patient is a 68 year-old man who was diagnosed with metastatic colon cancer to the caudate lobe of the liver. At operation, the left lateral segment was mobilized. The lesser sac was opened and ligament venosum was divided, thereby exposing the left hepatic vein. With the caudate lobe now exposed, the tumor was noted to have some direct extension to the lesser omentum and retroperitoneal fat. These attachments were taken en-bloc with the caudate lobe. The principle branch of the caudate was identified, doubly clipped and divided. Mobilization of the caudate was then pursed by retracting the hepatoduodenal ligament toward the patient’s right thus exposing the anterior surface of the inferior vena cava. Short hepatic draining veins were managed by metal clips and divided. The caudate process was then divided using a bipolar cautery device. The tumor was noted to be densely attached to the left lateral edge of the IVC and an en-bloc resection was required. The caudate process was completely divided up to the base and behind the conjoint middle and left hepatic vein. The caudate lobe was dissected free from the anterior surface of the IVC and reflected laterally leaving only ∼2 cm segment of tumor involvement with the IVC. This presumed direct tumor invasion was encircled using a silk suture and used for traction. A tangential partial caval resection was done using two loads of an EndoGIA vascular stapler. The specimen was retrieved via the midline periumbilical port.

LO-E.03 ANTE SITUM HEPATECTOMY AND REPLACEMENT OF INFERIOR VENA CAVA AND HEPATIC VEIN USING A TRANSITORY PORTO-CAVAL SHUNT A. Yanzon, A. Cristiano, L. Yazde, F. Rodriguez Santos, M. de Santibañes, J. Pekolj and E. de Santibañes Hospital Italiano de Buenos Aires, Ciudad Autonoma De Buenos Aires, BUENOS AIRES Introduction: Hepatic resection is considered the only curative treatment for both primary and metastatic disease of the liver. However, lesions located at the hepatocaval confluence compromising the inferior vena cava (IVC) are not resectable under conventional techniques. To overcome this problem, different surgical procedures have been developed for the last 25 years. Ex-vivo resection techniques provide excellent © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

accessibility to tumors placed around the IVC. The main problem regarding ex-vivo procedures is the hepatic low tolerance to warm ischemia and the splanchnic congestion secondary to vascular exclusion. Hypothermic hepatic perfusion and veno-venous bypasses are 2 strategies to avoid these complications. In this video, we report a novel technique to minimize vascular complications for ante-situ resection using a goretex graft for IVC replacement and a temporary shunt between the porta vein and the IVC using a cadaveric bank preserved venous graft. Method: We present the case of a 27 years old patient with an hepatic recurrence from a rectal cancer compromising hepatocaval confluence. To avoid complications of the total vascular exclusion and the extracorporeal veno-venous bypass, we performed a porto-sistemyc shunt between the portal vein and the vena cava. Results: The surgery accomplished an R0 resection, with low Morbidity. Total hospital stay was eighteen days. The patient, 7 months after surgery, has no evidence of recurrence. Conclusion: Ex-vivo resections techniques are a valid approach to treat these tumors. To our knowledge this report is the first description of the porto-caval transitory shunt to avoid the total vascular exclusion and venovenous bypass drawbacks.

LO-E.04 ROBOTIC PANCREATICODUODENECTOMY WITH ANOMALOUS HEPATIC ARTERIAL ANATOMY A. H. Zureikat, T. K. Nguyen and H. J. Zeh University Of Pittsburgh Department Of Surgery, Pittsburgh, PA Pancreaticoduodenectomies performed with robotic assistance is emerging as a safe alternative in experienced centers to the traditional open technique. This video demonstrates a robotic pancreaticoduodenectomy in a patient with a pancreatic head adenocarcinoma with an anomalous hepatic artery location. The robotic platform facilitates dissection of vascular structures using a no-touch technique, allowing meticolous hemostasis. Reconstruction of this 2 mm pancreatic duct is performed in duct to mucosa fashion under magnification. This patient was discharged on postoperative day number 8 with no complications.

LO-E.05 INTRAOPERATIVE USE OF GOOGLE GLASS FOR HEPATO-PANCREATO-BILIARY SURGERY S. L. Bokshan1, W. Chapman1,2 and M. Doyle1,2 Washington University School Of Medicine, St. Louis, MO; 2BJC Health System, St. Louis, MO

1

Google Glass is a modular technology that projects a small heads up display (Glass) into the peripheral visual field of the user. This heads up display receives user input from voice commands enhanced by bone conduction. In addition to receiving input commands from the user, Google Glass can also record 720p videos from its anterior camera. It boasts HPB 2014, 16 (Suppl. 1), 1–99

Abstracts Bluetooth and wireless connectivity allowing the user to download images to the device, connect to the user’s cell phone, video chat, and to upload images and videos. The heads up display is capable of projecting data into the user’s periphery. While all these technologies could potentially aid surgeons with intraoperative video recording, surgeons have yet to report the feasibility of using Google Glass intraoperatively. From Glass, the OR video is wirelessly streamed to an in-house video and multimedia management system known as Medicrate that we have developed. The video can be edited by whom ever is granted access and shared in order to maximize surgical teaching. The system also allows users to store videos, documents, presentations, and images to a local hospital cloud. These various forms of multimedia can be selectively shared by creating local groups on Medicrate and giving group access of personal multimedia. Finally, Medicrate allows users to generate custom quizzes associated with their content in order to assess student understanding. This video presentation will provide an in depth overview of the intraoperative feasibility and capabilities of Google Glass as well as the utility of the Medicrate system.

43

with current personal study strategies. 62% reported they primarily studied with focused intent (preparing for cases with immediate clinical relevance) versus following a prescribed curriculum(5%). 86% of fellows reported accessing information via computer and 43% via phone or tablet, while 67% of respondents also utilized textbooks. Figure 1 delineates online learning modalities respondents would utilize if available. Fellows indicated they would participate in an interactive discussion board once(47%) or 2–3 times(24%) weekly. Most respondents (86%) would meaningfully follow 1–2 discussions per month. Conclusion: HPB fellows utilize online learning tools. Improved implementation of the HPB fellows’ curricula is needed. This study informs learning modalities likely to be utilized by current HPB fellows, and guides design and development of learning tools and interactive components most likely to be adopted by fellows.

SATURDAY, FEBRUARY 22, 2014, 7:00AM–8:00AM PARALLEL LONG ORAL ABSTRACT F OTHER LO-F.01 DEVELOPMENT OF AN INTERNATIONAL ONLINE LEARNING PLATFORM FOR HPB TRAINING A NEEDS ASSESSMENT S. G. Warner1, S. J. Connor2, C. Christophi3, I. Azodo4, T. S. Kent5, D. Pier4 and R. M. Minter1 1 University Of Michigan, Division of HPB And Advanced GI Surgery, Ann Arbor, MI; 2Department Of General Surgery, Christchurch Hospital, Christchurch, .; 3 Department Of Surgery, University Of Melbourne, Austin Hospital, Heidelberg, VICTORIA; 4Royal Infirmary Of Edinburgh, Edinburgh, SCOTLAND; 5Beth Isreal Deaconess Medical Center, Boston, MA Introduction: Implementation of the well-developed AHPBA/ANZHPBA fellowship curricula remains limited. The current generation of trainees embrace online learning. The AHPBA and ANZHPBA are therefore developing an online interface to facilitate an interactive didactic experience for HPB fellows to augment the current curriculum. A needs-assessment was performed to identify content modalities to optimize fellow education. Methods: A 22-question survey querying fellows’ learning styles and use of various learning tools was disseminated electronically to 38 North American/Australasian HPB fellows after validation and revision in an initial 5-resident cohort for clarity and response time. Results: The survey response rate was 55%(n = 21). Median respondent age was 35 y(range 30–42 y) and 80% were male. Only 67% endorsed familiarity with their required HPB curriculum, and 43% indicated dissatisfaction HPB 2014, 16 (Suppl. 1), 1–99

LO-F.02 THE CURRENT STATE OF HPB FELLOWSHIP EXPERIENCE IN NORTH AMERICA H. Osman and D. Jeyarajah Methodist Dallas Medical Center, Dallas, TX Aim: The face of HPB training has changed over the past decade. The growth of focused HPB fellowships that are vetted with a rigorous accreditation process through the Fellowship Council (FC) has established these fellowships as an attractive mode of training in HPB surgery. This study looks at the volume of HPB cases performed during these fellowships in North America. Methods: After approval by the FC research committee, data from all HPB fellowships that had three years worth of complete fellow case log data were tabulated and reported (n = 12). Two year fellowships had their fellow logs tabulated at the completion of both years. Those programs that had transplant experience (n = 9) were reported. Results: Data for fellows’ case numbers is shown below. Table 1(liv-maj: hemi liver resection. liv-min: 2 segments or less.PD: Pancreaticoduodenectomy. DP: distal pancreatectomy). Conclusion: This study validates that FC accredited HPB fellowships have a robust exposure to complex HPB surgery. Fellows completing these fellowships should be well versed in the management and surgical treatment of HPB patients. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

44

Abstracts

LO-F.05 SURGICAL AND TOTAL COSTS OF HEPATO-PANCREATO-BILIARY (HPB) SURGERY FOR US PATIENTS ENROLLED IN PRIVATE INSURANCE J. R. Schubart, E. Schaefer, A. N. Kulaylat, C. S. Hollenbeak and N. J. Gusani Program For Liver, Pancreas, & Foregut Tumors, Penn State College Of Medicine, Hershey, PA

LO-F.03 SURGICAL MANAGEMENT OF METASTATIC COLON CANCER IN ELDERLY: ANALYSIS OF THE SURVEILLANCE, EPIDEMIOLOGY AND END RESULTS (SEER) DATABASE H. Khan1, A. Olszewski2 and P. Somasundar1 Roger Williams Medical Center, Providence, RI; 2 Pawtucket Memorial Hospital, Pawtucket, RHODE ISLAND

1

Background: The objective of this study was to evaluate survival after resection of primary tumor and/or metastatic sites, by age and the extent of nodal spread, using the population-based SEER registry. Methods: We extracted data on pts diagnosed with colon adenocarcinoma between 2000–2010 (N = 232,640). We analyzed extent of nodal involvement (N0-2) in cases with ≥12 lymph nodes examined (LNE), identified stage IV pts with recorded resection of primary tumor or of metastases. Association of overall survival (OS) with metastatectomy was evaluated in multivariable Cox models. Results: Among 41,590 pts with stage IV cancer, 67% underwent primary surgery and 11% metastatectomy. Older pts were more likely to have <12 LNE, but also to have N0 status when adequately staged (Table, P < .0001). Metastatectomy (N = 4,506) was less common in older patients (Table). OS after resection inversely associated with age, nodal involvement, tumor grade and suboptimal staging with <12 LNE (P < .0001). However, OS benefit of metastatectomy was significant in all age groups except those ≥85 y old (Table) and was independent of the extent of nodal spread (HR, 0.71, 0.68 and 0.77 for N0, N1 and N2 cases, respectively) or tumor grade. Conclusion: The survival advantage of metastatic site resection in colon cancer is present in ages up to 85 y. Adequate nodal resection favorably correlates with survival after metastatectomy, and older pts are more likely to be nodenegative at resection with ≥12 LNE. These results support offering curative metastatectomy and resection of the primary tumor in this age group.

Introduction: Costs of surgical care in private insurance systems are not well understood. In high-complexity surgical procedures, surgical and perioperative costs, as well as readmission and further therapies, are likely to drastically affect total costs incurred. Methods: The MarketScan® database, 2008–10, containing claims for >56 million employees/dependents under private insurance, was used to identify major pancreatectomy (n = 2254) and hepatobiliary (n = 1702) surgery patients who were enrolled >=120 days prior to surgery and after discharge. “Total costs” represent all payments paid to the provider for services reported within this timeframe, inflated to 2010 dollars. We also measured all payments for the initial surgery and subsequent hospital stay only (“surgical costs”). Results: Median age was 53, with perioperative mortality of 2.1% (pancreas), and 1.5% (liver), respectively. Median postoperative length-of-stay was 9 d for pancreatectomy and 6 d for hepatectomy. Fewer pancreatectomy patients were discharged to home compared to hepatectomy patients (70.9% vs 85.1%). 120 d readmission rates were higher for pancreatic surgery (31.1%) than for hepatic surgery (22.7%). Mean total costs were $105,200 (95% CI: $102,200-$108,200) for pancreatectomy and $88,600 (95% CI: $85,800-$91,500) for hepatectomy. Mean surgical costs were $70,700 for pancreatectomy, representing 67% of total costs, while mean surgical costs were $55,600 for hepatectomy, 63% of total costs. Conclusion: This study reports costs for surgery patients in private insurance. These patients are younger and likely to have fewer co-morbidities than those covered by Medicare. Total cost of surgical care was influenced largely by the index surgical admission, representing 63–67% of the total costs of care.

SATURDAY, FEBRUARY 22, 2014, 10:45AM–11:15AM BREAK VIDEO VB.01 ROBOTIC-ASSISTED PANCREATODUODENECTOMY FOR CANCER E. Vicente, Y. Quijano, B. Ielpo, H. Duran, E. Diaz, I. Fabra, C. Oliva, S. Olivares, R. Caruso, R. Ceron and J. Plaza Sanchinarro University Hospital, Madrid, MADRID Background: Pancreatic surgery is a challenge for most surgeons. This complexity is clearly increased with the use of the minimally invasive techniques. However, Robotic system may overcome some of the limitations of the laparoscopic procedure. Materials and Methods: In this video we show a case of Robotic assisted duodeno-cephalic pancreatectomy.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts Results: We show the different Robotic surgical steps of the pancreatoduodenectomy which includes the tumor resection and robotic surgical reconstructions (biliary, pancreatic and intestinal anastomosis). Conclusion: This video shows the efficacy and feasibility of robotic pancreatic surgery providing a three dimensional imaging, a greater instrumental movement and tremor filtering of the surgeon, thus reducing the minimally invasive surgery related complexity.

VB.02 SYNCHRONOUS TOTAL LAPAROSCOPIC RIGHT HEPATECTOMY AND RIGHT HEMICOLECTOMY S. Jayaraman University Of Toronto, Toronto, ON Synchronous resection of primary colon cancer and liver metastases has been shown to be safe and effective in selected cases. This video details the presentation of an asymptomatic patient with a colon cancer originating in the cecum as well as an isolated hepatic metastasis near the bifurcation of the right anterior and posterior inflow pedicles. In order to obtain total oncologic clearance, a right hepatectomy and right hemicolectomy was indicated. The operation begins with the hepatectomy being performed laparoscopically with out the aid of open or hand ports. Upon completion of the hepatectomy, the right liver specimen is placed in a large bag in the right upper quadrant. The laparoscopic right hemicolectomy ensues including a medial approach and high ligation of the ileocolic pedicle. Subsequently an intracorporeal laparoscopic ileocolic anastomosis is performed. Both the liver and colon specimen are removed through a Pfannensteil incision. The final pathology showed a pT3N1M1 colon cancer with all margins negative. The patient was discharged on the ninth postoperative day with an uneventful hospitalization. In one year follow up and after completion of adjuvant therapy, there is no evidence of recurrence.

VB.03 NEAR INFRARRED TECHNOLOGY GIVES A NEW TOOL TO THE SURGEONS TO IDENTIFY BILIARY ANATOMY F. Dip, M. Roy, C. Simpfendorfer, E. Lomenzo, S. Szomstein and R. Rosenthal Cleveland Clinic Florida, Weston, FLORIDA Introduction: Fluorescence-guided surgery is a novel technology that could facilitate the recognition of bile duct structures. Using adequate fluorescent dye and appropriate laparoscopic equipment, critical structures can be identified. Method: The procedure begins with administration of a single dose of 0.05 mg/kg of indocyanine green one hour prior to surgery. A Storz® xenon light source and laparoscope with a charge-coupled device that filters out light wavelengths except 830 nm with a specific 780 nm infrared light source is used intraoperatively. The light of the laparoscope can be easily changed to the infrared view using a pedal. When illuminated by infrared light, the dye manifests fluorescence. Abdominal inspection with standard trocar placeHPB 2014, 16 (Suppl. 1), 1–99

45

ment is performed. The fluorescence system is then activated using the pedal. Since the dye is excreted exclusively by the liver it creates the biliary structures can be identified because of fluorescence. Results: This video demonstrates the feasibility of the technique using indocyanin green as described above. Visualization of the cystic duct, hepatic ducts, accessory ducts and main bile duct is shown using fluorescence. Conclusion: Near infrared technology seems to be a feasible and easy method to delineate the biliary tree anatomy in real time.

SATURDAY, FEBRUARY 22, 2014, 2:30PM–4:30PM PARALLEL ORAL ABSTRACT G LIVER/OTHER O-G.01 EARLY IDENTIFICATION OF PATIENTS AT INCREASED RISK FOR HEPATIC INSUFFICIENCY, COMPLICATIONS, AND MORTALITY AFTER MAJOR HEPATECTOMY J. W. Etra1, M. H. Squires1, S. B. Fisher1, D. R. Rutz1, B. Martin1, D. A. Kooby1, K. Cardona1, J. M. Sarmiento2, C. A. Staley1, S. K. Maithel1 and M. C. Russell1 1 Department Of Surgery, Division Of Surgical Oncology, Winship Cancer Institute, Emory University, Atlanta, GEORGIA; 2Department Of Surgery, Division Of General And Gastrointestinal Surgery, Emory University, Atlanta, GEORGIA Background: A postoperative bilirubin >7 mg/dL is an accepted definition of postoperative hepatic insufficiency (PHI), as it is associated with increased complications and mortality after hepatectomy. Early recognition of these patients at increased risk may affect treatment strategy. Our aim was to identify a surrogate marker for PHI early in the postoperative course. Methods: A single institution database of patients undergoing major hepatectomy (≥3 segments) from 2000–2012 was retrospectively reviewed. Preoperative, intraoperative, and early postoperative factors were assessed for association with any complications, major complications (Clavien Grade IIIV), and 90-day mortality. Results: 607 patients undergoing major hepatic resection, excluding bile duct reconstruction, were identified. Postoperative complications and major complications occurred in 352(58%) and 136(22%) patients. 90-day mortality was 4%. A postoperative day 3 (POD3) total bilirubin ≥3 mg/dL was significantly (p < 0.001) associated with increased risk of post-operative complications (76% vs 54%), major complications (46% vs 18%), and 90-day mortality (16% vs 2%; Figure I). This association persisted on multivariate regression analysis for all 3 endpoints: any complications (HR:1.98, 95%CI:1.10–3.54,p = 0.022), major complications (HR:3.18, 95%CI:1.90–5.32,p < 0.001), and 90-day mortality (HR:8.11, 95%CI:3.00–21.92,p < 0.001; Table I). A POD3 bilirubin ≥3 was the only early perioperative factor associated with the development of a postoperative total bili© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

46

Abstracts

rubin >7 and/or ascites on multivariate analysis (HR:7.81, 95%CI:3.74–16.31,p < 0.001). Conclusion: Total bilirubin ≥3 mg/dL on POD3 after major hepatectomy is significantly associated with increased complications, major complications, and 90-day mortality. This marker may serve as a predictor in the early postoperative course for hepatic insufficiency, given its strong association with adverse outcomes and increased mortality.

higher in the obese group. EBL and OR time were greater for the overweight and obese groups, however only significant for the overweight group. Length of stay and post-operative complications were also statistically similar between groups. The multivariate model of factors significantly correlated with complications included weight class, ASA score, blood loss, and number of segments resected. In the multivariate model, morbid obesity (BMI >35) was significantly associated with complications (OR 3.125; p = 0.033), while BMI <35 was not significant (OR 0.973; p = 0.945). Conclusion: Similar to many open operations, weight may be associated with a slightly longer operating time and greater blood loss in laparoscopic hepatectomy. Overall complications as well as length of stay are similar for overweight and obese patients compared to normal BMI patients.

O-G.03 MORTALITY IN HEPATECTOMY: MODEL FOR END-STAGE LIVER DISEASE (MELD) AS A PREDICTOR FOR DEATH USING THE NATIONAL SURGICAL QUALITY IMPROVEMENT PROGRAM (NSQIP)

O-G.02 EFFECT OF OBESITY ON PERIOPERATIVE OUTCOMES FOLLOWING LAPAROSCOPIC HEPATECTOMY S. Lee, A. Hauch, E. Kane, E. Kandil and J. Buell Tulane Transplant Institute, Tulane University School Of Medicine, New Orleans, LA Background: Laparoscopic hepatectomy has become increasingly more accepted in the treatment of liver tumors in selected patients. Whether obese patients would benefit from laparoscopic hepatectomy has not been investigated. This study evaluates the effect of obesity on perioperative outcomes following laparoscopic hepatectomy. Methods: Retrospective review of 382 consecutive patients undergoing laparoscopic hepatectomy. Patients were divided into three weight classes: normal, overweight (BMI ≥25), and obese (BMI ≥30). Comparisons and analysis were made against the normal group on preoperative factors, operative variables, and perioperative outcomes. Results: There were 382 total patients, with 76 (20%) normal, 204 (53.4%) overweight, and 102 (23%) obese. The groups were equivalent with respect to age, presence of cirrhosis, segments resected, while the mean ASA score was © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

S. W. Ross, B. Oommen, M. Kim, V. A. Augenstein, A. L. Walter, B. T. Heniford, D. A. Iannitti, J. B. Martinie and R. Z. Swan Carolinas Medical Center, Department Of General Surgery, Charlotte, NC Introduction: Current literature is mixed on the predictive value of MELD score for mortality in hepatectomy. We hypothesized that MELD will independently correlate with 30-day mortality in patients undergoing hepatic resection. Methods: The American College of Surgeons NSQIP database was queried for hepatectomy. Original MELD, UNOS modification MELD(uMELD), integrated MELD(iMELD), and sodium corrected MELD(MELD-Na) were also calculated. Demographics, comorbidities, operative details, and patient outcomes were compared using standard statistical tests and examined by MELD type. Mortality was analyzed using logistic regression and the MELD types were compared using receiver operating curves (ROC). Results: There were 11,933 hepatectomies from 2005– 2011: 7,519 partial; 2,104 right; 1,210 left; and 1,100 trisegmentectomy. Demographics and mean MELD scores: age 58.7 ± 13.4, BMI 28.0 ± 6.3 kg/m2, MELD 3.4 ± 4.8, uMELD 7.9 ± 2.5, MELD-Na 8.2 + 2.9, and iMELD 31.2 + 5.0. 51.8% were male and 82.5% were Caucasian. Mean length of stay was 8.4 ± 22.0 days, and there were 275 deaths(2.4%). 30-day mortality rates by uMELD strata were: HPB 2014, 16 (Suppl. 1), 1–99

Abstracts 0–9(1.8%), 10–19(6.9%), 20–29(15.4%), 30 + (25.0%); and on bivariate analysis all MELD types were associated with higher mortality (p < 0.001). All meld types were independently associated with mortality (p < 0.001) and uMELD had the largest effect size (OR: 1.20, 1.17–1.24). iMELD had the narrowest confidence interval (OR: 1.18, 1.15–1.21) and largest area under the ROC. Conclusion: Our study shows that hepatectomy patients with higher MELD scores have increasing rates of 30-day mortality. For every point increase in uMELD there is a 20% increase in mortality risk. Additionally, iMELD has the best predictive characteristics for mortality when compared to other MELD scores.

O-G.04 ENUCLEATION OF HEPATIC NEUROENDOCRINE TUMOR METASTASES: A VALUABLE ADJUNCT TO SURGICAL CYTOREDUCTION N. N. Nissen1, V. G. Menon1, E. M. Wolin2, R. Yu2, S. D. Colquhoun1 1 Hepatobiliary And Pancreatic Surgery, Cedars-Sinai Medical Center, Los Angeles, CA; 2Carcinoid And Neuroendocrine Tumor Program, Cedars-Sinai Medical Center, Los Angeles, CA Objective: Hepatic neuroendocrine tumor (NET) metastases are often treated with cytoreductive surgery. The unique nature of these tumors may allow parenchymal preservation using the technique of enucleation, but this approach has not been systematically evaluated. Design: We reviewed our surgical experience in 100 consecutive patients with metastatic hepatic NETs between 2003–2013. Thirty-eight patients (38%) underwent planned enucleation (ENUC) as a component of surgery and these were compared to standard resection. Results: ENUC was performed more commonly in recent years, and was combined with anatomic hepatic resection in 20 patients (53%), tumor ablation in 14 (37%) and resection of primary abdominal tumor in 25 patients (66%). Mean number of lesions enucleated was 5 (range 1–26), and mean size of the largest enucleated lesion was 5.6 cm (range 1–19 cm). Laparoscopic ENUC was completed in 4 patients (11%). There was no mortality. Complications of Clavien grade 3 or more occurred in 22%, including 14% rate of bile leak, similar to patients undergoing standard resection. The ability to achieve maximal cytoreduction (>90% of tumor removed) was greater in the ENUC versus standard group (79 vs. 47%, p = .0028). Also, fewer patients in the ENUC group required additional hepatic therapy within 12 months compared to standard group (16 vs. 34%, p = .03). Both ENUC and standard groups showed excellent survival (92/ 88% and 88/72%) at 1/3 years respectively. Conclusion: Enucleation is a safe and effective tool in surgical management of hepatic NETs. Patients treated with a surgical strategy using this approach can often have nearcomplete cytoreduction and are unlikely to need additional hepatic therapies over the short term.

HPB 2014, 16 (Suppl. 1), 1–99

47

O-G.05 EXPRESSION LEVELS OF INSULIN-LIKE GROWTH FACTORS AND RECEPTORS IN HEPATOCELLULAR CARCINOMA Y. S. Chun1, M. Huang2 and L. Rink2 Virginia Piper Cancer Institute, Minneapolis, MN; 2Fox Chase Cancer Center, Philadelphia, PA

1

Introduction: The insulin-like growth factor (IGF) pathway is implicated in the pathogenesis of hepatocellular carcinoma (HCC) and may be particularly important in patients with non-alcoholic fatty liver disease (NAFLD) and metabolic syndrome, which is associated with insulin resistance and increased levels of bioavailable IGF-1. The aim of this study was to determine expression levels of IGFs and their receptors in patient with NAFLD-associated HCC. Methods: Tissue microarrays were constructed from patients who underwent hepatectomy for HCC. Immunohistochemistry was performed using antibodies for IGF-1, IGF-2 and their receptors, IGF-1R and IGF-2R. Slides were read by a pathologist blinded to clinical data. Median follow-up was 26 months (range, 2–136 months). Results: Among 27 patients with HCC, the most common underlying liver diseases included NAFLD (n = 13, 48%), hepatitis C (n = 7, 26%), and alcoholic hepatitis (n = 2, 7%). Expression levels of IGFs and their receptors were not associated with patients’ underlying liver disease. In all patients, IGF-2 expression was upregulated in tumor and adjacent non-tumorous liver. IGF-1 expression was low in adjacent non-tumorous liver in 6 of 10 patients (60%) with cirrhosis, compared to 2 of 17 patients (12%) without cirrhosis (p = 0.008). Higher IGF-1 expression in adjacent liver relative to tumor was associated with significantly poorer median survival of 22 months, compared to 72 months with equal or lower IGF-1 expression in adjacent liver relative to tumor (p = 0.006). Conclusion: Our preliminary results demonstrate significant associations between IGF-1 expression and liver cirrhosis and survival after resection in patients with HCC, independent of their underlying liver disease.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

48

Abstracts

O-G.06 POSITIVE CROSSMATCH AND THE ROLE OF DONOR SPECIFIC ANTIBODY IN LIVER TRANSPLANTATION R. J. Lynch, N. Lysak, S. Perez, T. Roberts-Wilson, R. A. Bray, H. M. Gebel and S. J. Knechtle Emory Transplant Center, Atlanta, GA Introduction: Although the importance of crossmatch and donor specific antibody (DSA) is clear in renal transplantation, their utility is unproven in liver transplantation. Recent studies have suggested that alloantibody could be implicated in acute reperfusion injury, and may initiate or potentiate graft rejection. Methods: We performed a retrospective analysis of liver transplants at our center from 2011–2013. Donor and recipient demographics, etiology of liver disease and graft ischemic time were recorded. Histocompatibility data were taken from a prospectively maintained database of crossmatch and alloantibody titers. Patient outcome was assessed through postoperative transaminase levels and incidence of readmission, biliary complications, and hepatic artery thrombosis. Histologically-confirmed rejection episodes were cataloged by type and severity. Results: 174 patients were followed for 318 ± 157 days. Immunosuppression consisted of calcineurin inhibitor, mycophenolate mofitil ± steroids. 48 patients had a positive crossmatch, of which 30 were found to be due to human leukocyte antigens (HLA). Of these, 16 had high (>10,000 MESF) anti MHC Class I antibody titers, and 9 had high MHC Class II titers. There were no significant demographic differences between groups. 24 hour transaminase measurements, as well as clinical complication rates, were similar. Rejection-free survival was slightly better in the positive crossmatch group, though this did not achieve statistical significance (32% vs 19%, p 0.08, Figure 1). Conclusion: Among liver transplant recipients, the presence of a positive crossmatch is not associated with adverse clinical or immunological outcomes. High titers of anti-donor antibody did not predict tissue injury or rejection over medium-term followup.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

O-G.08 INTRAOPERATIVE ULTRASOUND DURING RESECTION OF COLORECTAL LIVER METASTASES: IMPACT ON DETECTION OF LESIONS, CHANGE IN SURGICAL STRATEGY AND ACHIEVEMENT OF NEGATIVE RESECTION MARGINS S. A. Knowles1, K. P. Croome2 and R. Hernandez-Alejandro1 1 Division Of General Surgery, Western University , London, ONTARIO; 2Division Of Gastroenterologic And General Surgery, Mayo Clinic, Rochester, MINNESOTA Liver resections with negative margins improve survival in patients with colorectal liver metastases (CRLM). Intraoperative ultrasound (IOUS) is a valuable tool that can potentially give information about lesions that ultimately changes surgical strategy to ensure complete resection, which subsequently improves disease free survival (DFS). A retrospective chart review of patients who underwent a liver resection for CRLM from 2009 to 2012 was performed in order to determine the impact of IOUS. A total of 103 patients had a hepatic resection for CRLM. All of the patients had a preoperative triphasic CT and/or MRI to assist with operative planning. IOUS was performed in 72 cases. Surgical strategy changed in 31 (43.1%) cases with IOUS, compared to three (9.7%) with no IOUS (p < 0.001) (Figure 1). The most common reason for these changes was the location of the lesion as visualized by IOUS (N = 18, 25.0%), followed by detecting a new lesion not seen on preoperative imaging (N = 12, 16.7%). A higher proportion of nonanatomic liver resections were performed in the IOUS group (N = 45, 62.5%) compared to the non-IOUS group (N = 11, 35.5%). Achieving a negative resection margin (R0 resection) was comparable between the two groups. However, there was a trend toward improved DFS in the IOUS group. Despite advances in preoperative imaging, IOUS still demonstrates utility in detecting new lesions and giving novel information about these lesions that ultimately changes the operative plan, potentially improving DFS. It also assists the surgeon in performing parenchymal-preserving liver resections.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

49

O-G.09 NON-INVASIVE PREDICTION OF SINUSOIDAL INJURY IN PATIENTS TREATED WITH OXALIPLATIN FOR COLORECTAL LIVER METASTASES J. Urdzik1, T. Bjerner2, A. Wanders3, F. F. Duraj1, U. Haglund1 and A. Norén 1 Department Of Surgery, Uppsala University, Uppsala, SWE; 2Department Of Radiology, Uppsala University, Uppsala, SWE; 3Department Of Immunology, Genetics And Pathology, Uppsala University, Uppsala, SWE Introduction: Noninvasive methods to identify patients with sinusoidal injury (SI) after oxaliplatin based therapies for colorectal liver metastases (CRLM) have been proposed. Predictive values of APRI score (aspartate aminotransferase to platelets ratio index), spleen volume changes and portal vein hemodynamic changes were compared. Methods: APRI and magnetic resonance imaging flowmetry measuring portal vein hemodynamics were estimated the day before liver surgery in 29 patients. Spleen volume changes were calculated from segmented spleen volumes on CT or MR imaging before chemotherapy start and the day before surgery. Resected non-tumorous liver parenchyma was evaluated according to Complex Vascular Injury score (≥3 was recognized as relevant SI). Results: SI was observed in 6 of 29 patients. 5 of these 6 patients received oxaliplatin; however, the relation SI to oxaliplatin was not significant. Patients with SI demonstrated higher APRI score (p = 0.008), larger portal vein crosssection area (AREA) (p = 0.041) and lower mean portal flow velocity (VELO) (p = 0.016). VELO correlated with AREA (r = -0.552). Spleen volume changes correlated with APRI (r = 0.391) and AREA (r = 0.457), but did not change with SI. Logistic regression identified APRI and VELO as the only two independent predictors of SI. APRI ≥ 0.36 together with VELO ≤14.35 cm/s predicts SI with 100% sensitivity and 83% specificity, while area under receiver operating characteristic curve was 0.920. Conclusion: APRI combined with VELO can predict SI with high accuracy. Identified APRI threshold value was the same as in the previously published papers. Spleen volume changes and other portal vein hemodynamics parameters were less accurate in SI prediction.

HPB 2014, 16 (Suppl. 1), 1–99

O-G.10 PATIENT AND TUMOR BIOLOGY PREDICT SURVIVAL BEYOND THE MILAN CRITERIA IN LIVER TRANSPLANTATION FOR HEPATOCELLULAR CARCINOMA A. Andreou, P. Neuhaus and D. Seehofer Department Of General, Visceral And Transplant Surgery, Charité Campus Virchow Klinikum, Berlin, BERLIN Background: Milan criteria (MC) are currently the most established selection tool for liver transplantation (LT) in patients with hepatocellular carcinoma (HCC). Patients with HCC beyond MC are not considered for LT in many centers; however, LT may be the only treatment able to prolong survival in these patients with unresectable HCC. The aim of this study was to assess the role of recipient age and tumor biology expressed by DNA-index in the selection of HCC patients for LT. Methods: Clinicopathological data of 364 patients with HCC who underwent LT in a high-volume transplant center between 1989 and 2010 were evaluated. Overall survival (OS) was analyzed by tumor burden based on MC, patient age and DNA-index. Results: After a median follow-up of 78 months, the median survival was 100 months. Factors associated with OS in univariate analysis included MC, patient age, DNA-index and α-fetoprotein-level. In multivariate analysis, HCC beyond MC, DNA-index >1.5, and α-fetoprotein-level >200 ng/mL independently predicted worse OS. When stratifying patients by both MC and age, patients younger than 60 years with HCC beyond MC had OS comparable to that of older patients within MC (10-year-OS 32% vs. 37%, P = .008, Figure1). Patients < 60 years old with HCC beyond MC but DNA-index < 1.5 had excellent OS (Figure2), comparable to that of patients within MC. Conclusions: Patients younger than 60 years may undergo LT for HCC with favorable outcomes independently from their tumor burden. Additional assessment of tumor biology using DNA-index within this subset of patients can contribute to the selection of liver transplant candidates who may derive the most long-term survival benefit beyond MC. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

50

Abstracts

O-G.11 A 20 YEAR REVIEW OF LIVER TRANSPLANTATION: CHANGING INDICATIONS OVER 2 DECADES M. Doyle, N. Vachharajani, J. A. Lowell, S. Shenoy, J. R. Wellen , Y. Lin and W. C. Chapman Washington University School Of Medicine , St. Louis, MO

Conclusion: This 20 yr single center liver transplant program demonstrates excellent long term patient survival and outcomes. There has been a significant shift in the use of transplantation for HCC, HCV, and NASH, in the last 10-year period.

Introduction: Liver transplantation(OLT) has evolved over the past 20 years becoming a successful treatment for liver failure. Aim: We review our results with regard to indications, trends and failure. Methods: Using a prospectively database, we reviewed adult recipients undergoing OLT since 1993. Demographics, indications and survival were determined, as well as an analysis of graft loss and mortality. Results: Between 1/1/1993 and 12/31/2012, 1283 adult OLTs were performed. 37% (475/1283) transplants occurred before 2003. 1-, 5-, 10-, and 20 yr patient survival was 88.8%, 75.9%, 62.5% and 39.4% (median 16.1 yrs). Median graft survival was 14.5 yrs, with 30-day graft loss of 1.8%, usually related to HAT and PNF. 30-day mortality was 4.2% and was primarily due to bleeding or sepsis. Late deaths (>10 yrs) were primarily related to cardiac, hepatorenal, nonliver malignancy, or sepsis. Late graft loss was mainly due to cholangiopathy, rejection, chronic graft dysfunction and disease recurrence. 75 patients were retransplanted, with 1, 5- and 10-yr patient survival after retransplantation of 76.6%, 58.2% and 53.1%. OLT for cancer (35 vs 10.9%), HCV (43 vs31%) and NASH (7vs 0.9%) was more common in the second decade (>2003) and less common for PBC/ PSC (8.4 vs22%)(p < 0.001). Extended criteria donors were more likely to be used in the second decade (20.2% vs11.1% p < .0001). © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

O-G.12 A SIMPLIFED FRAILTY INDEX TO ACCURATELY PREDICT MORBIDITY AND MORTALITY AFTER HEPATIC RESECTION L. M. Louwers, G. Schnickel and I. Rubinfeld Henry Ford Health System, Detroit, MI Background: Frailty has been established as an important predictor of surgical outcomes. We hypothesized our modified frailty index (mFI) would be a helpful assessment for serious morbidity and mortality among patients undergoing hepatic resection. Methods: The National Surgical Quality Improvement Project (NSQIP) database was queried for all patients who had liver resections from 2005–2011. A previously described and validated mFI was calculated on the basis of NSQIP variables. The impact of frailty on 30-day serious postoperative morbidity (Clavien level 4, i.e. cardiac failure, respiratory failure, acute renal failure, sepsis, stroke, etc) and mortality was evaluated. Data was analyzed using chisquared and multivariate logistic regression. Analysis was performed using SPSS 21 software, with p < .05 defined as significant. Results: Within the 10,300 patients (6189 partial lobectomy, 3094 lobectomy, 1017 extended hepatectomy), the 30-day Clavien 4 complication and mortality rates were 7.5% (775/10300) and 2.3% (242/10300), respectively. Mean (range) mFI was 0.12 (0–0.67) in the Clavien 4 complication group and 0.13 (0–0.67) in the mortality group, versus 0.08 among patients who did not experience morbidity or mortality, p < 0.001 for both groups. Logistic regression using the mFI with NSQIP variables of age, ASA score, wound class, and emergency status demonstrated mFI was the strongest predictor of Clavien 4 complications (OR 2.61, p = 0.003) and mortality (OR 3.37, p = 0.022) Conclusion: Frailty may be an important predictor of morbidity and mortality among patients undergoing hepatic resection. The mFI is a simple and accurate tool, based on data collected in the patient history, to facilitate preoperative risk assessment.

O-G.13 ASSOCIATING LIVER PARTITION AND PORTAL VEIN LIGATION FOR STAGED HEPATECTOMY (ALPPS) VERSUS PORTAL VEIN EMBOLIZATION IN THE MANAGEMENT OF COLORECTAL LIVER METASTASES R. Hernandez-Alejandro1 and K. P. Croome1,2 Western University Canada, London, ONTARIO; 2Mayo Clinic, Rochester, MN 1

Background: Meticulous selection of those patients that can safely undergo the ALPPS procedure will be paramount in order to minimise the morbidity and mortality associated with the procedure. Methods: A total of 10 consecutive patients undergoing the ALPPS procedure and 15 patients undergoing portal vein embolization (PVE) with planned resection for colorectal liver metastases (CRLM) between the dates of May 1 2008 and May 1 2013 were identified. Results: A significantly faster and more extensive volume increase of the standardized future liver remnant (sFLR) was HPB 2014, 16 (Suppl. 1), 1–99

51

observed following the ALPPS procedure (89%), when compared to PVE (52%). All 10 ALPPS patients completed the two staged hepatectomy compared to only 11/15 in the PVE group. There was no significant difference in the total number of complications (p > 0.99) or severe complications (≥IIIB) (p = 0.84) in the two groups. No deaths were observed in the ALPPS group. Mean time off chemotherapy was shorter 130.25 ± 31.4 in the ALPPS group compared to 213.4 ± days in the PVE group (p < 0.001). Mean time until return to chemotherapy was 70.8 ± 8.1 days in the ALPPS group and 93 ± 45.9days in the PVE group (p = 0.15). At time of maximal follow-up in patients who completed the planned curative resection, recurrence occurred in 10% of patients in the ALPPS group and 36% of patients in the PVE group (p = 0.16). Conclusions: In appropriate selected patients the ALPPS procedure can be performed with low mortality and morbidity rates that are similar to those demonstrated with PVE. The ALPPS approach is a potential valid option to enable resection in patients with CRLM previously considered unresectable by standard techniques, offering several advantages over PVE.

O-G.14 COMPARISON OF SURVIVAL AND TOLERABILITY OF LIVER RADIOEMBOLIZATION IN ELDERLY VS. YOUNGER PATIENTS WITH METASTATIC COLORECTAL CANCER S. Tohme1, G. Nace1, D. Sukato1, J. Steele1, A. Zajko2, N. Amesur2, J. W. Marsh1, D. Geller1 and A. Tsung1 1 Department Of Surgery, University Of Pittsburgh School Of Medicine, Pittsburgh, PA; 2Department Of Radiology, University Of Pittsburgh School Of Medicine, Pittsburgh, PA Aim: To evaluate the outcomes among elderly (≥70 years) and younger patients (<70 years) with liver-dominant metastatic colorectal cancer (mCRC) who received radioembolization as a salvage therapy. Methods: A retrospective review of 107 consecutive patients with unresectable mCRC who were treated with radioembolization after failing first and second line chemotherapy. Demographics, treatment doses, biochemical and radiologic response, toxicities, and survival were examined. Results: Between 8/2002 and 4/2012, 44 elderly and 63 younger patients received radioembolization. Patients in both groups had similar previous extensive chemotherapy and liver-directed interventions. Using RECIST criteria, either stable or partial response was seen in 65.8% of the younger vs. 76.5% of the elderly patients (p = 0.357). Radioembolization was equally well tolerated in both groups and common procedure-related adverse events were predominantly grade 1–2 and of short duration. No significant difference was found with regard to overall median survival between younger (8.4months; 95% CI = 6.23–10.63) or elderly patients (8months; 95% CI = 4.51–11.43, p = 0.35). The presence of extrahepatic disease at the time of radioembolization was associated with a significantly worse median survival in both groups. A significant improved survival was seen in elderly patient whose CEA responded to radioembolization compared to those who didn’t (13.6 months; 95% CI = 2.37–24.77 vs. 3.0months; 95% CI = 2.76–3.24, p = 0.01). © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

52

Abstracts

Conclusion: Radioembolization appears to be as well tolerated and effective for the elderly as it is for younger patients with mCRC. Age alone should not be a discriminating factor for the management of mCRC patients.

O-G.15 WHAT IS THE CURRENT ROLE OF PERIHEPATIC DRAIN (PD) PLACEMENT IN LIVER SURGERY? A PROSPECTIVE ANALYSIS J. M. Butte, J. Grendar, O. Bathe, F. Sutherland, C. G. Ball and E. Dixon Foothills Medical Centre, , Calgary, ALBERTA Background: The standard use of PD in hepatic surgery is controversial and mainly supported by retrospective data.This study evaluates the role of PD in liver surgery. Methods: All patients included in a previous, randomized trial(RCT) that evaluated the role of mucomyst in liver resection were analyzed to determine the association between PD placement, postoperative complications (PC), and their initial treatment.A multivariate analysis (MVA) determined predictor factors of PC. A p < 0.05 was considered significant. Results: One hundred-ninety nine patients were analyzed [119 (60%) males; median age = 60 years (22–84)]. The majority of patients had colorectal liver metastasis (n = 114, 57%). PD (n = 87, 44%) were used most commonly in patients that had preoperative biliary instrumentation (p = 0.023), intraoperative bleeding (p = 0.01), Pringle’s maneuver (p < 0.001), and a more extensive resection (p = 0.001). Seventy-seven (39%) patients had a PC, which was associated with preoperative biliary instrumentation (p = 0.048), extent of resection (p = 0.002), and postoperative blood transfusions (p = 0.001). Patients with PD had a slightly higher rate of PC (46 vs. 33%) (p = 0.078) and a significant number of high grade PC (25% vs. 12%, p = 0.008). Nineteen patients(9.5%) developed a postoperative collection, with a similar incidence in those with (n = 10, 11.5%) and without (n = 9, 8%) PD, p = 0.47. Seven (8%) patients treated with and 9 (8%) treated without an PD needed a second drain after surgery (p = 1).Resection of ≥3 segment was the only factor found to be associated with PC (OR = 2, p = 0.025, 95% CI 1.1–3.7) in the MVA. Conclusions: Despite preferential PD use in patients with more complex tumors, they did not decrease the rate of PC, postoperative collections, and the need for the placement of a percutaneous postoperative drain.PD should be reserved for exceptional circumstances in liver surgery.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

SATURDAY, FEBRUARY 22, 2014, 2:30PM–4:30PM PARALLEL ORAL ABSTRACT H PANCREAS/TRANSPLANT/ BILIARY O-H.01 PANCREATIC FISTULAS AFTER PANCREATICO-DUODENECTOMIES: ARE PANCREATICO-GASTROSTOMIES SAFER THAN PANCREATICOJEJUNOSTOMIES? A QUASI-EXPERIMENT AND PROPENSITY-SCORE ADJUSTED ANALYSIS J. Abou Khalil, S. Dumitra, M. Jamal, N. Mayo, P. Chaudhury, P. Metrakos and J. Barkun McGill University Health Centers, Montreal, QC Background: Pancreatic fistula (PF) is a major contributor to morbidity and mortality after pancreaticoduodenectomy (PD). There remains a debate as to whether reestablishing pancreaticoenteric continuity through a PancreaticoGastrostomy (PG,) compared to a Pancreatico-Jejunostomy (PJ,) can decrease the risk of PF and total complications GOAL We compared surgical outcomes of PG and PJ after PD at our institution, where patient assignment to PG or PJ is consistent with a Quasi-Experiment. Methods: Data on pre-operative factors and post-operative complications was collected for patients undergoing PD between 1999 and 2011. We performed propensity-score adjusted logistic regression to identify the effect of PG on outcomes of PF, delayed gastric emptying (DGE), and total complications. We used both ISGPF and Strasberg and Linehan definition for PF and the ISGPS definition for DGE. Total morbidity was assessed using the Clavien-Dindo classification and the Comprehensive Comorbidity Index (CCI). Results: 23/103 and 20/103 (p = 0.49) of patients had PF and 74/103 and 55/103 patients had all-grades DGE in the PG and PJ groups respectively (p = 0.02). Groups did not differ with regards to Clavien-Dindo grade (p = 0.29) but did differ with regards to the CCI (38.4 vs. 31.4 for PG vs. PG respectively p = 0.02.) Propensity-score adjusted multivariate analysis showed no effect of surgical technique on PF (p = 0.89), DGE grades B/C (p = 0.9) or CCI (p = 0.41) but there remained an effect on all-grades DGE (p = 0.012). Conclusion: Patients undergoing PG did not have less PF or total complications than those reconstructed with PJ after PD at our institution; odds of all-grade DGE were higher with PG.

HPB 2014, 16 (Suppl. 1), 1–99

53

O-H.02 A NOVEL ONCOLYTIC ADENOVIRUS TARGETING C-MET RECEPTOR INCREASES INFECTIVITY EFFICIENCY AND ONCOLYTIC ACTIVITY IN HUMAN PANCREATIC ADENOCARCINOMA CELL LINES Q. D. Chu, G. Sun, L. Liu, B. Li, R. H. Kim and J. M. Mathis Louisiana State University Health Sciences Center-Shreveport, Shreveport, LOUISIANA Introduction: Virotherapy using oncolytic adenoviruses (Ads) is an innovative approach to cancer therapy. The c-Met receptor and its ligand, hepatocyte growth factor (HGF), are involved in a wide array of neoplastic diseases, including pancreatic cancer. We created a novel oncolytic adenovirus by incorporating NK2, a competitive antagonist of the HGF/ c-Met association, into the Ad serotype 5 knob fiber. We hypothesized that this new oncolytic Ad will result in efficient viral transduction and cell killing in an array of human pancreatic cancer cell lines. Methods: Human pancreatic cancer lines CFPAC-1, AsPC-1, and BxPC-3 were obtained from the American Type Culture Collection. Efficiency of Ad infection in the cells was determined using an vector construct expressing the green fluorescence protein (GFP) marker in place of the E1A gene (Ad5-RFP-FF-NK2) and quantified by flow cytometry. Oncolytic activity in the pancreatic cancer cells was determined using the oncolytic Ad5-NK2-E1A by a crystal violet staining method. Results: C-Met was overexpressed in all three pancreatic cancer cell lines. The oncolytic Ad5-NK2 was more efficient in transducing pancreatic cancer cell lines as well as cell killing than wild type adenoviruses (sample graphs below). Conclusion: The novel recombinant oncolytic Ad5-NK2 efficiently transduced and killed human pancreatic cancer cells. This novel construct represents a potential new therapy against pancreatic cancer.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

54

Abstracts

O-H.03 NEOADJUVANT FOLFIRINOX APPLICATION IN BORDERLINE RESECTABLE PANCREATIC ADENOCARCINOMA A. Paniccia, B. Edil, R. Schulick, C. Gajdos, J. T. Byers, C. Meguid and M. McCarter University Of Colorado Hospital, Aurora, CO Introduction: FOLFIRINOX has not been used extensively in the neoadjuvant setting because of concerns with toxicity. We evaluated our institutional experience with neoadjuvant FOLFIRINOX in borderline resectable pancreas adenocarcinoma (BRPAC). The primary endpoints were completion of therapy to surgery, and R0 resection rate. Methods: Retrospective analysis of patients with BRPAC treated with neoadjuvant FOLFIRINOX. Two pancreatic surgeons reviewed all diagnostic and preoperative radiographic imaging. Between August 2011 and June 2013, 17 patients with BRPAC treated with neoadjuvant FOLFIRINOX were identified. Results: 94% of patients were able to undergo surgery. One patient experienced toxicity. 10 pancreatic head lesions (63%), 5 body lesions (31%) and 1 synchronous head and tail lesion (6%). Abutment of PV/ SMV in 11 cases (69%), and short segment PV encasement in 2 cases (19%). SMA abutment in 1 case (6%), and concomitant abutment of the celiac axis and PV in 1 case (6%). Patients received a median of 4 cycles of FOLFIRINOX. At a median interval of 10 weeks from the first dose, 7 patients (44%) had partial radiologic response and 9 patients (56%) had stable disease. There were no cases of progression. Vascular resection was necessary in 7 cases (44%). All patients underwent margin negative resection (R0). Treatment effect in the pancreatic specimens © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

revealed poor or no response in 8 cases (50%), moderate response in 4 cases (25%) and marked response in 4 cases (25%). Conclusion: Neoadjuvant FOLFIRINOX toxicity did not prevent surgery and allowed a high rate of R0 resection.

O-H.04 PERIOPERATIVE RISK FACTORS FOR DELAYED GASTRIC EMPTYING IN PANCREATICODUODENECTOMY J. R. Robinson2, P. Marincola3, J. Shelton2, N. Merchant1, K. Idrees1 and A. Parikh1 1 Vanderbilt University Medical Center, Nashville, TN; 2 Vanderbilt University Medical Center, Nashville, TN; 3 Vanderbilt University School Of Medicine , Nashville, TN Introduction: Delayed gastric emptying (DGE) is a frequent complication of pancreaticoduodenectomy (PD) responsible for increased morbidity, prolonged hospital stay, and readmission. We sought to further define the incidence of DGE after PD based on the International Study Group of Pancreatic Surgery (ISGPS) definition and evaluate associated perioperative risk factors. Methods: 420 consecutive patients who underwent PD at our tertiary referral center were reviewed. Using a stepwise multivariate logistic regression model, perioperative factors associated with the development of clinically significant DGE defined by ISGPS grades B and C were assessed. Results: Clinically significant DGE occurred in 24% of patients (N = 98) with Grades B and C occurring at 13.5% (N = 55) and 10.5% (N = 43), respectively. Grade B or C DGE patients had significantly longer mean hospital stay (20 vs. 9 days, p < 0.001) and higher rate of 90-day readmission HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

55

(32% vs. 18%, p = 0.004) than those without DGE. By univariate analysis, white race, increasing ASA class, BMI > 35, OR length >6 hrs, preoperative weight loss and octreotide use were associated with increased risk of DGE. By stepwise multivariable regression, BMI > 35 (OR = 3.42), OR length >6 hrs (OR = 2.97), preoperative weight loss (OR = 2.05), and use of octreotide (OR = 2.33) were independently associated with increased risk of DGE. Conclusion: DGE remains a significant cause of morbidity, increased hospital stay, and higher readmission rates after PD. Our findings suggest patients with BMI > 35, preoperative weight loss, or longer OR times be considered for gastrostomy and/or jejunostomy tube placement and limited octreotide use to decrease the potential risk and consequences of DGE.

O-H.05 COST COMPARISON ANALYSIS OF OPEN VS. LAPAROSCOPIC DISTAL PANCREATECTOMY D. R. Rutz1, M. H. Squires1, S. K. Maithel1, J. M. Sarmiento2, J. W. Etra1, S. D. Perez2, W. Knechtle2, K. Cardona1, M. C. Russell1, C. A. Staley1, J. F. Sweeney2 and D. A. Kooby1 1 Department Of Surgery, Division Of Surgical Oncology, Winship Cancer Institute, Emory University, Atlanta, GEORGIA; 2Department Of General And Gastrointestinal Surgery, Emory University, Atlanta, GEORGIA Introduction: As compared with open distal pancreatectomy (ODP), laparoscopic distal pancreatectomy (LDP) carries fewer complications and shorter hospital stays for selected patients, but cost data between the two approaches are limited. We analyzed operative and hospital costs of both approaches at our institution. Methods: Records from all DPs from 1/2009–6/2013 were reviewed and stratified by complexity (standard vs. high). Patient factors and outcomes were recorded. Total variable costs (TVC) were tabulated for each patient, and then by category (e.g. “floor,” “OR,” “radiology,” etc . . .). Costs were compared between LDP and ODP groups for index admissions and any readmissions. Differences in means between groups were compared via independent samples T-tests. Categorical variables were analyzed by chi-square test. Results: Of 153 procedures, 115 were ‘standard’ complexity (70 LDP, 45 ODP) and were analyzed. The LDP and ODP groups had similar patient demographics and complication rates (Table). TVC of the index admission was $3,418 less per patient in the LDP group ($10,484 vs. $13,903, p = 0.059). Although operating costs were significantly greater in the LDP cohort ($5,756 vs. $4,900, p = 0.017), the shorter hospitalization for the LDP group (5.2 vs. 7.7 days, p = 0.009) resulted in lower overall cost. Total cost of index hospitalization combined with readmission was significantly less in the LDP group ($11,106 vs. $14,803, p = 0.05). Conclusion: For appropriately selected patients, the laparoscopic approach to distal pancreatectomy is more costeffective than the open approach. Increased OR cost associated with LDP is offset by the shorter hospitalization. These data provide targets for further cost reduction.

HPB 2014, 16 (Suppl. 1), 1–99

O-H.06 FAILURE OF NARCOTIC WEAN AFTER TOTAL PANCREATECTOMY WITH ISLET AUTOTRANSPLANTATION K. A. Morgan, S. M. Owczarski, H. Wang, W. Balliet, J. J. Borckardt and D. B. Adams Medical University Of South Carolina, Charleston, SC Background: Pancreatectomy with islet autotransplantation (TPIAT) is an effective means of pain relief in selected patients with debilitating pain from chronic pancreatitis. Patient selection remains challenging. Despite significant improvements in quality of life, many patients continue to require daily narcotic medication postoperatively. We sought to identify predictive factors for success at postoperative narcotic weaning. Methods: A prospectively collected database of patients undergoing TPIAT between March 2009 and March 2013 was reviewed, with attention to patient characteristics, daily morphine equivalents, and quality of life (SF-12), from the preoperative period and up to 36 months postoperative. Results: 104 patients (81 women, age 41) underwent TPIAT and had at least 6 months follow-up data. Mean physical QOL (pQOL) improved from 27.3 preoperatively to 35.1 (p < 0.001) and mean mental health QOL (mhQOL) improved from 38.4 preoperatively to 44.7 (p = 0.001) at last postoperative follow-up (mean 1.7 years). Despite this, 27% of patients (29/104) did not wean significantly their narcotic requirement. Preoperative daily narcotic doses were 277 MEs (range 0–2518). Comparison of patients successful at narcotic wean (SNW) with those who failed (FNW), demonstrated they were similar in gender, age, disease duration, and etiology. While preop pQOL was similarly poor (28.6 v 27.1), FNW patients had lower preoperative mhQOL (33.5 v 38.6 p = 0.05). Conclusions: TPIAT results in significant improvements in QOL, but narcotic weaning is more complex. Patients with low preoperative mhQOL are less likely to wean narcotics postoperatively. Behavioral psychology intervention should be focused in this patient population. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

56

Abstracts

O-H.07 RISK FACTORS FOR DEVELOPMENT OF STEATOSIS AND STEATOHEPATITIS AFTER PANCREATICODUODENECTOMY L. K. Fonkoua, R. Tappouni, N. H. Bhayani, E. T. Kimchi, K. F. Staveley-O’Carroll, J. T. Kaifi and N. J. Gusani Program For Liver, Pancreas, & Foregut Tumors; Penn State College Of Medicine, Hershey, PA Introduction: A significant number of patients undergoing pancreaticoduodenectomy (PD) develop Non-Alcoholic Fatty Liver Disease (NAFLD) postoperatively, and some patients even progress to Non-Alcoholic Steatohepatitis (NASH). The etiology of these fatty changes in the liver after PD remains unclear. Methods: Patient demographics, operative procedures, pathology, imaging, and perioperative clinical variables for 102 patients undergoing PD between 2002 and 2010 were retrospectively examined, and the most influential risk factors for development of postoperative liver steatosis were identified by uni- and multivariate analyses. Results: Of 102 patients who underwent PD, 21 (20.6%) developed hepatic steatosis, as measured by a significant change of liver-to-spleen attenuation ratio (L/S ratio) on unenhanced computed tomography (CT) images during the follow-up period. Overall, patient demographics, preoperative findings, operation types, and final pathological findings did not differ significantly between the steatosis patients and the controls; however, the frequency of preoperative BMI >30 was higher in the steatosis group (48% vs 19%, p = 0.0098). In addition, this group had a lower proportion of pancreatic ductal adenocarcinoma patients (24% vs 76%, p = 0.0001) and a higher proportion of ampullary/duodenal carcinomas (48% vs 7.4%, p = 0.003) compared to the no steatosis group. A multivariate logistic regression analysis identified pancreatic ductal adenocarcinoma (p = 0.005) and postoperative chemoradiation therapy (p = 0.035) as independent risk factors for the development of hepatic steatosis. Conclusion: Postoperative hepatic steatosis occurs in 20.6% of patients after PD. Risk factors for development of steatosis include pathologic subtype of pancreatic ductal adenocarcinoma and receipt of adjuvant chemoradiation therapy.

O-H.08 CASE CONTAMINATION REPORTING IN HPB SURGERY: CAN A SINGLE WORD CHANGE PUBLIC PERCEPTION OF INSTITUTIONAL PERFORMANCE? P. J. Speicher, D. Nussbaum, J. E. Scarborough, R. R. White, C. R. Mantyh, D. S. Tyler and B. M. Clary Department Of Surgery, Durham, NC Introduction: The drive to improve patient outcomes and the inevitability of mandated public outcomes reporting necessitate uniform reporting processes and accurate administrative databases. While conducting an internal QI process, we identified non-uniformity within NSQIP in the reporting of case contamination status in patients undergoing HPB surgery and queried what the impact these inconsistencies would have on quality metrics used as benchmarks for comparison. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

Methods: 27,376 hepatic and pancreatic resections were identified from the 2005–2011 NSQIP Participant User Files. The effect of case contamination status on postoperative SSI rates was determined through stepwise forward multiple logistic regression modeling. The impact of alternative casecontamination reporting policies on observed : expected ratios for SSI was then simulated for poor, average, and high performing institutions. Results: Significant heterogeneity in contamination status reporting was present. For example, despite clear NSQIP guidelines prompting at least “clean-contaminated” designation, 8% of all patients were coded as “clean”. Although typically homogeneous in conduct, over 10% of pancreaticoduodenectomy and partial hepatectomy (with biliary/intestinal procedures) cases were classified as “contaminated” or “dirty” instead of “clean-contaminated”. Contaminated (AOR 1.39, p = 0.001) and dirty (AOR 1.42, p = 0.02) cases were associated with higher odds of SSI, while clean-contaminated were not (p = 0.99). O/E ratios were highly sensitive to modest changes in the proportion of HPB resections classified as “contaminated” or higher (Table 1). Conclusion: NSQIP SSI O : E ratios are easily modified by institutional contamination classification policies. As institutions work to improve outcomes and prepare for public reporting, it is imperative that all adhere to consistent outcomes reporting practices.

O-H.09 TARGETED INHIBITION OF THE AKT PATHWAY IN CHOLANGIOCARCINOMA BY MK2206 J. M. Wilson, S. Kunnimalaiyaan, M. Kunnimalaiyaan and T. Gamblin Medical College Of Wisconsin, Department Of Surgery, Division Of Surgical Oncology, Milwaukee , WI Introduction: Cholangiocarcinoma (CCA) is a rare disease with limited effective treatment options. The PI3K/Akt pathway represents an attractive therapeutic target due to its frequent dysregulation in CCA. MK2206, an allosteric Akt inhibitor, has been shown to reduce cellular proliferation in other cancer types. We hypothesized that MK2206 mediated inhibition of Akt in human CCA cells would impact cell viability and sought to establish a dose and time correlate. Methods: Human CCLP and SG231 CCA cell lines were treated with 0–2 μM MK2206. MTT assay was used to assess cellular viability at 96 hours. Expression levels of apoptotic markers and cell cycle proteins were determined via western blot. HPB 2014, 16 (Suppl. 1), 1–99

Abstracts Results: CCLP viability was significantly reduced at MK2206 concentrations of 0.5, 1, and 2 μM by approximately 44%, 53%, and 64% respectively (p < 0.01). SG231 also demonstrated significant reduction in viability of 32%, 32%, and 42% respectively for the same treatment concentrations (p < 0.05). Western analysis revealed a strong decrease in phosphorylated AKTSer473, while phosphorylated AKTThr308 remained unchanged in both cell lines. In addition, cleaved PARP expression was increased while pro-caspase 3 and 9 expression was reduced. Interestingly, survivin expression increased with treatment in both cell lines. Conclusion: Our study clearly demonstrates that by blocking phosphorylation of Akt at serine 473, CCA cellular growth is reduced. This growth suppression is mediated via apoptosis induction. Importantly, the concentrations needed for growth reduction in CCA are less than those needed in other cancer types. It is anticipated that these concentrations could be achieved in vivo with limited toxicity.

O-H.10 ARE GENERAL SURGERY RESIDENTS ADEQUATELY PREPARED FOR HEPATOPANCREATICOBILIARY (HPB) FELLOWSHIPS? QUESTIONNAIRE BASED STUDY H. Osman1, J. Parikh2, S. Patel1 and D. Jeyarajah1 Methodist Dallas Medical Center, Dallas, TX; 2Indiana University Hospital, Indianapolis, IN

1

Objectives: HPB fellowship addresses organ diseases that are not heavily covered during residency. Moreover, it is a technically demanding field. We aim to identify challenges that HPB fellows have experienced during the initial part of their HPB training. Methods: An online questionnaire was designed and distributed electronically to all HPB fellows (n = 20) in fellowship council accredited programs in the academic year 2012–2013. Responses were then analyzed. Results: Response rate was 45% (n = 9). Prior to their fellowship, 55.6% were in surgical residency, 33.3% were in another fellowships, while 1 fellow (11.1%) was in general surgery practice. 77.8% were graduates of university based residency programs. 77–88% felt comfortable performing basic general surgery procedures independently at the end of residency training. When uncomfortable performing basic or advanced surgical procedures, 77.7% cited case volume and lack of autonomy during residency as reasons. 44.4% cited inability to adequately perform preoperative workup for general surgical diseases as their biggest fear entering practice. 77.8–88.9% felt adequately prepared handling benign and malignant HPB diseases at the beginning of fellowship. Minimal invasive surgery and transplant rotations were cited as non-HPB rotations more likely to prepare residents for their upcoming HPB fellowship. Conclusion: HPB fellows felt overall to be adequately prepared for their fellowship. Case volume and intraoperative autonomy are the 2 biggest reason when fellows felt insufficiently prepared. General surgery residents who plan on pursuing HPB fellowship may benefit from spending extra rotation on certain subspecialties in addition to focus on preoperative workup and management.

HPB 2014, 16 (Suppl. 1), 1–99

57

O-H.11 PORTAL VENOUS SYSTEM DECOMPRESSIVE PROCEDURES DURING PANCREATIC RESECTION: PERIOPERATIVE AND LONG-TERM OUTCOMES M. J. Truty1, R. M. Thomas3, J. E. Lee2, M. H. Katz2 and J. B. Fleming2 1 Mayo Clinic College Of Medicine, Rochester, MN; 2 University Of Texas, MD Anderson Cancer Center, Houston, TX; 3University Of Florida College Of Medicine, Gainseville, FL Background: Sinistral portal hypertension may occur following splenic vein ligation and SMV/PV resection during pancreatectomy in the absence of adequate gastro-splenic collaterals. Varices associated with occluded SMV/PV increase risk of intraoperative hemorrhage due to mesenteric venous hypertension. Distal splenorenal shunts (DSRS) and temporary mesocaval shunts (MCS) can prevent such respective complications. Design: Patients that required portal decompressive procedures during pancreatectomy were identified. Clinicopathologic outcomes, shunt patency, and postoperative gastro-splenic complications were reviewed. Results: Twenty-one decompressive procedures were performed in 18 patients who required SMV/PV reconstruction. 16 patients had pancreatic adenocarcinoma. SMV/PV reconstruction was performed primarily in 7 (39%) patients and with interpositional grafting in 11 (61%) patients. 4 (24%) patients required simultaneous hepatic arterial resection/ reconstruction. The R0 resection rate was 94%. DSRS was constructed in 16 patients and 2 patients underwent splenoportal anastomoses. Temporary MCS was performed concomitantly in 3 patients. Mean (median) operative time, EBL, and LOS were 664 (643) min, 1111 (900) ml, and 13 (11) days, respectively. Major complications occurred in 5 (28%) patients. One patient developed bleeding gastropathy from acute sinistral hypertension due to splenoportal anastomotic thrombosis requiring emergency splenectomy. Postoperative imaging revealed mean (median) shunt patency of 12.8 (5.7) months. No patients developed late hypersplenism. Conclusion: Portal decompressive procedures during pancreatectomy are safe and may minimize intraoperative bleeding and the development of early/late sinistral hypertension after splenic vein ligation for SMV/PV reconstruction. DSRS may be preferable to splenoportal anastomoses due to thrombosis risk. The excellent early patency may minimize acute complications until appropriate collaterals develop.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

58

Abstracts

O-H.12 COST-EFFECTIVENESS AFTER PANCREATICODUODENECTOMY: BOLSTERING THE VOLUME ARGUMENT

Conclusions: For PD, lower volume centers have higher perioperative mortality, 30-day readmission rates, longer LOS and 10.9% higher cost per patient. Performance of PD at higher volume centers can lead to both better outcomes and substantial cost savings.

J. M. Sutton, G. C. Wilson, K. Wima, R. C. Quillin, I. M. Paquette, D. J. Hanseman, J. J. Sussman, M. J. Edwards, S. A. Ahmad, S. A. Shah and D. E. Abbott Department Of Surgery, University Of Cincinnati Medical Center, Cincinnati, OH Background: The cost implication of variability in pancreatic surgery is not well described. We hypothesized that for pancreaticoduodenectomy (PD), lower volume centers demonstrate worse perioperative outcomes at higher costs. Methods: 9,883 patients undergoing PD from 2009–2011 were identified from the University Healthsystems Consortium database and stratified into quintiles by hospital and surgeon volume. A decision analytic model was constructed to assess cost-effectiveness. Costs were based on Medicare cost : charge ratios and included readmission costs when applicable. Results: When both index and readmission costs were considered, the median per-patient cost at the lowest-volume centers was $23,005, or 10.9% ($2,263 per case) more than at the highest-volume center. Additionally, the lowest volume centers demonstrated inferior perioperative outcomes with a higher median overall length of stay (11 vs. 8 days, p < 0.001), higher perioperative mortality rate (3.5% vs. 1.3%, p < 0.001), and higher 30-day readmission rate (20% vs. 16%, p = 0.003) compared to the highest-volume centers. Similar to hospitals, low volume surgeons had higher perioperative mortality (3.1% vs. 1.2%; p = 0.004) and readmission rates (19% vs. 16%; p = 0.008). One-way sensitivity analyses adjusting for perioperative mortality (1.3% at all centers) and readmission rates (16% at all centers) did not materially change the cost-effectiveness analysis. Differences in cost were largely recognized in the index admission; readmission costs were similar across quintiles. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

O-H.13 PANCREATICODUODENECTOMY IN FLORIDA: DO 20 YEAR TRENDS DOCUMENT SALUTARY BENEFITS OF CENTRALIZATION OF CARE? T. W. Wood, S. B. Ross, A. E. Smart, C. E. Ryan, P. B. Sukharamwala and A. S. Rosemurgy Florida Hospital Tampa, Tampa, FL Introduction: Concentration of care has been promoted as fostering superior outcomes. This study was undertaken to determine if concentration of care is occurring in Florida for pancreaticoduodenectomy, and if so, is it having a salutary effect. HPB 2014, 16 (Suppl. 1), 1–99

Abstracts Methods: The data for pancreaticoduodenectomy were obtained from the Florida Agency for Health Care Administration for three 3-year periods:1992–1994, 2001–2003, 2010–2012; data were sorted by surgeon volume of pancreaticoduodenectomy during these periods and correlated to postoperative length of stay (LOS), in-hospital mortality, and hospital charges (adjusted to 2012 dollars). Results: Relative to 1992–1994, in 2010–2012 46% fewer surgeons undertook 115% more pancreaticoduodenectomies with significant reductions in LOS and in-hospital mortality, and higher charges (p < 0.001 for each) (Figure). In 1992– 1994, 62% of pancreaticoduodenectomies were undertaken by surgeons undertaking ≤1 per year; this fell to 13% in 2010–2012. Similarly, from 1992–1994 to 2010–2012 there was an 18-fold increase in the number of pancreaticoduodenectomies by surgeons undertaking ≥12 per year (n = 45 to n = 806, respectively). During 2010–2012, the more frequently surgeons undertook pancreaticoduodenectomy, the shorter LOS, the lower in-hospital mortality, the greater likelihood of discharge home, and the lower hospital charges (p < 0.03 for each) Conclusion: Over the last 20 years concentration of care has occurred in Florida with substantially fewer surgeons undertaking many more pancreaticoduodenectomies with dramatic improvements in LOS and in-hospital mortality, albeit with increased hospital charges. During 2010–2012, busiest surgeons got the best results, since the frequency with which surgeons undertook pancreaticoduodenectomy favorably impacted LOS, in-hospital mortality, the likelihood of discharge home, and hospital charges.

59

Multivariable Cox-proportional hazards modelling was used to evaluate the independent prognostic value of the two staging systems. Results: 302 patients were analyzed. The median overall survival was 38 months and 5-year survival was 37%. In the 7th edition 64% of patients were upstaged from the 6th edition. Both staging schemes were strong prognostic factors on univariate analysis: p = 0.0007 (6th) and P < 0.00001 (7th). The 7th edition more effectively stratified survival by stage: median survival for stage 0 (carcinoma in situ) was 80 months; stage I (confined to bile duct) 90 months; stage II (beyond bile duct) 47 months; stage III (node positive) 23 months; and stage IV (invasion main portal vein, common hepatic artery, second-order bile ducts, or distant metastasis) 24 months. Neither the 6th nor the 7th editions were independent predictors of survival in a multivariable model including margin status, nodal status, tumor differentiation, and perineural invasion. The AJCC staging (7th edition) was an independent prognostic factor only for patients with an R0 resection. Conclusion: The 7th edition of the AJCC staging system for HCCA improves stratification, but its prognostic value is lost in the face of other, more potent variables.

O-H.14 AJCC STAGING FOR RESECTED HILAR CHOLANGIOCARCINOMA: 7TH EDITION BETTER THAN 6TH, BUT NOT AN INDEPENDENT PROGNOSTIC FACTOR B. Groot Koerkamp1, J. K. Wiggers2, P. J. Allen1, O. R. Busch2, M. I. D’Angelica1, R. P. Dematteo1, Y. Fong1, D. J. Gouma2, T. P. Kingham1, T. M. Van Gulik2 and W. R. Jarnagin1 1 Memorial Sloan-Kettering Cancer Center, New York, NEW YORK; 2Academic Medical Center Amsterdam, Amsterdam, NOORD-HOLLAND Background: This study evaluates the prognostic value of the 6th and 7th edition of the AJCC staging system for patients with resected hilar cholangiocarcinoma (HCCA). Methods: Patients with HCCA resected between 1991 and 2012 were identified from prospective databases from two institutions. Kaplan-Meier median overall survival was compared between stage groups with the logrank test. HPB 2014, 16 (Suppl. 1), 1–99

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

60

Abstracts

O-H.15 HIGH-VOLUME SURGEONS VERSUS HIGH-VOLUME HOSPITALS: ARE BEST OUTCOMES MORE DUE TO WHO OR WHERE? S. B. Ross1, A. N. Igartua1, M. DeGori1, A. F. Teta1, K. Luberice1, P. G. Toomey2, P. B. Sukharamwala1 and A. S. Rosemurgy1 1 Florida Hospital Tampa, Tampa, FL; 2University Of South Florida Department Of Surgery, Tampa, FL Introduction: High-volume hospitals are purported to provide ‘best’ outcomes. High-volume hospitals and highvolume surgeons are inextricably related. We undertook this study to evaluate the outcomes after pancreaticoduodenectomy when high-volume surgeons relocate to a lowvolume hospital. Methods: Outcomes after the last 50 pancreaticoduodenectomies at a high-volume hospital in 2012 were compared to the outcomes after the first 50 pancreaticoduodenectomies undertaken at a low-volume (i.e., no pancreaticoduodenectomies in >5 years) hospital in 2012–13. Data are presented as median (mean ± SD). Results: Patients undergoing pancreaticoduodenectomies at a high-volume vs. a low-volume hospital were not different relative to sex and age (Table). Patients operated upon at the low-volume hospital had lower ASA class, had shorter operations with less blood loss, spent less time in the ICU, and had shorter length of stay (p < 0.05 for each, Table); 30-day mortality and 30-day readmission rates were not different (Table). Conclusion: The salutary benefits of being a high-volume hospital for pancreaticoduodenectomy are transferred to a low-volume hospital when high-volume surgeons relocate. The ‘best’ outcomes associated with high-volume hospitals may be primarily due to the high-volume surgeons who work there, and the ‘best’ results follow the high-volume surgeons. The ‘best’ results seem to be related more to who does the pancreaticoduodenectomy rather than where the pancreaticoduodenectomy is undertaken.

SATURDAY, FEBRUARY 22, 2014, 6:15PM–7:45PM ORAL POSTER II (LIVER III, IV; PANCREAS III, IV; TRICKS OF THE TRADE I; BILIARY II) OP-II.01 EXPRESSION OF FATTY ACID BINDING PROTEINS IN A MOUSE MODEL OF OBESITY-ASSOCIATED HCC K. J. Thompson, S. S. Nazari, D. A. Iannitti, I. H. McKillop and D. Sindram Carolinas Medical Center, HPB Surgery, Charlotte, NC HCC is the 3rd leading cause of cancer-related mortality. Risk factors for HCC include viral hepatitis and chronic alcohol consumption, with obesity emerging as an important risk factor. The aim of this study was to investigate expression of fatty acid binding proteins (FABPs) as potential biomarkers of obesity-associated HCC. C57BL/6 mice were injected with vehicle (Veh) or diethylnitrosamine (DEN; 5 mg/kg) at 21–24 d. At 5 wks, mice were placed on control diet (CD; 10% kcal%/fat) or high fat diet (HFD; 60% kcal%/fat). At 42 wks livers examined for gross tumor incidence and histology performed. To assess FABP expression, RT-qPCR and immunoblots were performed. Animals on HFD-Veh formed large tumors in 30% of animals, an effect exacerbated by DEN administration (90%), compared to small tumors in 60% in CD-DEN. Expression of FABP1 was upregulated in CD-DEN and HFD-DEN compared to Veh-injected animals. FABP2 was unchanged between CD-Veh and –DEN and HFD-Veh animals. However; there was a 2-fold down regulation of FABP2 in livers of HFD-DEN non-tumor tissue (NT) compared to HFD-DEN tumor (T) tissue. There was ∼1000-fold upregulation of FABP4 in HFD animals compared to CD animals, an effect exacerbated in DEN-injected animals. FABP5 expression was downregulated in HFD-Veh and HFD-DEN-NT tissue; however, HFD-DEN-T tissue expressed 3-fold higher levels. FABPs are dysregulated in obesity-associated HCC. Interestingly, there is a reciprocal relationship between expression of FABPs 1, 2 and 5 in tumor tissue from HFD-DEN mouse livers compared to NT liver tissue. These findings suggest FABPs may be biomarkers for obesity-associated HCCs.

OP-II.02 SURVIVAL FOLLOWING MULTIDISCIPLINARY MANAGEMENT OF EXTENSIVE COLORECTAL LIVER METASTASES J. Pasko, J. Wagner, O. Lafi, B. Sheppard, S. Orloff, G. Vaccaro, C. Lopez and K. G. Billingsley Oregon Health And Sciences University, Portland , OR Introduction: Developments in surgical technique and utilization of peri-operative oxaliplatin based chemotherapy have provided the opportunity of resection to patients with extensive colorectal liver metastases (CRLM). The aim of this study is to assess survival and disease recurrence following resection for patients with advanced CRLM. Methods: This is a single institution retrospective review of patients with colorectal liver metastases from June 2003 to © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts May 2011. During the study 168 patients underwent liver resection. This study included 28 patients with >4 liver metastases. Survival is calculated using the Kaplan Meier estimate. This study was approved by this institution’s IRB. Results: The median age was 59 years, 71% were male. Seventy-five percent had synchronous metastases at the time of initial operation; 61% received preoperative chemotherapy. Operations included: trisegmentectomy 46%, lobectomy 25%, segmental resection 29%. The sizes of largest tumor ranged from 1.1 cm to 9.2 cm and median size was 3.2 cm. Positive margins after resection occurred in 39%. Seven patients (28%) underwent a second liver resection. Recurrent disease developed in 26 patients, 9 (35%) liver only, 17 (65%) with extrahepatic disease. The 30 day mortality was 3%. Survival at 3 and 5 years is 46% and 18%. The overall median survival was 38 months. The median disease free survival was 10 months. The median follow-up was 34 months. Discussion: These results indicate few patients with extensive CRLM achieve long-term cure. However, resection is safe and most patients experience extended treatment-free periods. High tumor number alone should not contraindicate resection.

OP-II.03 PROSPECTIVE ANALYSIS OF TACE AS PRIMARY THERAPY IN HCC MANAGEMENT K. Collins, N. Saad, N. Vachharajani, K. Fowler, K. Korenblat, M. Doyle and W. Chapman Washington University School Of Medicine, St. Louis, MO Introduction: Transarterial Chemoembolization remains first-line therapy for patients with HCC, however the effectiveness of response is unknown. We conducted a prospective analysis of TACE as primary therapy in patients with HCC and chronic liver disease. Methods: Between 7/1/2009 and 12/31/2012, we performed 761 TACE in 398 HCC patients. 10 cases of biphenotypic tumors were excluded, as were 17 cases of radioembolization.359 were newly identified HCC cases whereas 39 were recurrent HCC who had one of the following treatments in the past with complete response at that time: OLT (1 case); Resection (12 cases); RFA (1 cases); TACE (25 cases). 192 cases were outside of Milan at the time of treatment. (Stage III – 104; Stage IVA1 – 58; Stage IVA2 – 30). Results: 25/192 (13%) initially beyond Milan have undergone OLT and (4.7%) are currently listed. 5 (2.6%) patients died while waiting on list and 1 (0.5%) needed to be taken off-list due to disease-progression. 68 cases could not be listed for OLT. Reasons were as follows: Disease-extent (30 (15.6%) cases), comorbidities (9 (4.7%) cases), social/ insurance/non-compliance related problems (13 (6.8%) cases), unwilling for transplant (13 (6.8%) cases), patient died before the evaluation was completed (3 (1.6%) cases). Of the remaining 208 patients that came to us as WithinMilan cases, 75 have been transplanted. Conclusion: TACE is an effective therapy for patients within Milan and for those beyond Milan, both for OLT candidates and for those who may downstage to within Milan. This has become the standard of care for HCC in non-operative candidates. HPB 2014, 16 (Suppl. 1), 1–99

61

OP-II.04 MINIMALLY INVASIVE MICROWAVE ABLATION OF EARLY STAGE HEPATOCELLULAR CARCINOMA F. Alemi, K. Specht, K. Kowdley, A. Siddique, R. Crane, E. Hauptmann, M. Fotoohi, A. Alseidi, T. R. Biehl, S. Helton and F. G. Rocha Virginia Mason Medical Center, Seattle, WA Background: Efficacy of microwave ablation (MWA) for liver tumors has been difficult to compare given the heterogeneity of patient selection, therapies, and tumor types in published studies. Herein we report our experience with minimally invasive MWA in a singular cohort of early stage hepatocellular carcinoma (HCC). Methods: All patients with HCC who underwent percutaneous or laparoscopic MWA from 2011–2013 were retrospectively analyzed. Treatment decisions were based on Barcelona Clinic Liver Cancer staging criteria after discussion at a multidisciplinary tumor board. Standard power-time parameters were utilized and imaging was obtained 30 days post-procedure. Results: Forty-eight HCC patients underwent MWA of 63 lesions, 33 percutaneously and 30 laparoscopically. All patients had cirrhosis, 19 exhibited portal hypertension with a mean Child’s score of 6.3 and MELD of 10.2. Mean tumor size was 2.0 cm (range 0.7–3.5) with an average of 1.43 ablations/tumor. Length of stay was 1.1 days in the laparoscopic group while all percutaneous treatments were outpatient. There were four complications (2 pneumothorax, 1 wound infection, and 1 arrhythmia) and four local recurrences (6.3%; three percutaneous and one laparoscopic, p = 0.61). Thirteen patients developed a new lesion at a mean of 6.6 months. There was no difference in total cost between laparoscopic or percutaneous groups ($36,508 versus $30,510, p = 0.15). Overall survival was 87.5% with all deaths due to end-stage liver disease. Conclusion: MWA of early stage HCC is safe and effective through laparoscopic and percutaneous routes. Further studies should be focused on selecting patients who may derive benefit and value from these approaches.

OP-II.05 LIVER RESECTION WITH HEPATIC VEIN RECONSTRUCTION IS A VIABLE OPTION IN SELECTED PATIENTS R. Smoot, J. Conneely, P. Serrano, D. Cavallucci and I. McGilvray University Of Toronto, Toronto, ON Introduction: The morbidity and mortality associated with liver resection has continued to decrease and venous reconstruction experience has continued to increase. This allows consideration of extended liver resection for tumors, which includes resection and reconstruction of the major hepatic venous outflow in selected patients. Methods: Patients undergoing liver resection with major hepatic venous reconstruction between the years 2005–2013 were reviewed. Results: For the study years a total of 18 patients were identified. Median age was 49.5 years (24–67) for 11 men and 7 women. The pathology included colorectal cancer © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

62

Abstracts

(7 patients), cholangiocarcinoma (6 patients), sarcoma (3 patients), hepatocellular carcinoma (1 patient), and perivascular epitheliod tumor (1 patient). Resection included the inferior vena cava in 8 patients, and 7 patients underwent reconstruction with ringed PTFE grafts. The venous outflow reconstruction included left hepatic vein (LHV) to graft (4 patients), LHV to cava (3 patients), Right hepatic vein (RHV) to graft (3 patients), RHV to cava (4 patients), middle hepatic vein (MHV) to cava (1 patients), RHV and MHV to cava (1 patient), MHV reconstruction with vein patch (1 patient), and LHV reconstruction with vein patch (1 patient). Perioperative mortality was 6% (1 patient) and a complication rate of 78% (14 patients). Six patients had died at the time of last followup, four of those patients with tumor recurrence. The cohort had a median overall survival of 15.4 months. Conclusion: In highly selected patients resection and reconstruction of the major venous outflow can be completed with reasonable morbidity and mortality.

OP-II.06 NOVEL RADIOFREQUENCY ABLATION SYSTEM INCORPORATING MAGNETIC TRACKING IN PORCINE MODEL M. Brown, M. A. Cassera, W. Johnston and C. W. Hammill Hepatobiliary And Pancreatic Surgery Program, Portland, OREGON Introduction: Ultrasound-guided laparoscopic radiofrequency ablation (RFA) requires advanced skills in hand-eye coordination and spatial positioning. The objective of this study was to assess whether a novel magnetic tracking device (MTD), providing real time three-dimensional guidance on a stereoscopic monitor, increases proficiency over ultrasound guidance alone. Methods: Fiducial markers were placed at three points in the liver of a live porcine model. Two surgeons targeted the markers using the MTD in conjunction with ultrasound, and under ultrasound guidance alone. The markers were targeted with the RFA needle in-line with the ultrasound, off axis 45-degrees from the ultrasound, and off axis 90-degrees from the ultrasound. Results: The times for each trial are summarized in Table 1. Discussion: Surgeon #1 showed a decreased time to target the marker with the MTD over ultrasound alone when the target was approached from out-of-plane, with a larger impact at 90 degrees. In addition, time with the MTD decreased as they became more familiar with it. Surgeon #2 showed a decrease in time taken to target the marker with the MTD when in-plane and 90 degrees out-of-plane, but did not show a clear benefit when 45 degrees out-of-plane. The use of the MTD did not show a clear benefit. However, for both surgeons, the use of the MTD decreased the amount of time it took to accurately target the marker when the RFA probe was 90 degrees out of plane with the ultrasound. This is classically the most difficult method used for targeting.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

OP-II.07 PATTERNS OF RECURRENCE FOLLOWING MARGIN-POSITIVE RESECTION OF COLORECTAL LIVER METASTASES C. W. Michalski, J. Wagner, B. Sheppard, S. L. Orloff, G. Vaccaro, C. Lopez and K. G. Billingsley Oregon Health And Science University, Portland, OR Introduction: Positive surgical margins remain a persistent problem in liver resection for colorectal liver metastases (CRLM). The aim of this study was to define the patterns of recurrent disease in patients with positive margins following hepatectomy. Methods: This is a retrospective analysis of prospectively collected data for hepatectomies for CRLM performed at our institution (2003–2012). The study population is limited to patients with microscopic positive margins. Survival is calculated using the Kaplan-Meier estimate. The study is approved by the OHSU IRB. Results: Out of 196 hepatectomies for CRLM, 21 patients with a positive resection margin were identified. Twenty of these patients had received neoadjuvant chemotherapy. The median number of liver metastases per case was 4 and the average size of the largest metastasis was 4 cm. Concomitant intraoperative radiofrequency ablation was performed in 6 cases. Median overall survival following liver resection was 39 months and median disease-free survival was 12 months. 4 patients recurred in the liver only. Of the liver only recurrences only one patient had disease limited to the site of resection. In 15 patients, either distant metastasis, liver and distant recurrence, or liver recurrence followed by shortly afterwards distant recurrence was found (one patient did not recur and one patient was lost to follow-up). Conclusion: Although surgeons must aim for negative margins, these results suggest that most patients with positive margins have unfavorable disease biology and their greatest risk is systemic recurrence. Additional therapy, such as radiation to the involved margin is unlikely to add to survival in patients with positive margins.

OP-II.08 THE SLIPPERY SLOPE OF CROSS CLAMPING VS OPEN SURGERY: IS THE CHOICE OBVIOUS? T. Dominioni, J. Viganò, L. Cobianchi, A. Peloso, J. Ferrario, A. D’addiego, I. Benzoni, P. Dionigi and M. Maestri Fondazione IRCCS Policlinico San Matteo, Pavia, PV Liver resection is the best chance of cure for primary and metastatic liver cancers. Extreme liver surgery has been proven feasible when the remaining parenchyma is sufficient to support metabolic needs and the stress of liver regeneration. Intermittent cross clamping is used to decrease the risk of excessive blood losses in liver surgery. 89 patients underwent major liver surgery between January 2009 and January 2012. Such cases were enrolled in a study to evaluate the effect of intermittent cross clamping (CC) of the liver hilum vs. open flow surgery (OF) . Patients where assigned to group 1 (n = 50, CC) and group 2 (n = 39, OF). Blood losses, hospital stay, postoperative liver function and complications were evaluated. Statistical analysis was performed by non parametric techniques. Survivals were compared by Kaplan HPB 2014, 16 (Suppl. 1), 1–99

Abstracts Meier curves and post-hoc tests. All the patients survived and were discharged within the 14th POD. Blood losses were lower in the CC group compared to OF (p < 0.05). However, the CC group showed a significant increase of LFTs after surgery, from the 1st POD to the 5th. While all cases recovered to normality, hypoxia was the key factor to determine this observation. Survival analysis did not demonstrate any significant difference between CC and OF. The onset of complications was remarkably similar in both groups. While survivals do not differ, the use of cross clamping should be carefully evaluated and its use reserved to cases with special technical difficulties.

OP-II.09 TREATMENT RESPONSE OF HEPATOCELLULAR CARCINOMA BY ETIOLOGY OF LIVER DISEASE R. M. Seshadri, D. J. Niemeyer, J. Lee, R. Z. Swan, D. Sindram, J. B. Martinie, M. Russo and D. A. Iannitti Carolinas Medical Center, Charlotte, NC Introduction: In the US, HCC is usually associated with Cirrhosis from either Alcohol; hepatitis C (HCV) or Nonalcoholic Fatty Liver Disease (NAFLD). Therapy while waiting for transplant includes transarterial chemoembolization (TACE), microwave ablation (MWA), radioembolization using Yttrium 90 and Sorafenib. Our aim was to determine if there is an association between etiology of cirrhosis and treatment response to MWA and TACE for HCC. Methods: Retrospective analysis was performed from our patient database and 332 patients were reviewed of which 118 of them were found to have HCC. 71 patients developed HCC from HCV and 21 patients from NAFLD. The first CT/MRI after treatment was reviewed using the RECIST (response evaluation criteria in solid tumors) criteria. We compared RECIST response, 1 year mortality, type of treatment and demographic data between the 2 groups. Conclusion: NAFLD patients with HCC had higher response rates to locoregional therapy but our sample size may have been too small to demonstrate statistical significance. The HCV group was younger compared to the NAFLD group and the NAFLD group was more likely to have high AFP. Further study is warranted because these findings may have implications of when to initiate screening and utilize AFP based upon etiology of liver disease.

63

drift in incidence and severity of post-operative complications arising from laparoscopic compared to open liver resection. Methods: Retrospective review of 615 patients managed by a single surgeon from 2000 to 2013. Analysis contrasts the demographics and post-operative outcomes between both groups. Complications were examined and compared using the Clavien-Dindo Classification for severity. Results: 753 hepatic resections in 615 patients were examined. Open surgery was reserved for smaller lesions (4.5 vs. 5.1 cm: p = 0.064) and predominately in malignant lesions (44.8 vs. 78.2%: p < 0.001). Complications were more frequent in open procedures (40.2 vs. 20.2%: p < 0.001) although the severity of complications was not significantly different (p = 0.846). Concordantly an open approach resulted in a higher percentage of patients requiring critical care (89.7 vs. 12.9%: p < 0.001) resulting in a longer hospitalization (8.5 vs. 3.4 days: p < 0.001). Conclusion: This study confirms laparoscopic hepatic resection carries a lower incidence of complication. Notably the severity of the complications is equivalent between the laparoscopic and open groups. The higher utilization of critical care services and prolonged hospitalization may also be attributed to an increased complexity of open resections not performed laparoscopically.

OP-II.11 AN ENHANCED RECOVERY IN LIVER SURGERY (ERILS) PATHWAY FOR MINIMALLY INVASIVE AND OPEN HEPATIC RESECTION DEMONSTRATES EARLIER RETURN TO NORMAL FUNCTION AND INTENDED ONCOLOGIC THERAPIES T. A. Aloia, A. Cooper, C. Conrad, J. Vauthey and V. Gottumukkala UT MD Anderson Cancer Center, Houston, TX

OP-II.10 REDUCUCTION OF COMPLICATIONS: THE PROGRESSION OF LAPAROSCOPIC LIVER RESECTION A. Hauch, S. Lee, A. Golden, E. Kane, E. Kandil and J. Buell Tulane Transplant Institute, New Orleans, LA Introduction: Laparoscopic hepatic resection has been described as safe and efficacious. This study examines the HPB 2014, 16 (Suppl. 1), 1–99

Background: Although multiple fast-track and enhanced recovery (ER) pathways are currently being evaluated in abdominal surgery, few have focused on the benefits of ER in liver surgery and no studies have evaluated the impact of ER on return to intended oncologic therapies. Patients/Methods: Short-term outcomes for patients managed with an ERILS pathway (preoperative education, non-narcotic premedication, epidural catheter for open approach, and postoperative multimodality non-narcotic analgesia) were prospectively recorded. Milestones including length of stay, time to return of baseline performance status and to intended oncologic therapies were compared to historical norms. Results: Between June-September 2013, 11 patients were treated with ERILS. The median age was 58 years (range: © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

64

Abstracts

42–74), and 7 (63.3%) were male. Indications included 3 HCC/Cholangio, 5 colorectal metastases and 3 other metastases with median tumor number of 1 (range:1–2) and median tumor size of 4.0 cm (range:1.3–8.5 cm). Resection approaches and procedures included 4 minimally invasive (MIS), 7 open and 5 major hepatectomies. The median lengths of stay (LOS) for the MIS and open groups were 2.3 days and 4.3 days, respectively, with one complication (urinary retention) and no readmissions. The median time to return to baseline function was 12 days and the median time to reinitiation of intended oncologic therapy (RIOT) was 24 days. These benchmarks compare favorably to historical MIS LOS (3.2 days), Open LOS (6.4 days), and RIOT times (34 days). Conclusion: This early experience with a narcotic-sparing ERILS pathway demonstrates safety, feasibility and encouraging reductions in time to return to normal activity and to intended oncologic therapies.

OP-II.12 COST EFFECTIVE ANALYSIS OF ALTRUS VERSUS CUSA FOR LIVER RESECTIONS G. R. Beck, K. P. Croome and R. Hernandez-Alejandro University Of Western Ontario, London Health Science Centre, London, ONTARIO Various techniques are available for hepatectomy parenchymal division including “clamp-crush”, electrocautery, ultrasonic or waterjet ablation, staplers and tissue sealing/fusion/ cutting devices from different energy sources. Each have advantages/disadvantages leading to equivocal best method for hepatectomy. With limited data, studies have failed to demonstrate benefit of these technologies over “clampcrush”. We report the first use of a novel tissue thermal fusion device ALTRUS(CONMED) in a retrospective comparison to our standard of CUSA for hepatectomy. From January 2010 to August 2013, single surgeon, 120 patients analyzed and after exclusions, ALTRUS (n = 49) and CUSA (n = 53) compared. Both groups were comparable with respect to age, gender, ASA, indication for resection (mainly CRC). The ALTRUS group had fewer segments resected (more parenchymal preservation), with similar number liver lesions in both groups. In the ALTRUS group, operating room time was significantly shorter (minutes, 220.5 ± 92.2 vs. 324.4 ± 113.5, p < 0.001). There was less calculated blood loss (cc, 376.1 ± 399.1 vs. 710.8 ± 663.2, p = 0.01), less transfusions intraoperatively (units PRBC, 0.34 ± 1.41 vs. 1.45 ± 2.50, p = 0.02) and total in hospital stay (units PRBC, 0.53 ± 1.46 vs. 2.34 ± 3.28, p = 0.004). Accounting for quicker OR time and equipment, cost was significantly less (CAN$, 5764.20 vs. 8066.20, p < 0.001) excluding surgical billings. Resection margin and hospital stay were similar. This retrospective comparison of the novel ALTRUS demonstrates it is faster, less expensive to the health care system, efficient and oncologically safe. It permits more parenchymal preservation. Given its novelty, other studies(including basic lab) may be warranted to formalize its safety and efficacy.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

OP-II.13 IMMUNOHISTOCHEMICAL TUMOR MARKERS IN THE DIAGNOSTICS OF FOCAL LIVER LESIONS J. Vilmanis, I. Strumfa, A. Ozolins, R. Stucka and J. Gardovskis Pauls Stradins Clinical University Hospital, Riga, RIGA Introduction: Correct diagnosis is mandatory for treatment of primary or secondary liver tumors. In most cases, biopsy is considered the gold standard. The aim of the present study was to evaluate the role of selected immunohistochemical tumor markers in the diagnostics of focal liver lesions. Materials and methods: Sequential patients presenting with focal liver lesions were enrolled in the study. The clinical history and radiological findings were reviewed. Biopsy and/or operation material were examined by routine histology followed by immunohistochemical visualization of alpha-fetoprotein (AFP), TTF-1, hepatocyte antigen (HepPar), CD10, cytokeratin (CK)7, CK20, CDX2, mammaglobin, chromogranin A and prostate specific antigen. Results: In total, 71 patients, aged 22–86 years (mean, 61.9) were enrolled. The final diagnosis of hepatocellular carcinoma (HCC) was established in 17 cases and of metastatic colorectal cancer (mCRC) – in 12 cases. In addition, metastatic lung cancer (3) and neuroendocrine tumours (2) were diagnosed. Metastatic pancreatic cancer (6) and cholangiocarcinoma (4) were suspected; however, no definite molecular proof was reached in these cases. The panel including HepPar, TTf-1 and AFP allowed diagnosing all cases of HCC by showing positivity of at least 1 marker. CDX2 in combination with expression of CK20 and negativity for CK7 was informative in all mCRC cases. Benign conditions were diagnosed in 9 cases. Conclusions: Tissue evaluation in conjunction with tumor marker analysis can provide useful diagnostic information. HCC can be reliably diagnosed with panel including HepPar/ TTF-1/AFP, but mCRC – by CDX2/CK20. Additional new markers are necessary to diagnose cholangiocarcinoma and pancreatic cancer.

OP-II.14 LAPAROSCOPIC MANAGEMENT OF CYSTIC DISEASE OF THE LIVER A. Hauch, S. Lee, E. Kane, E. Kandil and J. Buell Tulane Transplant Institute, Tulane University School Of Medicine, New Orleans, LA Introduction: Cystic disease of the liver can occur in isolation or in combination with polycystic disease of the kidney. The treatment of symptomatic cysts has evolved from an open surgical procedure to a laparoscopic approach. Methods: Retrospective review of 58 patients managed by a single surgeon from 2000 to 2013. Analysis was performed to compare and contrast with the previously reported Pittsburgh series. Results: Surgically resected patients underwent either subtotal cystectomy or a total resection for lesions with suspicious cyst linings. Median age was 61 years of age, BMI of 27, with 19% being obese (>35), with an ASA of 3. The median cyst size was 10.8 cm. In the 54 laparoscopic patients, 42% underwent major hepatic resection (formal lobe) with a median resection of 3 segments in the full HPB 2014, 16 (Suppl. 1), 1–99

Abstracts cohort. Resection required hand-assist in 60% of cases with a median operative time of 120 min, 100 cc blood loss and 12 staple loads. Five cystadenomas were encountered. Four cases early in the cohorts experience were approached via an open technique for suspected cystadenoma and complexity. Incidence of complications was 11% (bile leak n = 3, pulmonary n = 2, ileus n = 2). Conclusion: This experience confirms laparoscopic resection is the gold standard for the management of hepatic cysts. In our experience open resection was rapidly abandoned for laparoscopy. An early incidence of bile leak was identified and addressed by deployment of staplers for cyst wall transection. Both series confirm a laparoscopic approach can be utilized with a low incidence of complications and recurrence.

65

ment. We hope that by presenting these cases, we have reinforced the need to consider surgical shunting in appropriately selected patients.

OP-II.18 EPIDEMIOLOGY OF METASTATIC COLON CANCER AMONG SAUDIS F. Al-alem1, E. Simoneau2, T. A. Traiki1, N. Al-malki1, K. Alkhayal1, A. Zubaidi1, O. Obaid1 and M. Hassanain1,3 1 Department Of Surgery , King Saud University, Riyadh, CENTRAL; 2Department Of Surgery , McGill University Health Center, Montreal, QUEBEC; 3Department Of Oncology , McGill University Health Center, Montreal, QUEBEC

OP-II.16 WHEN SURGICAL SHUNTING IS THE BEST OR ONLY OPTION: A CASE SERIES L. M. Louwers1, V. Raofi2, S. Cohn2 and A. Koffron2 Henry Ford Health System, Detroit, MICHIGAN; 2 Beaumont Health System/Oakland University School Of Medicine, Royal Oak, MICHIGAN

1

Background: Surgical portosystemic shunting (PSS) has become a less frequently used modality for control of portal hypertension (PHTN) in the era of transjugular intrahepatic portosystemic shunting (TIPS). In certain instances, however, TIPS may not be anatomically feasible or the best long-term management strategy. Objectives: To present a series of patients highlighting scenarios where surgical PSS may have a more favorable outcome compared to TIPS placement or medical management, reinforcing the role of surgical shunting in the postTIPS era. Methods: A retrospective review of surgical PSS performed at a tertiary care center from 2009–2011. Results: Six cases were selected to highlight circumstances wherein surgical shunting was the preferred or only available modality in complicated benign disease states. The underlying etiologies of PHTN included nodular regenerative hyperplasia, polycystic liver disease, chronic pancreatitis, non-alcoholic steatohepatitis, and alcoholic pancreatitis with left-sided PHTN. Detailed preoperative and operative descriptions provide enhanced understanding of the benefits of PSS in these situations. Conclusion: In this series, we highlight two groups where surgical shunting may still be the preferred, and in some cases the only approach. The first group is patients who will need long term patency of their shunt given the fact that they are either well compensated cirrhotics or have non-cirrhotic portal hypertension. The second, and probably more important, are the patients with underlying aberrant anatomy which precludes or cannot be completely addressed by TIPS placeHPB 2014, 16 (Suppl. 1), 1–99

Introduction: The incidence of colorectal cancer in Saudi Arabia is rising, being the most common cancer in the male population. However, little is known about the epidemiology of metastatic colon cancer. The country has a population of over 26 M people with a median age and life expectancy of 25 and 75 years old respectively. Our study aim is to present the incidence and management patterns of metastatic colorectal cancer in Saudi Arabia. Methods: A retrospective analysis of prospectively collected data was done using the Saudi Cancer Registry (SCR) for colorectal cancer with focus on our representative tertiary care center, between 2006–2009. Results: In 2008, SCR reported 904 cases with male predominance. The median age was 59 , 29.2% of patients had synchronous disease. At our tertiary center, there were 191 new colon cancer cases with male predominance (57%) and a mean age at diagnosis of 58. Median follow up was 3 years. Rectal cancer was the primary site in 33% of cases. 64 (33%) patients had synchronous disease 39(20%) were hepatic and 23 (12%) were lung metastasis. Total 15 follow up patients had metachronous disease 7% hepatic and 4% was lung. Out of the total liver metastasis 20% were resected. Conclusion: Metastatic colon cancer is a considerable problem occurring in a relatively young population. More effective screening, possibly at an earlier age with population-specific guidelines, and better treatment strategies are needed to improve this important public health burden.

OP-II.19 MALIGNANT MASQUERADE OF PET-CT AVID HEPATIC ADENOMA F. Alemi, T. Mohammed, P. Sicuro, K. Kowdley, A. Siddique, A. Alseidi, T. Biehl, S. Helton and F. Rocha Virginia Mason Medical Center , Seattle, WA Background: Positron emission tomography-computed tomography (PET-CT) can be a useful adjunct imaging © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

66

Abstracts

modality to help differentiate malignant from benign liver lesions. Known false positive lesions on PET imaging include focal steatosis or hepatic abscesses. However, eight cases of PET-avid hepatic adenomas have been published in the literature. We describe our experience with this condition and report our outcomes. Methods: All patients who underwent evaluation for hepatic malignancy from 1995-present with PET imaging were reviewed focusing on patient presentation, imaging, pathology and outcome. Results: Two patients were identified with PET-avid hepatic adenomas (see Figure). A 53 year-old woman with a history of breast cancer and BRCA2 mutation developed a 2 cm mass in segment 5 on surveillance imaging. PET-CT scan showed a lesion with a standardized uptake value (SUV) of 10.7. Subsequent percutaneous biopsy revealed a hepatic adenoma (negative beta-catenin staining). A 49 year-old woman who developed abdominal pain was found to have a 4 cm juxtacaval lesion in segment 7 with liver atrophy. The lesion was indeterminate with standard imaging and PET-CT scan showed a lesion with an SUV of 4.9. Given concern for malignancy, she underwent right hepatectomy and pathology was consistent with well-differentiated hepatic adenoma with negative beta-catenin staining but positive HNF1alpha mutation. Both patients are doing well with no progression or recurrence. Conclusion: Despite its rarity, hepatic adenomas may demonstrate PET-CT activity. Clinicians should consider this benign etiology in young women with PET-avid liver lesions and pursue further diagnostic tests if indicated prior to intervention.

from 2004–2011. Chart review included comorbidities, hospital course, trends in serum lactate levels, APACHE scores, MELD scores (Table 1). Ten cirrhotics with PI were identified. Nine were managed conservatively. One patient underwent exploration and died. Results: In total, in-hospital mortality was 70%. The mean MELD in patients who died was higher than in those who survived (28.6 vs. 14.0). Mortality was 100% in cirrhotic patients with PI whose MELD was greater than 16. Mean serum lactate levels (91.4 mg/dL vs 20.8 mg/dL) and mean APACHE scores (28.3 vs. 15.3) were also higher in patients who died. Serum white blood cell count and bicarbonate were not consistently abnormal in those who died. Conclusion: Our results suggest that MELD and APACHE scores and lactate levels are important predictors of mortality in cirrhotics with PI. Given the high risk of surgical mortality and the natural course of PI in our population, surgical intervention may be futile in cirrhotics with a MELD score greater than 16. Larger studies are required to determine whether there is a role for surgery in cirrhotics with lower MELD scores.

OP-II.23 PANCREATICODUODENECTOMY IN THE OLDEST OF OLD: RESULTS BEYOND OPERATIVE MORTALITY

OP-II.20 MELD PREDICTS MORTALITY IN CONSERVATIVELY MANAGED CIRRHOTICS WITH PNEUMATOSIS F. Karipineni, J. Pang, A. Parsikia and J. Ortiz Albert Einstein Medical Center, Philadelphia, PA Introduction: Pneumatosis intestinalis (PI) is a radiologic finding associated with multiple conditions and a high mortality rate. This is the first study addressing the management of PI in cirrhotics. While conservative treatment is an accepted approach to PI in the general population, it is unknown whether the same is true when treating cirrhotics. Methods: We retrospectively identified cases of PI found on computed tomography (CT) scans performed on cirrhotics © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

R. C. Langan1, P. R. Bucciarelli1, W. Chapman1, K. Oza1, C. Fehring1, P. G. Jackson1, T. M. Fishbein1, R. Jha2, N. Haddad2, F. Al-Kawas2, J. Carroll2, J. Hanna1, A. Parker1, W. B. Al-Refaie1 and L. B. Johnson1 1 Department Of Surgery: Georgetown University Hospital, Washington, DC; 2Department Of Gastroenterology: Georgetown University Hospital, Washington, DC Introduction: The projected expansion of the oldest old (≥80 years) will have significant implications on patients, providers, hospital service lines and payers. Methods: Perioperative outcome measures were compared between 33 patients (≥80 years) and a non-randomized cohort of 33 patients less than 70 years who underwent pancreaticoduodenectomy between 2005 and 2013. Results: Median age was 62 years (range, 34–70) and 83 years (range, 80–87) respectively. Pathology (young vs. HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

67

elderly) included pancreatic ductal adenocarcinoma (21/33 vs. 20/33), duodenal carcinoma (0/33 vs. 2/33), ampullary carcinoma (2/33 vs. 3/33), cholangiocarcinoma (1/33 vs. 2/33), pancreatic neuroendocrine tumor (2/33 vs. 2/33) and benign lesions (7/33 vs. 4/33). Median tumor size (young vs. elderly) was 2.8 cm and 3.25 cm, respectively. The mean estimated blood loss in the young vs. elderly group was 485 mL vs. 567 mL (p = 0.42). There were no differences in 30-day operative mortality, overall perioperative complications (12/33 vs. 15/33) or 60-day re-admission rates (8/33 vs. 8/33). However, the oldest old experienced longer hospital stays than their younger counter parts (13.7 vs. 8.4 days, p = 0.007) and were more likely to be discharged to a skilled nursing facility or sub-acute rehabilitation compared to home (0/33 vs. 9/33, p = 0.002). Conclusion: Despite their comparable operative mortality and adverse events, the oldest old experienced higher rates of prolonged length of stay and discharge to a skilled facility. These results have future research and policy implications to overcome some of the challenges of this growing cohort of the US population.

OP-II.24 ANALYSIS OF PATENCY RATES IN PORTOMESENTERIC VENOUS RECONSTRUCTION DURING PANCREATIC RESECTION M. Brown Providence Portland Medical Center, Portland, OREGON Introduction: In this study, we sought to evaluate our experience with portomesenteric venous reconstruction (PMVR) and long-term durability. Methods: From May 2007 to June 2012, 33 cases were identified in which PMVR’s was performed at out Institution. Our analysis focused on long-term patency, interval between reconstruction and thrombosis or significant stenosis, requirement of salvage procedure. Results: 33 patients with en bloc pancreatic resections also, required PMVR. 24 (72.3%) patients had pancreatic adenocarcinoma, 4 (12.1%) neuroendocrine tumor and other neoplasms. Reconstructive techniques included lateral venorrhaphy (14) (42%), primary end-to-end anastomosis (11) (33.3%), venous interposition graft (Left internal jugular vein harvest) (2) (6%), Bovine pericardial patch (6) (18.2%). A vascular surgeon assisted with 7 of the 33 cases. Outcomes focused on long-term patency, interval between reconstruction and thrombosis or significant stenosis, requirement of salvage procedure. Early <3 months, intermediate 3–12 months and late >12 month patency rates were analyzed, which are 87.8%, 75%, 42.4% respectively. 13 (39%) deaths occurred over the duration of our study. Discussion: Results of this study are consistent with most recent published literature on PMVR with respect to patency rates and survival. Our data suggests that early thrombosis/ stenosis which ultimately resolve is less likely due to technical error.

HPB 2014, 16 (Suppl. 1), 1–99

OP-II.25 TIME TO ADJUVANT THERAPY FOR LAPAROSCOPIC PANCREATIC RESECTION T. Sachs, M. Cooper, T. Pawlik, M. Choti, L. Zheng, J. Herman, D. Laheru, J. Cameron, C. Wolfgang, M. Makary and M. Weiss JOHNS HOPKINS MEDICAL INSTITUTION, Baltimore, MARYLAND Background: The laparoscopic approach is being increasingly applied to pancreas surgery, however the oncologic advantage remains undefined. Methods: We performed a retrospective analysis on patients who underwent upfront pancreatic resection for pancreatic cancer (2009–2012). We compared demographics, perioperative variables, time-to-adjuvant therapy (median ± IQR) and survival between open (O) and laparoscopic (L) surgery goups. Results: We identified 226 patients (L:25 laparoscopic, O:201). There were no differences in demographics or comorbidities between groups. The majority underwent pancreatoduodenectomy (L:76%, O:80%; P = .64), and were stage IIB upon resection (L:64%, O:73%; P = .28). Margin positivity was less, but not significant (L:8%, O:22.4%; P = .09). Only 62% of patients ultimately received adjuvant therapy (L:56%, O:63%; P = .56). Median time-to-adjuvant therapy was less, but not significant (L:45 ± 32, O:54 ± 19 © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

68

Abstracts

days; P = .81). When comparing more recent patients, (after 2011), this gap widened (L:39 ± 37, O:55 ± 22; P = .47). More laparoscopic patients received adjuvant therapy within 6 weeks of surgery (L:72%, O:52%; P = .05). Within the study period, 37.2% of patients died (L:44%, O:36.3%; P = .45), with similar median survival (L:288 ± 236, O:314 ± 319 days; P = .88). Open operations had far greater overall complications (L:24%, O:70%; P = .01), and major wound complications (L:4%, O:24%; P = .02). Among patients with no complications, median time-to-adjuvant therapy (L:43 ± 29, O:50 ± 16 days; P = .27) was less, but did not achieve significance. Conclusion: Laparoscopic patients have fewer wound complications, and less overall morbidity. While gains are being made in time-to-adjuvant therapy, more patients and a broader understanding of the advantages of laparoscopy within the medical community are necessary if we are to attain oncologic benefits.

OP-II.26 IMPACT OF MULTIDISCIPLINARY HEPATO-PANCREATO-BILIARY CASE CONFERENCE ON THE MANAGEMENT OF PANCREATIC CANCER K. Mayer1,2, Y. Kuo2,3, J. M. Davis1,2, A. Adams1, O. Cabasso1,3, J. J. Vernick1,2 and E. J. Hagopian1,2 1 Department Of Surgery, Jersey Shore University Medical Center , Neptune, NJ; 2Department of Surgery, Robert Wood Johnson Medical School, New Brunswick, NJ; 3 Office Of Academic Affairs, Jersey Shore University Medical Center, Neptune, NJ Background/Aim: Prospective case discussion allows for the dissemination of the best practices in patient care. The aim of this study was to determine if the initiation of a monthly Multidisciplinary Hepato-Pancreato-Biliary (HPB) Case Conference has influenced the approach of the initial method of tissue diagnosis of pancreatic cancer at our institution. Methods: All patients initially diagnosed with pancreatic cancer at our institution between 2007–2012 were queried from the Tumor Registry. Patient charts were reviewed for demographic and clinical information. Patients were classified as resectable, locally advanced/unresectable, or metastatic at the time of presentation. The method of initial tissue diagnosis was determined. The HPB Case Conference log © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

was queried for the number and type of cases presented. Data was analyzed using Chi-square. Results: Since 2010, a total of 116 cases, including 42 pancreatic cases, have been presented during the Multidisciplinary HPB Case Conference. A total of 171 pancreatic cancer cases, including 91 female and 80 male, diagnosed between 2007–2012 were identified from the Tumor Registry. When comparing the three-year time frame prior to (2007–2009) and after (2010–2012) the initiation of the HPB conference, there was a significant increase in the use of a surgical or EUS approach for resectable pancreatic cancer (P = 0.03), EUS approach for locally advanced/unresectable (P = 0.03), and metastatic biopsy site approach for metastatic (P = 0.02) as compared to other methods as the initial method of tissue diagnosis. Conclusion: The initiation of the Multidisciplinary HPB Case Conference has positively influenced patient care and physician practices at our institution.

OP-II.28 IS THERE A JULY EFFECT FOR HPB CASES? S. Patel, S. Cheek, T. Winston, H. Osman and D. Jeyarajah Methodist Dallas, Dallas, TX Background: The presence of a “July Effect” has an increased propensity to negatively affect patient care have been widely debated. However, there are no studies looking at the July phenomenon related to surgical HPB fellowship. This study aims to evaluate if there is a “July Effect” in pancreatic cases in HPB fellowship. Methods: Retrospective analysis from 2006–2012 was conducted for pancreaticoduodenectomies (PD) and laparoscopic distal pancreatectomies (LDP) performed in July/August (group A) and April/May (group B). Intraoperative and postoperative parameters were compared. Variables were compared using 2 sample t test, WilcoxonMann-Whitney test, and Fisher exact test. Results: 93 Pancreaticoduodenectomies (PD) and 18 Laparoscopic distal pancreatectomies (LDP) were conducted during the time frames stipulated. See Table. Conclusion: This is the first study to show a “July effect” in PD with regards to operative time. However, there was no difference in oncologic parameters examined, validating the lack of impact on quality of care. LDP, on the other hand, did not show such an effect. One possibility is that there is a steeper learning curve in LDP. HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

69

OP-II.30 GROOVE PANCREATITIS: A RARE MIMICKER OF PANCREATIC CANCER WHICH CAN BE TREATED WITH PANCREATICODUODENECTOMY S. Rabi, J. Cameron, M. Makary, K. Hirose, N. Rezaee, T. Pawlik, R. Hruban, C. Wolfgang and M. Weiss The Johns Hopkins Hospital, Baltimore, MD

OP-II.29 THE USE AND UTILITY OF ABDOMINAL CT SCANS IN THE IMMEDIATE POSTOPERATIVE PERIOD FOLLOWING PANCREATIC RESECTIONS S. Zenoni1, P. Veldhuis 1, S. Eubanks1, S. G. De La Fuente1,2 and P. Arnoletti1,2 1 Florida Hospital Orlando, Orlando, FL; 2University Of Central Florida, Orlando, FL Background: Pancreatic resections are associated with significant morbidity and re-admission rates. Abdominal computerized tomography (CT) scans are often obtained in the immediate postoperative period to rule out potential complications. We sought to analyze the frequency and indications of postoperative abdominal CT scans among pancreatectomy patients in an attempt to elucidate the true impact of such diagnostic imaging studies on clinical management. Methods: A prospectively maintained database of all consecutive patients undergoing pancreatic resections during a 10-month period was reviewed. Patients that had undergone a CT scan during the immediate 30-days postoperative period were included. Outcome variables included type of CT scan performed, indication for CT scan, resultant modification in clinical care (drain placement, dietary modifications, antibiotic therapy, surgical re-intervention), length of hospital stay and readmission rates. Results: A total of 62 patients underwent pancreatic resections during the study period. Median length of hospital stay for that cohort was 10 days. Hospital re-admission rate was 30%. Postoperative abdominal CT imaging was obtained in 48% (n = 30) patients at an average time interval of 11.3 days following surgery. Postoperative CT imaging led to insertion of a percutaneous drain in 11% of patients and surgical re-intervention in 5% of patients. There were 76% of patients with unremarkable CT findings. Conclusion: Postoperative CT scans seldom result in clinically significant interventions and are over-utilized following pancreatic resections. Abdominal imaging in pancreatectomy patients should be selectively employed to avoid the increased cost and potential harmful consequences of repeated radiographic studies. HPB 2014, 16 (Suppl. 1), 1–99

Background: Groove pancreatitis (GP) is a rare condition that occurs in the groove region between the head of pancreas, the duodenum and the common bile duct (CBD) The diagnostic features and best treatment remain unknown. Our goal was to identify clinical features that may differentiate GP from malignancy and evaluate long-term outcomes of pancreaticoduodenectomy (PD) for GP. Methods: We reviewed all patients who underwent PD at our institution from 2005–2013 and identified those with GP on pathology. The demographic features, diagnostic findings, surgical complications and long-term outcomes were characterized. Results: Among 1424 patients who underwent PD, 14 (1%) were diagnosed with GP on pathology. The majority (71%) were male. Six (43%) abused alcohol, 4 (29%) tobacco and 3 (21%) both. The most common symptoms were abdominal pain (64%), decreased appetite (50%), and weight loss (50%). Nine (64%) patients had a clinical history or radiological features of pancreatitis upon presentation. Imaging demonstrated a cystic pancreatic head mass in 8 (57%), focal enlargement of the pancreatic duct and CBD in 8 (57%) and 5 (36%), respectively. The overall complication rate was 57.1% including two (14.3%) 30-day mortalities. At a median follow-up of 14.2 months, the remaining patients are asymptomatic and without radiologic evidence of recurrence. Conclusion: GP is a rare pathologic finding. We identified no reliable pre-operative features to distinguish GP from malignancy. In the light of the uncertain diagnosis, symptomatic presentation, and inability to reliably distinguish GP from pancreatic cancer, we consider PD to be an acceptable treatment of GP.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

70

Abstracts

OP-II.33 PERIOPERATIVE MORBIDITY AND MORTALITY AFTER PANCREATECTOMY: A SYSTEMATIC REVIEW OF PREDICTION SCORES, MODELS AND NOMOGRAMS 1

2

2

J. Aubin , R. Lapointe , F. Vandenbroucke-Menu and G. Martel1 1 Liver And Pancreas Unit, Division Of General Surgery, The Ottawa Hospital, University Of Ottawa, Ottawa, ON; 2 HPB Surgery And Liver Transplantation Unit, Hal Saint-Luc, CHUM, Montreal, QC Pancreatic resection is associated with a high rate of postoperative morbidity and mortality. Numerous scoring systems and predictive models have been developed to forecast the postoperative course of individual patients. A systematic review of the literature pertaining to perioperative morbidity and mortality risk prediction scores after pancreatic resection was carried out. A review protocol was utilized. Papers presenting or utilizing prediction scores, models and nomograms were retrieved. The characteristics of each score were reviewed, as were their validation history. 788 de-duplicated studies were screened. 58 studies fit the criteria for inclusion, yielding 44 different scores or prognostic models. Among these scores, 54.5% of papers presented novel scores that were not further evaluated or utlized in the literature. 45.5% of scores were pre-existing and being validated for or applied to pancreatectomy, such that, notably, POSSUM was presented in 32.8% of papers, and P-POSSUM and E-PASS in 6.9% of papers, each. Identified scores predicted several different types of outcomes, including mortality, overall morbidity, and specific morbidity such as pancreatic fistulas and post-operative diabetes. 38.6% of scores are validated for pancreatectomy; 30% of pre-established scores and 45.8% of novel scores. This work identified 44 different scores to predict post-operative morbidity and mortality after pancreatectomy. Many scores were generic, while others pertained to specific populations or outcomes. The overall degree of validation was poor. Further prospective studies are needed to validate the remaining scores. The usefulness of individual scores and their ability to predict an outcome of interest was highly variable.

OP-II.34 THE IMPACT OF ABERRANT ANATOMY OF THE HEPATIC ARTERY ON RESECTION MARGIN STATUS IN PANCREATICODUODENECTOMY P. Casey, R. De Souza, S. Lapsia, Z. Ali, K. Elhadd, A. Kausar, D. Chang, A. Krige, J. Thomas and D. Subar Department Of Hepatopancreatobiliary Surgery, Blackburn, LANCASHIRE Several studies have demonstrated that resection status (R) affects survival after resection of pancreatic adenocarcinoma. This study aims to assess the association between the presence of abnormal arterial anatomy (namely, a replaced or aberrant right hepatic artery) and resection margin status. Between 2007 and 2012 we analysed the pathology and pre-operative cross sectional images of all patients undergoing Pancreaticoduodenectomy for pancreatic adenocarcinoma. Aberrant arterial anatomy was classified by a single Radiologist as per Michels classification. Statistical analysis was conducted using SpSS (IBM) statistical program. Onehundred and fourty eight (148) pancreaticoduodenectomies were performed of which 75 were for adenocarcinoma of the head of the pancreas. The male to female ratio was 45 : 30 with a mean age of 65. Fifty-nine patients had a positive resection margins with 16 having clear margins. The medial (SMV) margin was the most commonly involved. Fifteen patients (21%) had aberrant arterial anatomy of the hepatic artery course. The presence of aberrant arterial anatomy did not have a statistically significant impact on the R Status (p = 0.67), involvement of the medial (p = 0.34), or Transection margin (p = 0.22). In concordance with other studies, R status correlated with overall survival (p = 0.02) suggesting that the presence of arterial variation does not affect survival in adenocarcinoma of the pancreas. There appears to be no increased risk of a positive resection margin in pancreaticoduodenectomy in the presence of variant arterial anatomy of the right hepatic artery. Identification of vascular anomalies by pre-operative imaging will reduce the risk of intraoperative injury.

OP-II.35 DIFFERENCES IN OUTCOMES IN PANCREATICODUODENECTOMY (PD) IN THE MODERN ERA: DOES A MALIGNANT DIAGNOSIS MAKE A DIFFERENCE? S. Cheek, S. Patel, H. Osman and D. Jeyarajah Methodist Dallas Medical Center, Dallas, TX Objective: Historically, there has been less enthusiasm for doing PD for benign conditions than malignant disease because of high morbidity and mortality rates. With improved outcomes PD is more commonly being preformed for benign indication, including pancreatitis and premalignant lesions. The aim of this study is to evaluate the number of PD done for benign disease and the differences in outcomes between pancreatic resection for malignant versus benign disease. Method: 321 patients who underwent PD between September 2005-May 2013 were identified retrospectively from our database. Diagnoses were divided into malignant and benign conditions. Malignant conditions include adenocarcinoma, © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts neuroendocrine tumors, and metastatic disease. IPMN, pancreatitis, and adenoma constitute 75% of benign condition. Cases were analyzed for operative time, operative blood loss, length of stay (LOS) and complication. Results: 226 (70%) had malignant disease (group A) and 95 (30%) patients had benign disease (group B). There was no statistically significant difference in median operative time, blood loss, and length of stay (LOS) between malignant and benign group. Conclusion: Operative time, blood loss, and LOS, were not statistically significant between the two groups. PD can be performed at a tertiary referral center for patients with both benign and malignant conditions with excellent and comparable outcomes.

OP-II.36 THE BENEFITS OF LAPAROSCOPIC PANCREATECTOMY: IT DOES MAKE A DIFFERENCE S. Cheek, S. Patel, H. Osman and D. Jeyarajah Methodist Dallas Medical Center, Dallas , TX Objective: Laparoscopic approach for distal pancreatecomy has been increasing recently. The aim of this study is to evaluate the operative and oncologic measures of laparoscopic distal pancreatectomy (DP) at a non-university setting. Methods: 169 patients who underwent a distal pancreatectomy at our tertiary referral center from September 2005May 2013 were identified from an institutional database. These cases were evaluated based on method of resection, pathologic diagnosis, intraoperative blood loss, OR time, and LOS. In malignant cases, margin of resection weas evaluated. Results: A total of 169 distal pancreatectomies were performed. An open resection (group A) was done 56.8% (96) of the time and a laparoscopic resection (Group B) was done 43.2% (73). Group B had statistically significant less operative blood loss and shorter length of stay. There was a higher rate of negative margins in patients with malignant diagnosis in group B (p < 0.0001). (Table 1) Conclusion: This study demonstrates that LDP can result in decreased blood loss and LOS compared to open resection. Moreover, in carefully selected patients, excellent oncologic parameters can be achieved.

HPB 2014, 16 (Suppl. 1), 1–99

71

OP-II.37 LONG-TERM SURVIVAL AFTER RESECTION OF PANCREATIC CANCER: SIGNIFICANCE OF EARLY DETECTION AND HISTOLOGICALLY CURATIVE RESECTION T. Yamamoto, H. Kinoshita, Y. Sakamoto, K. Okada, K. Inoguchi, S. Yao, S. Kaihara, R. Hosotani and S. Yagi Kobe City Medical Center General Hospital, Kobe, HYOGO Background: The prognosis of patients with pancreatic cancer remains poor. The number of long-term survivors, however, has been increasing. In this study, we retrospectively analyzed patients with pancreatic cancer who received pancreatic resection and assessed the factors that affected long-term survival and the clinicopathological characteristics of long-term survivors. Methods: From January 2000 to December 2011, 195 patients underwent pancreatic resection in our hospital. We first analyzed the prognostic factors of pancreatic cancer after pancreatic resection in these patients, of whom 96 patients were observed for more than 5 years. We analyzed the clinicopathological characteristics of patients who survived for 5 years or more and compared them to those of the patients who died within 5 years. Results: Using multivariate analysis, tumor size ≤ 20 mm (hazard ratio [HR] 0.38; confidence interval [CI] 0.16–0.76) and negative surgical margin (R0 resection) (HR 0.46; CI 0.29–0.73) were independent favorable prognostic factors. Comparison of 5-year survivors (n = 20) to non-survivors (n = 76) revealed that lower levels of pancreatic cancerassociated antigen (DUPAN-2) (1318 ± 4706 vs. 1655 ± 4210), tumor size ≤ 20 mm (35% vs. 8%), R0 resection (95% vs. 61%), and no lymph node metastases (60% vs. 18%), were significantly associated with 5-year survival. Conclusions: Negative surgical margin and tumor size ≤20 mm were independent favorable prognostic factors. Histologically curative resection and early tumor detection are important in treating pancreatic cancer and achieving longterm survival.

OP-II.38 SOCIOECONOMIC STATUS AND SURGICAL OUTCOMES AFTER TOTAL PANCREATECTOMY H. Schmidt, E. M. Kilbane, A. P. Turner, M. G. House, N. J. Zyromski, A. Nakeeb, C. M. Schmidt and E. P. Ceppa Indiana University School Of Medicine, Indianapolis, IN Background: A paucity of data exists between socioeconomic status (SES) and surgical outcomes. Patients deemed “poor” are presumed to suffer from higher rates of postoperative morbidity. The aim of this study was to determine if the level of Median Household Income (MHI) correlated with increased morbidity following total pancreatectomy (TP). Methods: Consecutive cases of TP (n = 21) were reviewed at a single high-volume institution between 2008 and 2012. Annual MHI was derived from the patient’s zip code as determined by U.S. Census data; patients groups were divided by high or low SES being above or below $50 K. All cases were monitored with complete 30-day outcomes © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

72

Abstracts

through the American College of Surgeons-National Surgical Quality Improvement Program. Electronic medical records were used for 90-day outcomes. Two-way statistical analyses were performed between the patient categories. Results: The high SES group (n = 10) as compared to the low SES group (n = 11) had a significantly higher MHI (p < 0.001). No significant differences existed in preoperative risk, intraoperative variables, surgical pathology, nor in either 30-day and 90-day readmissions or mortality. The lower SES group had a higher rate of total postoperative morbidity at 30-day (p < 0.01); no difference was seen at 90-day morbidity (p = 0.09). Conclusion: Low SES imparts a higher short-term morbidity to patients undergoing TP, although the long-term disadvantage is not evident. This may suggest that lower SES is not a reliable criterion when assessing surgical risk for this patient population. Further study of this topic is warranted with a larger patient sample and for regional pancreatectomy.

OP-II.39 OUTCOMES IN OPEN VERSUS LAPAROSCOPIC DISTAL PANCREATECTOMY R. Pieretti- Vanmarcke1,3, J. Wargo1,3, C. Fernandez-Del Castillo1,3, H. Zheng2,3, A. Merrill1,3, S. Thayer1,3, K. Lillemoe1,3 and C. R. Ferrone1,3 1 Department Of Surgery, Massachusetts General Hospital, Boston, MA; 2Department Of Biostatistics, Massachusetts General Hospital, Boston, MA; 3Harvard Medical School, Boston, MA Objective: To compare the outcomes of patients undergoing either open (ODP) or laparoscopic (LDP) resection of the distal pancreas. Methods: A single institution retrospective analysis of 429 patients undergoing distal pancreatectomies performed between 1/2004–12/2012. Demographics, tumor characteristics, operative technique and long-term outcomes were examined. Results: A total of 429 patients underwent distal pancreatectomy of whom 43% were male with a median age of 60. Patients underwent LDP (n = 104) or ODP (n = 325). 18 patients (17%) were converted to an open procedure. Operative time was significantly longer for LDP than ODP (p = 0.0016). The most common pathologic diagnoses were neuroendocrine tumor (21%), ductal adenocarcinoma (18%), mucinous cystic neoplasm/adenoma (9%) and intraductal papillary mucinous tumor (8%). Median tumor diameter was 3.1 cm and 3.0 cm, respectively. BMI was significantly higher in patients who underwent LDP than ODP (p = 0.0048). Length of stay was longer in patients who had an ODP (5.5 days, stdev = 2.81 vs 6.98 days, stdev = 5.61) (p = 0.0003). Total number of complications was also higher in patients who had an ODP (p = 0.016). Positive margins were higher in ODP (p = 0.05). © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

Conclusion: In the era of cost containment laparoscopic distal pancreatectomy is a safe operation which offers a shorter LOS and lower morbidity than ODP, despite a longer operative time.

OP-II.40 OUTCOME AFTER REVISION OF THE PANCREATICOJEJUNOSTOMY: A SINGLE INSTITUTION EXPERIENCE A. I. Gagnon, R. Birkett and T. Schnelldorfer Lahey Clinic, Burlington, MA Introduction: Long-term complications derived from the pancreatic anastomosis after pancreatoduodenectomy are fortunately infrequent. When this type of complication occurs operative revision is commonly needed, which can represent a significant challenge to the surgeon. Methods: The records of all patients who underwent a pancreatic operation at our institution between 2003 and 2012 were retrospectively reviewed and analyzed. Results: Three patients originally underwent pancreatoduodenectomy with a final diagnosis of inflammatory mass in two and IPMN with low grade dysplasia in one. The median age was 49 years. The indication for operation was anastomotic stricture in one and chronic pancreaticocutaneous fistula in two. Two were revised to pancreaticogastrostomy and one underwent revision to pancreaticojejunostomy. Peri-operative morbidity occurred in one patient without any mortality. Median estimated operative blood loss was 250 mL. There were no reports of major vascular injury, such as portal vein injury, during the operation. Median length of hospital stay was 7 days (range 5 to 9 days). Median follow-up was 1 year (range 0 to 4 years). At last follow-up, all patients had complete clinical resolution of their symptoms without any evidence of pancreatic duct obstruction. Conclusion: Revision to pancreaticojejunostomy seems preferable, but revision to pancreaticogastrostomy should be considered for patients at risk for ductal strictures from chronic pancreatitis, thus allowing endoscopic access. Operative revision of the pancreatic anastomosis, while technically challenging, is overall tolerated well. The operation can be performed with minimal blood loss, low perioperative morbidity, and long lasting patency of the anastomosis in suitably selected patients.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

OP-II.41 INCREASING INSTITUTIONAL EXPERIENCE AND EXPERTISE IMPROVES DIAGNOSTIC YIELD OF ENDOSCOPIC ULTRASOUND (EUS) IN SOLID PANCREATIC TUMORS S. D. Kachare1,2, A. G. Kaul1, T. L. Fitzgerald1,2,3 and E. E. Zervos1,2,3 1 East Carolina University, Brody School Of Medicine, Greenville, NORTH CAROLINA; 2Vidant Medical Center, Greenville, NORTH CAROLINA; 3Leo W. Jenkins Cancer Center, Greenville, NORTH CAROLINA Background: The diagnostic yield of EUS is multifactorial including: the endoscopist’s /cytopathologist’s expertise and underlying patient and tumor characteristics. The purpose of this study was to determine the diagnostic yield of repeat attempts at EUS for solid pancreatic neoplasms while controlling for patient and tumor factors. Methods: A prospective database of patients undergoing repeat EUS for solid pancreatic neoplasms was queried to determine its impact on clinical course decision making. All secondary EUS were performed and interpreted at a high volume center whereas initial EUS occurred early in our institution’s learning curve. Subjective impressions and pathology were compared to surgical pathology or clinical course to determine the utility of repeat EUS. Results: Ten patients underwent second and third (n = 3) attempts to achieve diagnosis. Final EUS yielded a pathologic diagnosis or impression that changed clinical course in 4 patients. One diagnosed subjectively with pseudopapillary tumor continues to be followed, one had adenocarcinoma confirmed on surgical pathology and two diagnosed with adenocarcinoma were ultimately proven to have PanIN and benign disease. The final six patients had inconclusive findings on repeat EUS, four of whom underwent successful percutaneous or open biopsies while the final two remain undiagnosed due to refusal of further diagnostic measures. Conclusion: Repeat EUS in experienced hands changes clinical decision making 40 % of the time. Final pathology, however, was discordant from EUS pathology in half of these cases. While the yield is low, referral for salvage EUS in this study justified resection or continued surveillance in highly select patients.

OP-II.43 10 STEPS TO SALVATION – CONTROLLING AN IVC BLEED DURING MAJOR HEPATECTOMY J. B. Conneely1 and S. Jayaraman1,2 University Of Toronto, Toronto, ONTARIO; 2HPB Surgery Service, St. Joseph’s Health Centre, Toronto, ONTARIO 1

This vignette describes an approach to controlling major hemorrhage encountered during a total-laparoscopic right hepatectomy for metastatic breast cancer. A major reason for surgeons’ reluctance to embrace laparoscopic major hepatectomy, is the concern about how to deal with significant intra-operative hemorrhage. In particular, bleeding from the Hepatic veins or inferior vena cava (IVC) is especially ominous, however, certain fundamental surgical principles for safety hold true regardless of the approach. These fundaHPB 2014, 16 (Suppl. 1), 1–99

73

mental principles are as follows: maintaining control of the bleeding site; improving access to the surgical field; and increasing operative space. We describe herein, our approach to controlling and resolving a major bleed from the Right Hepatic Vein – IVC confluence encountered in the final stages of parenchymal division during a formal right hepatectomy. Specifically, we document 10 Steps we took to ensure safe, timely control of a major bleed without compromizing patient safety or oncologic outcome, or having to resort to urgent conversion to open surgery. We propose that these steps can safely be applied to similar intra-operative scenarios.

OP-II.44 SAFE AND EASY PRINGLE FOR LAPAROSCOPIC LIVER RESECTION M. M. Dua1, D. J. Worhunsky1, K. J. Hwa2, B. Siu1, G. A. Poultsides1, J. A. Norton1 and B. C. Visser1 1 Stanford University School Of Medicine, Stanford, CA; 2 VA Palo Alto Health Care System, Palo Alto, CA Introduction: A primary concern during laparoscopic liver resection (LapLR) is bleeding during parenchymal transection. Intermittent vascular occlusion minimizes blood loss; however, is challenging to reproduce during laparoscopy. Described techniques are difficult to achieve emergently or are incompletely occlusive. This study describes the safety and efficacy of a facile method for Pringle maneuver during LapLR. Methods: 145 consecutive patients who underwent LapLR from 2007–2013 were retrospectively reviewed. For Pringle, a laparoscopic articulating esophageal-dissector is used to encircle the hepatoduodenal ligament with an umbilical tape. The tape is externalized through an 12-F flexible Rumel tourniquet which runs parallel to and alongside a port used for the operation. The internal end of the catheter is close to the pedicle and the external end is extracorporeal, allowing for easy pringle (Figure A). Pringle maneuver (PM, n = 82) was compared to “no occlusion” (NO, n = 63) for intraoperative characteristics and postoperative complications. Results: Annual application of the tourniquet and use of Pringle increased from 27% to 80% (p = 0.002) and 20% to 60% (p = 0.04), respectively. Median (IQR) occlusion time was 23 minutes (15–34). There was no morbidity associated © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

74

Abstracts

with placement of the tourniquet. Although operative time and EBL were higher in PM, there were no significant differences noted in transfusion or liver function postoperatively (Table 1). There were only two conversions to open surgery. Conclusion: Extracorporeal tourniquet placement in LapLR is a quick and safe method of gaining control for inflow occlusion. Routine adoption of laparoscopic Pringle facilitates low conversion rates and increased safety.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

OP-II.45 TRICKS OF THE TRADE: OPERATIVE MANAGEMENT OF PORTAL ANNULAR PANCREAS IN PATIENTS UNDERGOING PANCREATICODUODENECTOMY A. V. Maker and V. K. Maker University Of Illinois At Chicago, Chicago, IL Background: Portal annular pancreas (PAP) is an asymptomatic and under-recognized congenital anomaly present in ∼ 2% of the population. It is characterized by pancreatic encasement of the portal vein (PV) where the uncinate fuses to the dorsal surface of the body. It is critical to recognize this anomaly before pancreaticoduodenectomy to appropriately manage the pancreatic anastomosis, minimize post-operative fistula/complications, and to provide an appropriate oncologic resection. Description of the technique: The pancreas encircles the main PV superior to the splenic confluence and the main pancreatic duct bifurcates at the level of the SMA circumscribing the portal vein (Jones modified type II). It is critical to trace the ductal anatomy to determine the appropriate location to divide the pancreas in order to avoid a retroportal anastomosis. Intraoperative ultrasound tracing the ducts is employed and demonstrated. Operative video demonstrates the pancreas encircling the PV (see figure), division of the neck of the pancreas, identification of the ventral duct for an anterior pancreatico-jejunostomy, followed by division of the retroportal pancreas at the level of the SMA, and closure of the dorsal duct. Wide tumor margins are obtained. Conclusion: It is important to be familiar with the anatomy of the PAP in order to recognize this anomaly and avoid technical problems at the time of resection. This video, associated illustrations, and intraoperative photos demonstrate the critical components of the anomaly and illustrate a useful technique to manage tumors requiring pancreaticoduodenectomy in the setting of PAP.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

OP-II.46 TRANCYSTIC BILIARY DESCOMPRESSION WITH INTERNAL-EXTERNAL BILIARY DRAINAGE PLACED LAPAROSCOPICALLY AFTER BILE DUCT EXPLORATION. PRELIMINARY REPORT OF AN INNOVATIVE TECHNIQUE G. E. Pfaffen, P. Barros Schelotto, P. Romero and G. E. Gondolesi Fundacion Favaloro, Buenos Aires, BUENOS AIRES It is part of a good HPB surgical practice to assure an adequate biliary drainage into the duodenum after any laparoscopic common bile duct exploration (LCBDE). Often early after LCBDE, edema or spasm are responsable of the initial Spinchter of Oddi dysfunction (SOD). Pharmacological, endoscopic or surgical procedures have been reported to resolve this event. The aim of this report is to present our preliminary results of three consecutive patients with poor biliary drainage secondary to SOD treated by placement of a Laparoscopic Trancystic Transpapillary Internal-External Biliary Drainage (LTTIEBD). We report 3 patients operated between june 2011 – april 2013, who underwent LCBDE for common bile duct (CBD) stones. Once choledochal lithiasis was removed, the intraoperative cholangiogram (IOC) showed absence or delay empty into the duodenum in all of them. In two patients a LTTIEBD was performed and the postoperative was uneventful. In the 3nd patient, abdominal pain and secondary bile leak through the primary choledochorraphy was diagnosed early after bile duct exploration. Consequently, we reoperated this patient removing the ligature of the cystic stump and a LTTIEBD was used, the patient was discharged 48 hours later. The LTTIEBD was removed in all patient 45 days after placement without complications. During a median follow-up period of 12 months (4–25) no clinical or laboratory abnormalities were observed. The excellent outcome of these patients demonstrates that LTTIEBD could be a useful tool for the management SOD after LCBDE. A larger cohort is needed to validate these results.

HPB 2014, 16 (Suppl. 1), 1–99

75

OP-II.48 INTRAHEPATIC CHOLANGIOJEJUNOSTOMY FOR THE TREATMENT OF COMPLEX BILIARY STENOSIS AFTER PEDIATRIC LIVING-DONOR LIVER TRANSPLANTATION R. Sanchez Clariá, F. A. Alvarez, N. Resio, E. De Santibañes and M. Ciardullo General Surgery And Liver Transplant Unit, Hospital Italiano De Buenos Aires., Ciudad Autonoma De Buenos Aires, BUENOS AIRES Background: The treatment of complex biliary stenosis after pediatric living-donor liver transplantation (LDLT) is challenging, especially when there is an intrahepatic stenosis of the hepaticojejunostomy. Intrahepatic cholangiojejunostomy (IHCJ) has been classically applied for the treatment of malignant disease. We describe an innovative technique of IHCJ for the treatment of complex biliary stenosis after pediatric LDLT. Technique: The first step is to replace a previous percutaneous biliary drainage, usually from an 8 fr catheter to a 12 fr catheter. Once the drain tract is consolidated (3–4 weeks) and a patent hepatic artery has been certified by angiography, the surgery is carried out. During surgery, the previous anastomosis is revised to confirm the absence of extrahepatic bile ducts. The drain is removed and a flexible metal stylet is introduced trough the tract into the bile ducts. Subsequently, the most superficial aspect of the biliary tree is recognized by palpation of the round tip of the stylet in the liver surface. The liver parenchyma is then transected until the bile duct is reached and opened longitudinally (Figure). A side-to-side anastomosis to the previous Roux-en-Y limb is performed using a 6–0 polypropylene suture in an interrupted fashion over a silicone stent. Three pediatric recipients were successfully treated with this approach. Conclusion: The presented technique of IHCJ was feasible and safe in experienced hands. This novel approach allows treating cholestasis in pediatric LDLT recipients with a lack of extrahepatic ducts due to complex biliary stenosis that failed percutaneous treatment, thus avoiding liver retransplantation.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

76

Abstracts

OP-II.49 ULTRASOUND-GUIDED INTRAOPERATIVE TRANSDUODENAL PANCREATIC HEAD BIOPSY A SAFE AND EFFECTIVE PROCEDURE TO EXCLUDE MALIGNANCY FOR OBSTRUCTIVE JAUNDICE PATIENTS: TECHNICAL NOTE N. Ikoma, T. Karahashi and Y. Hosoda Department Of Surgery, Saitama Social Insurance Hospital, Kitaurawa, SAITAMA Background: Despite advanced imaging and endoscopic technologies, differentiating between chronic pancreatitis and pancreatic adenocarcinoma is challenging. At times patients undergo “blind” pancreatic resections without histological confirmation of diagnosis, and consequently bear the significant risk associated with pancreatic surgery for benign disease.

OP-II.50 CONCEPT OF THE MESOPANCREAS: COINING A NEW SURGICAL ANATOMICAL SPACE

Methods: Ultrasound-guided intraoperative transduodenal pancreatic head biopsy was performed for a patient with a pancreatic head tumor with obstructive jaundice, who had equivocal findings on pre-operative evaluation with imaging studies and ERCP, between pancreatic cancer and tumorforming pancreatitis. Results: Frozen section was negative for malignancy and choledocoduodenostomy was performed without complication. The patient has been in good condition with resolution of obstructive jaundice at 24 months follow-up. Conclusion: Ultrasound-guided intraoperative transduodenal pancreatic head tumor biopsy is an alternative strategy to establish an accurate diagnosis before proceeding with extensive pancreatic surgery. This procedure suits the patient with a pancreatic head tumor and obstructive jaundice who requires a biliary drainage procedure.

is an important step into achieving R0 resection which can translate into prolong survival.

S. Patel, S. Cheek, H. Osman and D. Jeyarajah Methodist Dallas, Dallas, TX Background: The concept of the mesorectum has become well accepted. It represents a known space of oncological relevance. In the same light, we believe that the mesopancreas represents essentially an anatomical space with similar anatomical rationale as the mesorectum. The mesopancreas we describe as extending from the posterior aspect of the pancreatic head to behind the mesenteric vessels. It contains lymphatics, nerve plexus, adipose tissue and vascular structures. This concept of the mesopancreas has not been anatomically classified in the surgical literature, though there have been few studies describing the concept. Techniques: After initial exploration, the duodenum is kocherized with a generous Cattell-Brasch maneuver. After division of the pancreatic neck, the proximal jejunum is also divided and passed to the right side underneath the superior mesenteric artery (SMA). The specimen is then retracted anteriorly and medially and the mesopancreas is visualized. The mesopancreas is incised along the SMA heading towards the portal vein in an SMA first approach. The mesopancreas is taken down with the electrocautery bovie on a right angle clamp. In this manner, all tissue in the retropancreatic space and in the interface between SMA and portal vein is included in the specimen. We believe this © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

OP-II.51 FACTORS PREDICTIVE OF CHOLEDOCHOLITHIASIS. OPTIMIZING MANAGEMENT STRATEGIES C. Rengifo, D. Chang, R. Watson, A. Kausar and D. Subar Department Of HPB Surgery, Blackburn, LANCASHIRE Background: The investigation and management of patients with potential bile duct stones vary between institutions and depends on resources. HPB 2014, 16 (Suppl. 1), 1–99

Abstracts Aim: The aim of this study is to develop a risk stratification process for determining the presence of choledocholithiasis in patients with gallstones who do not demonstrate common bile duct stones(CBDS) on ultrasound or have gallstone pancreatitis or cholangitis. Methods: This is a retrospective study of all patients presenting to our institution from March 2011-December 2012 for symptomatic gallstones who had MRCP or ERCP. Risk factors including age, sex, dilation of common bile duct(>7 mm), elevated alanine transaminase(ALT), elevated alkaline phosphatase(ALP) and elevated bilirubin were analysed for predictive significance for the presence of common bile duct stones based on MRCP/ERCP results. Each risk factor was then scored as 0 or 1 based on cut-off value. The cumulative scores were compared to the presence of bile duct stones on ERCP/MRCP. Results: On univariate analysis, age, dilated common bile duct on US(>7 mm), ALP >200 IU/L, ALT >53 IU/L and bilirubin >22 μmol/L were all statistically significant. The positive predictive rate for individual factors ranged from 53.3%-43.5%. A cumulative score of 5 was 100% predictive for CBDS. Scores of between 4- 1 had a predictive probability of between 52.9% and 20%. Patients scoring a zero had no bile duct stones. Conclusion: Patients scoring a 5 on identified risk factors for CBDS should go straight to ERCP. Patients scoring 4-1 should be considered for MRCP. Patients scoring zero should be considered for laparoscopic cholecystectomy without preoperative MRCP/ERCP.

OP-II.53 COMPLICATIONS AFTER SURGICAL REPAIR OF BILE DUCT INJURY D. Zamora-Valdés, M. Vilatoba, A. G. Contreras, I. Dominguez-Rosado, A. L. Carrete, M. Gonzalez, C. Kaufmann, E. M. Saucedo, S. Mier and M. A. Mercado INCMNSZ, Mexico City, DF Background & aim: The surveillance of surgical complications may lead to improvement in medical care. The aim of this paper is to study the incidence and risk factors for surgical complications in patients undergoing bile duct injury (BDI) surgical repair. Methods: We review our prospective database of patients with BDI undergoung surgical repair from January 2003 to December 2011. Surgical complications were graded according to the Clavien-Dindo (CD) and Accordion systems. Comprehensive Complication Index (CCI) was calculated. Complication rate was compared with two other series in the literature. Multivariate analysis was performed to evaluate risk factors associated with major morbility. Results: Surgical repair was performed in 425 patients; Roux-en-Y hepaticojejunostomy (HJ) in 299 patients; neoconfluence in 101 patients (23.8%), portoenterostomy in 25 patients (5.9%) and double barrel HJ in 2 patients (0.5%); 202 patients (47.5%) had complications. Major morbidity occurred in 17.4% of patients. According to CD, postoperative complications were: I 4%, II 26.1%, IIIa 4.2%, IIIb 3.1%, IVa 4.5%, IVb 4.0% and V 1.6%. According to Accordion classification, postoperative complications were: I 4%, II 26.1%, III 4.5%, IV 3.1%, V 8.2% and VI 1.6%. Mean CCI was 15.52 ± 19.36. Risk factors associated with major HPB 2014, 16 (Suppl. 1), 1–99

77

morbility were age, preoperative serum albumin, preoperative pericholangitic abscesses, ASA classification, operative time (hours) and repair other than HJ. Conclusion: Morbidity associated with surgical repair of BDI is noteworthy. Directed strategies should be planned to reduce it. Main factors associated with major morbidity are patient status and complexity of bile duct lesion.

OP-II.54 COMPLICATIONS OF PERCUTANEOUS CHOLECYSTOSTOMY TUBE (PC) IN THE TREATMENT OF ACUTE CHOLECYSTITIS (AC) A. D. Guerron, A. Abdel Azim, M. Moorman, G. McLennan and M. Loor Cleveland Clinic Foundation, Cleveland, OH Introduction: Acute cholecystitis (AC) is a common surgical emergency, often affecting patients with serious medical comorbidities. Procedure related complications and patients risks have not been correlated with outcome. As an initial step in the development of an evidence based care path for AC, we review our series of patients treated with PC. Methods: An IRB approved, retrospective analysis of a prospectively maintained database was conducted. Results: A total of 345 patient underwent PC between 2003 and 2012 were included in the analysis. Males comprised 57% and females 42%. Periprocedural complications occurred in 20% of patients, the most prevalent of which was dislodgement in 36%, followed by bleeding and occlusion (8 and 6%). These were typically handled with tube replacement. In patients who experienced a tube-related complication, mortality was 52%, compared to 44% in the no complication group. A total of 83 patients (27.8%) eventually underwent cholecystectomy at a median of 99 days post PC. Complications presented at a median of 83 days post PC. Patients with tube related complications were less likely to ultimately have a cholecystectomy. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

78

Abstracts

Conclusion: In this study, we identified the rates of complications associated with PC. Cholecystostomy dislodgement or dysfunction are common and predict patient outcome. A minority of patients ultimately underwent cholecystectomy. Further studies regarding timing of these complications are required. This information is useful in guiding the management of patients with severe acute cholecystitis.

OP-II.55 PREDICTING (AVOIDING) INCIDENTAL GALLBLADDER CANCER K. Patel, A. Sautter, T. Buddensick, B. Wu, H. Ferdosi, D. Narducci, M. Siddique, L. Setiawan, H. Shaukat, G. Sulkowski, S. O. Farooqui, G. C. Kowdley and S. C. Cunningham Saint Agnes Hospital And Cancer Institute, Catonsville, MD Introduction: Incidental gallbladder cancer (IGBCa) is an uncommon finding following cholecystectomy. Consequences of discovering IGBCa postoperatively include need for repeat operation and worse median survival if bile spillage was allowed to occur, both of which could be avoided if IGBCa were suspected preoperatively. We hypothesized that criteria may exist to increase the chance of predicting IGBCa. Methods: Retrospective single-institution review identified patients with IGBCa. Precholecystectomy ultrasound (US) images for both IGBCa cases and matched controls were re-reviewed by a blinded radiologist. Statistical analysis was done using chi-square test and t-test. Differences were considered significant(*) when p < 0.05. Results: Among 5770 cholecystectomies performed 2000– 2011, there were 22 (0.38%) IGBCa cases. Patients with IGBCa were older (72 vs 52),* but similar regarding gender and race. Conversion to open cholecystectomy was more common with IGBCa (18% vs 3.5%).* Preoperative imaging tests, including CT and ERCP were more commonly performed with IGBCa.* CBD diameter was higher in IGBCa cases (7.2 mm vs 5.0 mm).* Alkaline phosphatase, bilirubin, and prothrombin time were all significantly higher, and albumin lower, in IGBCa cases.* On blinded review of US imaging, gallbladder wall thickening (GBWT), but not pericholecystic fluid (PCCF) or gallstones, was significantly more common in IGBCa cases (70.0% vs 21.3%)* and mean GBWT was significantly greater in IGBCa cases (7.5 mm vs 3.8 mm).* Conclusion: Advanced age, elevated alkaline phosphatase, bilirubin, and prothrombin time, the existence of multiple preoperative imaging tests, and isolated GBWT without other signs of cholecystitis (eg, PCCF and gallstones) are risk factors that may raise suspicion for IGBCa.

OP-II.56 SHORT TERM PHYSIOLOGICAL EFFECTS OF A HEPATOPANCREATICOBILIARY RESERVOIR A. Yusuf2 and B. Asiyanbola1 University Of Maryland, Easton Memorial, Easton, MD; 2 Johns Hopkins Center For Talented Youth, Baltimore, MD

1

Introduction: Recent reports of the surgical placement of hepatopancreatobiliary (HPB) reservoirs have been pub© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

lished. Such reservoirs are potentially of use in targeted or regional management because unlike systemic therapy, targeting or regionalizing appears to be more efficient and less costly. Reservoirs have been used to manage HPB malignancies and diabetes. However, there are scant reports of the physiologic effect of creating reservoirs. The objective is to investigate what short term physiologic changes occur with creation of a gallbladder reservoir. Methods: Approval from the Animal Care and Use Committee was obtained. Five female swine including two controls and three test subjects were used. Gallbladder reservoirs whose walls were cystic mucosa and serosa containing flurescent nanomicelles were created, while controls had incision of the gallbladder. Drainage was through the cystic then portal veins into liver. Serum samples were taken for liver function tests (LFT’s) and inflammatory markers (IF) at 0, 12, 24, and 36 hours. Ultrasound was carried out at 12 and 24 hours. Euthanasia was at 36 hours and samples sent for histology. Results: Similar to the swine that acted as controls, serum LFT’s and IF rose in the subject swine 12 hours after surgery then normalized at 24–36 hours. Ultrasound revealed gallbladder contraction on feeding at 24 hours in both groups. Both groups showed histological inflammation and microscopy in subject swine revealed nanomicelles. Conclusion: In this short term study, there was no gross adverse effect on the physiology of the gallbladder. Validation of these findings as well as more research is needed.

OP-II.57 NEUROENDOCRINE NEOPLASMS OF THE EXTRAHEPATIC BILE DUCT: REPORT OF SIX CASES N. Hong and H. Kim Department Of Radiology, Asan Medical Center, University Of Ulsan College Of Medicine, Seoul, SONGPA-GU Neuroendocrine neoplasms (NENs) are rarely occurred in the extrahepatic bile duct with a small number of cases reported to date. We herein report 6 more cases of NENs of the extrahepatic bile duct which are diagnosed postoperatively in our single institution over a 12-year period (from January 2001 to June 2012) and discern a set of characteristic features through retrospective review of their medical records. There were three male and three female with average age of 59.2 years (range, 41–68 years). The most common presenting symptoms were right upper quadrant or epigastric discomfort and obstructive jaundice. According to the recent WHO classification, five patients were diagnosed neuroendocrine carcinomas (NECs) and one patient was mixed adenoneuroendocrine carcinoma (MANEC). Five patients had their tumors in the common bile duct, and only one patient in the cystic duct. Morphologically, four tumors were solid mass, one was cystic mass, and the other one was segmental duct wall thickening. Most patients received adjuvant chemotherapy and recurrences were detected in four patients involving liver, lymph node, pancreas and peritoneum within 2–7 months after surgery. Only one patient has survived beyond 3 years without recurrence. NECs of the extrahepatic bile ducts are such aggressive neoplasms with poor prognosis and although preoperative diagnoses are not easy, should be considered in differential diagnosis of malignant bile duct obstruction, to be treated with different stratHPB 2014, 16 (Suppl. 1), 1–99

Abstracts egy from ordinary bile duct cancer. However, further studies are needed to comprehend their own clinical behaviors.

OP-II.58 INCIDENTAL GALLBLADDER CANCER ON CHOLECYSTECTOMY: IS THERE MORE THAN MEETS THE EYE? S. Patel, S. Cheek, T. Winston, H. Osman and D. Jeyarajah Methodist Dallas, Dallas, TX Background: Gallbladder cancer (GBC) remains a rare and underdiagnosed malignancy. This study aims to describe the experience at non-university tertiary care center (NUTCC) and to evaluate if patients with incidental GBC on cholecystectomy benefited from further evaluation and radical resection. Methods: Patients with GBC from 2006–2013 were retrospectively identified in the cancer registry. Patient were divided into 2 groups; patients with incidental GBC on cholecystectomy prior to referral to our NUTCC (Group A) and patients who were referred prior to surgical intervention based on suspicion from radiographic studies (Group B). Pathological staging was then analyzed. Results: 36 patients were identified, 23 (64%) in Group A and 13 (36%) in Group B. 77% were females and 23% were males. In Group A; 4 (17%) were found to have incidental stage II GBC after cholecystectomy, 13 (56%) had stage III, and 6 (26%) had stage IV. All patients in Group A were then further evaluated at our center; 6 (26%) were deemed unresectable based on imaging. 17 patients underwent radical resection and residual disease was found in 10 (58%) patients. 50% of patients with initial stage II had residual disease. In Group B 7 (54%) were unresectable and 6 (46%) had stage III (T3 or N1) disease. Conclusion: GBC remains an aggressive disease. Majority of patients with incidental GBC on cholecystectomy will have further disease, including patients with early disease on initial pathological staging. All these patients should be evaluated for radical resection. Furthermore, radiographic findings suspicious for GBC was an ominous sign for advanced stage.

OP-II.59 NOVEL STRATEGY FOR SUSPECTED PT2 GALLBLADDER CARCINOMA BY LAPAROSCOPIC APPROACH O. Itano1, G. Oshima2, M. Shinoda1, M. Kitago1, Y. Abe1, T. Hibi1, H. Yagi1, S. Aiko2 and Y. Kitagawa1 1 Keio University School Of Medicine, Tokyo, TOKYO; 2 Eiju General Hospital, Tokyo, TOKYO Purpose: The purpose of this study was to evaluate the results of laparoscopic approach for suspected pT2 gallbladder carcinoma. Methods And Procedures: From December 2007 to January 2013, the laparoscopic approach was performed for treating 12 patients with suspected pT2 gallbladder carcinoma. First, staging laparoscopy was done to further exclude peritoneal dissemination, liver invasion and liver metastases. After dissection of Calot’s triangle, surgical margin of the HPB 2014, 16 (Suppl. 1), 1–99

79

cystic duct was examined by intraoperative pathological examination. Cholecystectomy and gallbladder bed resection were performed by laparoscopic procedure. Intraoperative pathological examination was performed again to confirm the presence of carcinoma and the depth of invasion. Surgery was completed when the diagnosis was benign disease. Additional lymphadenectomy was performed according to the depth of invasion. Biliary tract reconstruction was performed through 10-cm subcostal incision. Results: There was no conversion to laparotomy. For two patients with benign lesion, surgery finished without lymphadenectomy. For two patients with T1b carcinoma, lymphadenectomy was performed at the hepatoduodenal ligament. Other 8 patients with T2 were added lymphadenectomy around the pancreas head. Two patients underwent bile duct resection due to aberrant pancreaticobiliary duct junction. The mean operative time was 320 ± 123 minutes and the mean blood loss was 96.5 ± 140 ml. There were no postoperative complications and the mean postoperative hospital stay was 6.8 ± 1.8 days. After a mean follow–up of 21 months, all patients with gallbladder carcinoma survived without recurrence or metastasis. Conclusion: Our strategy for suspected pT2 gallbladder carcinoma is feasible and safe.

OP-II.60 CHOLECYSTECTOMY IN CIRRHOTIC PATIENTS. WHEN TO OPERATE? L. Sarotto1,2, J. Paleari1,2, H. Todeschini1,2 and P. Ferraina2 Instituto Quirurgico Del Callao, Capital Federal, BUENOS AIRES; 2Universidad De Buenos Aires, University Of Medicine, Capital Federal, BUENOS AIRES

1

Background: The incidence of vesicular lithiasis in patients with cirrhosis is high implying higher morbidity and mortality.The hepatic cirrhosis is a severe risk factor due to the general patient’s condition and the technical surgical difficulties. Ojbective: To assess the experience in our laparoscopic cholecystectomy ward with cirrhotic patients and to determine the best time to operate on these patients. Design: Retrospective and descriptive study. Material and Method: 24 patients with hepatic cirrhosis subjected to cholecystectomy due to vesicular condition at the “ Instituto Quirurgico del Callao” between June 2007 and June 2013. were included. Each patient was assessed regarding their sex, age, clinical , biochemical, surgical and child’s classification parameters, post operative evolution, hospital stay and histological diagnosis. Results: The group consists of 15 women and 9 men with an average of 60.2 years. The first cause of cirrhosis was viral (58,3%). The surgical indications were symptomatic colelithiasis in 19 patients and severe cholecystitis in 5 patients. The pre-surgical diagnosis for cirrhosis was performed on 18 patients and intraoperative diagnosis was permormed on 6 patients. 16 patients were subjected to laparoscopic cholecystectomy. All the patients operated who had been prediagnosed with cirrhosis presented a type A Child’s classification. 3 patients presented mild ascities as a post surgical complications. The average hospital stay was 4 days. Conclusions: The laparoscopic cholecystectomy is a safe and efficient method in patients with early and compensated © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

80

Abstracts

cirrhosis. At present, cirrhosis is not a counter – indication laparoscopic cholecystectomy if performed by experimental surgeons.

SUNDAY, FEBRUARY 23, 2014, 7:30AM–8:30AM PARALLEL COMPETITIVE VIDEOS A LIVER CV-A.01 TOTALLY LAPAROSCOPIC RIGHT HEPATECTOMY BY CAUDATE HEPATOTOMY TECHNIQUE C. U. Corvera and V. Kuo University Of California San Francisco, San Francisco, CA The patient is a 59 year-old woman with synchronous metastatic colon cancer to the liver. Cross-sectional imaging showed several hepatic tumors distributed in the right lobe. Given the location and distribution, a right hepatectomy was planned. This video describes our preferred anterior approach for a formal laparoscopic right hepatectomy– A generous caudate hepatotomy is used to help gain access to the main right portal vein. The coronary ligament is incised and the hepatic veins are exposed. The hilar dissection was commenced by exposing the right hepatic artery lying posterior to the common hepatic duct. The main RHA is clipped and divided. The main portal vein was dissected until the bifurcation is clearly identified. A caudate hepatotomy was done in order to allow isolation and division of the main RPV by a single firing of a vascular stapler. The line of demarcation along the interlobar plane was marked using electrocautery and the parenchymal transection was started. The caudate hepatotomy was extended further posteriorly beyond the hilum and then the anterior liver was divided. As the liver was opened, the right bile duct was transected intrahepatically. Coursing branches of the middle and right hepatic veins were also managed intrahepatically using a vascular stapler. The parenchymal transection plane was aligned along the right lateral edge of the IVC until the dissection reached the base of the Right hepatic vein. The RHV was identified and divided and the right hemi-liver was mobilized from diaphragm. The specimen was extracted via a low transverse incision.

CV-A.03 LAPAROSCOPIC RIGHT POSTERIOR SEGMENT LIVER RESECTION M. M. Augustine and T. P. Kingham Memorial Sloan Kettering Cancer Center, New York, NEW YORK The purpose of this video presentation is to highlight one case from our experience performing laparoscopic, right posterior segment liver resection. We illustrate the postioning of patient, port placement, and the use of laparoscopic intraoperative ultrasound to define tumor position and approach to resection. Through our right side mobilization, © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

which includes right lateral port placement through the chest, we show how we can access the right posterior liver. The use of an ultrasonic generator, bipolar cautery, and stapler for transection of the liver parenchyma are demonstrated. Lastly, we present extraction from a transvaginal route when combined with a gynecologic laparoscopic operation.

CV-A.04 ROBOTIC-ASSISTED RIGHT HEPATECTOMY E. Vicente, Y. Quijano, B. Ielpo, H. Duran, E. Diaz, I. Fabra, C. Oliva, S. Olivares, R. Caruso, R. Ceron and J. Plaza Sanchinarro University Hospital, Madrid, MADRID Background: Different technical and technological advances have significantly reduced the morbidity and mortality of hepatic resection in the past 20 years. It has been proved that the use of minimally invasive procedures in liver surgery improved the postoperative comfort of the patient as well as some perioperative complications, like blood loose. Among minimally invasive surgeries, laparoscopy has occupied a place of privilege. However, it still has important limitations that Robotic technique evolved to overcome. Materials and Methods: We present a Robotic-assisted right hepatectomy in a patient affected by colorectal liver metastasis. Results: It is showed the different steps of the procedure which includes the liver mobilization and hilar dissection, the liver transection and right supra-hepatic vein section. Conclusion: This video shows the efficacy and feasibility of Robotic liver surgery providing a three dimensional imaging, a greater instrumental movement and tremor filtering of the surgeon, thus reducing the minimally invasive surgery related complexity.

CV-A.05 LAPAROSCOPIC LEFT HEPATECTOMY COMBINING CUSA AND HARMONIC SCALPEL DISSECTION J. B. Conneely, R. Smoot and S. Cleary Toronto General Hospital, Toronto, oNTARIO The development and increasing availability of laparoscopic versions of the so-called ‘energy devices’ for vessel sealing and division has been a boon to laparoscopic surgeons worldwide. Liver resection in particular has been aided by the improved performance and evolving ergonomics of these devices. However, there remains great benefit to be derived from the use of the laparoscopic version of the Cavitron Ultrasonic Surgical Aspirator, or ‘CUSA’, for many years, the workhorse of open liver resection. Herein, we describe our technique of parenchymal division during major hepatectomy, for which we employ the combined benefits of the CUSA and the Harmonic Scalpel. During a formal left hepatectomy, we demonstrate the precise parenchymal dissection made possible by the laparoscopic CUSA. We also demonstrate how the use of the Harmonic Scalpel in tandem with the laparoscopic CUSA can facilitate efficient parenchymal transection by avoiding unnecessary clip application which can be slow and may affect the use of laparoscopic staplers. We propose that this technique of liver resection is safe, efficient and precise. HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

SUNDAY, FEBRUARY 23, 2014, 7:30AM–8:30AM PARALLEL COMPETITIVE VIDEOS B PANCREAS/BILIARY CV-B.01 COMPLETION RADICAL CHOLECYSTECTOMY AND PORTAL LYMPHADENECTOMY: MINIMALLY INVASIVE TECHNIQUES FOR MANAGEMENT OF INCIDENTAL GALLBLADDER CARCINOMA S. Lee, M. I. DAngelica, W. R. Jarnagin and T. P. Kingham Memorial Sloan-Kettering Cancer Center, New York, NY – NEW YORK Introduction: Gallbladder cancers staged T1b or higher are treated with a liver resection and portal lymphadenectomy. Methods: We present 2 patients and their surgical videos to illustrate our experience. The patients had preoperative staging imaging and index pathology was reviewed. Diagnostic laparoscopy was performed prior to proceeding with liver resection and lymphadenectomy. A 30-degree laparoscope, bipolar, surgical staples and harmonic scalpel were used to perform the segment 4b/5 liver resection and portal lymphadenectomy. The use of the da Vinci Surgical System(Si system, Intuitive Surgical Inc., Sunnyvale, CA) was illustrated in one case. Frozen section and intraoperative ultrasonography assisted in determining extent of resection. Results: Four patients underwent a prior laparoscopic cholecystectomy at an external institution with an incidental diagnosis of T1b or higher gallbladder adenocarcinoma were treated with a minimally invasive approach for completion radical cholecystectomy and portal lymphadenectomy. After a staging CT scan to exclude metastatic disease, a laparoscopic/robotic portal lymphadenectomy, frozen section of the cystic duct remnant and segment 4b/5 liver wedge resection was performed. The mean estimated blood loss was 150 cc (range: 50 to 400 cc), operative time was 225 minutes (range: 165 to 290 mins), length of stay was 5 days (range: 4 to 7 days) and there were no complications. The median number of lymph nodes harvested was 6 and there were no residual cancer seen in resected liver segment 4b/5 in all the cases. Conclusion: A minimally invasive approach to early gallbladder cancers can be safely utilized to perform a completion radical cholecystectomy and portal lymphadenectomy.

CV-B.02 SURGICAL ASPECTS OF THE MESENTERICOPORTAL VENOUS RESECTION IN HEPATO-BILIO-PANCREATIC SURGERY E. Vicente, Y. Quijano, B. Ielpo, H. Duran, E. Diaz, I. Fabra, C. Oliva, S. Olivares, R. Caruso, R. Ceron, A. Cubillo and J. Plaza Sanchinarro University Hospital, Madrid, MADRID Neoplasm of the hepatobiliopancreatic (HPB) field may be deemed locally unresectable because they appear to encase the superior mesenteric vessels. Limited data exist on the ability of high-quality contrast-enhanced CT to accurately predict the need for venous resection at the time of resection. HPB 2014, 16 (Suppl. 1), 1–99

81

Intra-operative evaluation of the relationship of a tumor of the superior mesenteric vein (SMV) or portal vein is prone to inaccuracy. SMV resection and reconstruction at the time of HPB tumors add significant complexity to an already lengthy operation. However, with proper patient selection and surgeon experience, SMV resection and reconstruction can be performed safely. The need of venous resection does not significantly impact survival duration. The video show a variety of methods for venous resection and reconstruction.

CV-B.03 PANCREATIC NEUROENDOCRINE TUMOR WITH CELIAC AXIS INVOLVEMENT V. M. Cano Busnelli, F. A. Alvarez and J. Pekolj HPB Surgery Section And Liver Transplant Unit. Hospital Italiano De Buenos Aires., Capital Federal, CAPITAL FEDERAL Introduction: Pancreatic neuroendocrine tumors (PNETs) are rare neoplasms. Complete surgical resection is the only curative treatment and remains the cornerstone therapy for these patients. Surgically aggressive approaches have resulted in improved long-term survival in locally advanced disease. Methods: We present a 41-year-old male with diagnosis of a pancreatic neuroendocrine tumor of the body and tail that was biopsed during elective open cholecystectomy. Chromogranin and synaptophysin were positive and the Ki67 was 60%. A CT-scan revealed compromise of the celiac axis and the posterior gastric wall. Selective common hepatic artery embolization was performed to increase flow to the Proper Hepatic artery via the Gastroduodenal artery. A preoperative angiography certified flow on the Proper Hepatic Artery from the pancreatoduodenal arcades from the Superior Mesenteric artery. A distal esplenopancreatectomy with en bloc total gastrectomy, transverse colectomy and celiac axis resection was carried out. Digestive reconstruction was accomplished with a circular stapled Rouxen-Y esophagojejunostomy. A linear stapler was used for duodenal and pancreatic section. Results: The patient developed a duodenal fistula that did not require further treatment rather than a longer hospital stay. The histo-pathological report confirmed a neuroendocrine origin. Conclusion: The presented Appleby operation was feasible and safe in experienced hands with satisfactory results. In patients with tumoral involvement of the celiac axis, preoperative embolization of the hepatic artery offers the possibility to achieve complete tumor removal; therefore offering a chance of cure to patients previously declared unresectable.

CV-B.04 LAPAROSCOPIC COMMON BILE DUCT EXPLORATION FOR IMPACTED CHOLEDOCHOLITHIASIS S. Jayaraman University Of Toronto, Toronto, ON – ONTARIO The indications for common bile duct exploration are limited with the widespread availability of endoscopic retrograde cholangiopancreatography (ERCP). However, in selected cases, surgical clearance of choledocholithiasis is necessary due to anatomic factors or failed ERCP. This video demonstrates the case of an elderly woman with several medical © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

82

Abstracts

comorbidities. She had ERCP with stent insertion and papillotomy for a bile duct that was impacted with gallstones. Dozens of stones were removed endoscopically at that time; however, a second ERCP was performed due to recurrent jaundice and symptoms. The first stent had migrated and the duct was still impacted despite removing even more stones. She was referred for laparoscopic common bile duct exploration. This video demonstrates all of the steps required for total clearance of the common duct in a challenging case: duct exposure, choledochotomy, stone and stent extraction, complete choledochoscopy and stone extraction, duct closure, cholecystectomy, and cholangiography. A T-tube was not inserted at the conclusion of the operation as the patient had a recent papillotomy and her anatomy was normal (no Roux-en-Y).

CV-B.05 LAPAROSCOPIC PARTIAL SLEEVE DUODENECTOMY (PSD) FOR DUODENAL ADENOMA J. A. Stauffer and H. J. Asbun Mayo Clinic, Jacksonville, FL The video submission is a case presentation of a 65-year-old female who was found to have a circumferential adenoma of the infra-ampullary duodenum. She was taken to surgery and underwent a laparoscopic partial sleeve duodenectomy (PSD). The video shows technical considerations and steps of laparoscopic PSD. Laparoscopic PSD spares resection of the head of the pancreas by separating the duodenum from the head of the pancreas under direct magnified vision. Intraoperative upper endoscopy was performed to ensure complete adenoma removal and patency of the ampulla prior to transection of the duodenum just distal to the ampulla. Reconstruction was performed with a side to side duodenojejunostomy. The patient was found to have a 5-cm tubulovillous adenoma with high grade dysplasia and recovered without any complications.

SUNDAY, FEBRUARY 23, 2014, 10:30AM–12:30PM PARALLEL MINI ORAL A LIVER/TRANSPLANT MO-A.01 OVERALL SURVIVAL BY PATTERN OF RECURRENCE FOLLOWING CURATIVE INTENT SURGERY FOR COLORECTAL LIVER METASTASIS A. N. Kulaylat, J. R. Schubart, A. L. Stokes, E. T. Kimchi, K. F. Staveley-O’Carroll, N. J. Bhayani, J. T. Kaifi and N. J. Gusani Section of Surgical Oncology, Penn State Hershey Cancer Institute, Hershey, PA Introduction: Following curative intent surgery (CIS) for colorectal liver metastasis (CRLM), patterns of recurrence and subsequent survival outcomes are not widely reported. Methods: An institutional database (January 2002December 2012) was reviewed to evaluate patterns of recurrence following CIS for CRLM. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

Results: 183 patients with CRLM underwent CIS (liver resection, ablation, or both). Median follow-up and progression-free survival (PFS) was 28 and 16 months, respectively. 5-year overall survival (OS) was 51%. After initial CIS, 102 (56%) patients recurred: 44% liver (5-year OS 55%), 15% lung (5-year OS 45%), and 41% other/ multifocal (5-year OS 20%). While OS for liver vs. lung was not significantly different, both were better compared to other/multifocal (p = 0.018 for liver vs. other, p = 0.086 for lung vs. other). Wider liver resection margins reduced the risk of general (RR 0.95, p = 0.018) and isolated liver (RR 0.93, p = 0.043) recurrence. Other/multifocal recurrence was more likely with extrahepatic disease at time of liver resection (RR 7.83, p = 0.045) and bilateral liver metastasis (RR 3.61, p = 0.041), but less likely with staged resections (RR 0.11, p = 0.016), and increasing liver resection margin (RR 0.90, p = 0.012). Of 102 1st recurrences, 53 patients were able to undergo repeat CIS (in 71% liver recurrences, 67% lung, and 24% other/multifocal). When repeat CIS was possible, 5-yr OS was 67%. Conclusion: Patients with liver- or lung-only recurrences after CIS for CRLM have high rates of repeat CIS and excellent survival. Repeat CIS offers improved survival and should be pursued where possible.

MO-A.02 INTRA-OPERATIVE LIVER BIOPSY PATHOLOGY IN PATIENTS UNDERGOING WEIGHT LOSS SURGERY A. Dan1,3, M. Subichin1,3, J. Clanton1,3, M. Makuszewski3, A. Bohon3, M. Pozsgay1,3, R. Shenoy1, S. M. Thompson2,3, A. Venkataramani4, C. Sievers4 and J. Zografakis1,3 1 Summa Health System Department Of Surgery, Akron, OH; 2Summa Health System Department Of Pathology, Akron, OH; 3Northeast Ohio Medical University, Rootstown, OH; 4Summa Health System Department Of Gastroenterology, Akron, OH Introduction: The incidence of liver disease in the morbidly obese varies, but most studies to date are small series. To determine the degree of liver pathology at the time of weight loss surgery, liver biopsy is often performed. We sought to review the utility of routine liver biopsy and determine the frequency of abnormal liver pathology in a large sample of morbidly obese patients. HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

83

Methods: A retrospective chart review of 1,000 consecutive patients undergoing weight loss surgery was performed. Data collected included: age, sex, BMI, liver function tests and liver biopsy pathology results. Student’s t-test was used to compare groups. Results: The average age of the study population was 48 years, with a 4 : 1 female to male ratio. All patients had a BMI greater than 35 kg/m2. Of these patients, 81.2% had liver pathology related to obesity seen on pathology, including 70.3% with steatosis (Grade 1–3), and 4.6% with nonalcoholic steatohepatitis (NASH). The mean BMI of patients with steatosis was 47.7 compared to a BMI of 48.7 without steatosis (p = 0.088). Mean BMI of patients with and without NASH was 49.9 and 47.9, respectively (p = 0.11). Conclusion: Significant liver pathology is highly prevalent in the obese population, but is not statistically associated with higher BMI. Pathologic examination is an excellent modality to delineate and specify the extent of the disease. This data represents the largest single center analysis of consecutive liver biopsies of patients with morbid obesity. Liver biopsy remains a valuable component in the treatment of patients undergoing weight loss surgery.

MO-A.03 ALTERATIONS IN HIPPO SIGNALING PATHWAY DURING LIVER REGENERATION J. Grijalva1, M. Huizenga2, F. Camargo1, G. Sadri-Vakili2, K. Vakili1 1 Boston Children’s Hospital, Boston, MASSACHUSETTS; 2 Massachusetts General Hospital, Charlestown, MA Background: The Hippo signaling pathway has been implicated in organ size regulation and hepatocyte proliferation, however, its role in liver regeneration (LR) is unclear. The Hippo pathway modulates activity of the transcriptional co-activator YAP which regulates a proliferative transcriptional program. Phosphorylation of YAP (pYAP) results in its degradation. Thus, we hypothesized that the Hippo pathway modulates LR and we sought to characterize its signaling alterations during LR. Methods: 70% rat partial hepatectomy (PH) model was used. Remnant livers were collected 1–30 days post-PH (n = 4–8/group). Liver/body weight ratios were calculated. Western blot analysis was performed on YAP, pYAP, Mst1 and phosphorylated Mst1. YAP target gene expression of Cyr61 and AmotL2 was measured using RT-qPCR. Each animal served as its own control. Results: The liver/body weight ratio approached preoperative levels 7 days post-PH. A Significant increase in total YAP was observed 1–7 days post-PH (P < 0.05) followed by a significant decrease in pYAP from days 3–14 (p = 0.045). A Significant decrease in Hippo kinase Mst1 phosphorylation (a ctivated) was observed 1–3 days post-PH (P < 0.05) (Figure 1). Significant increase in Cyr61 (P < 0.005) and AmotL2 (P = 0.0167) mRNA expression was observed 1–3 days post-PH. Conclusions: Our findings demonstrate dynamic changes in the Hippo pathway following PH that stabilize when liver/ body weight ratio reaches normal levels. Additionally, our findings demonstrate increased levels of active YAP and increased YAP target gene expression during LR. These results suggest that this pathway may be an important mechanism underlying hepatocyte proliferation during LR and provide a potential therapeutic target to enhance LR. HPB 2014, 16 (Suppl. 1), 1–99

MO-A.04 VALIDATION OF THE SELECTION CRITERIA FOR SAFE HEPATECTOMY FOR HEPATOCELLULAR CARCINOMA: ANALYSIS OF 312 CONSECUTIVE HEPATECTOMIES M. Donadon, F. Procopio, M. Cimino, G. Costa, D. Del Fabbro, A. Palmisano, A. Gatti and G. Torzilli Liver Surgery Unit, Department of General Surgery, Humanitas Clinical And Research Center, Rozzano, MiLAN Introduction: Hepatectomy is the treatment of choice for hepatocellular carcinoma(HCC). The estimation of functional liver reserve is the stronghold to achieve good results. The aim of this study was the validation of the selection criteria for safe hepatectomy for HCC. Methods: Clinical, pathological and short-term data were reviewed in a prospective cohort of patients who underwent hepatectomy between 2004–2013. All consecutive patients were submitted to hepatectomy based on the same selection criteria, which included bilirubin(BIL), cholinesterases (CHE), the presence of ascites, the presence of esophageal varices, and the residual liver volume. Univariate, and multivariate analyses were performed. Results: A total of 312 patients were analyzed. One hundred three patients (33%) had thoracoabdominal approach, but only 36 (12%) had major or extended resections. The median tumor number was 1(range 1–33), while the median tumor size was 4,5 cm (range 0.8–28). BCLC were 0-A in 53%, B in 25%, and C in 22%. Of those patients, 94 (29%) had complications, of which 6% were major(Dindo > IIIa). The 30- and 90-days mortality were 1.2 % and 1.8% respectively. The MELD, APRI, and CPT-scores were tested. None of these were found to be statistically significant at the logistic © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

84

Abstracts

regression analysis, while the use of the combination of BIL > 1 and CHE≤5.900 was the best combination to detect complications (P = 0.007), and specifically liver-related complications such as ascites (P = 0.000). Conclusions: This study shows that our selection criteria that counts mainly on two simple, fast, and cheap parameters, BIL and CHE, lead to identify patients potentially at risk of complications after hepatic resection for HCC.

MO-A.05 READMISSION AFTER HEPATIC RESECTION AT A QUATERNARY ACADEMIC CENTER A. Ejaz, G. Spolverato, D. Lucas, M. Weiss, C. Wolfgang, K. Hirose, M. Choti and T. M. Pawlik Johns Hopkins Hopsital, Baltimore, MD Introduction: Reducing readmission has become a key quality improvement target for policymakers. We sought to define the incidence and identify factors associated with readmission after hepatic resection at a quaternary academic center. Method: Readmission within 30 days of discharge after hepatic resections between 2008 and 2012 at a major academic center was analyzed. Modified Poisson regression with a robust variance estimator was used to calculate risk ratios for readmission. Results: Among 357 patients indications for surgery included colorectal liver metastasis(42.6%), non-colorectal liver metastasis(35.3%), primary hepatic tumors(17.7%), and benign disease(2.8%). Comorbidities were common: 74.2% were ASA class 3–4. Operations included non-anatomic resection(69.5%), hemi-hepatectomy(17.1%), and trisegmentectomy(12.9%). Median length-of-stay was 5 days; 48.2% patients experienced at least one inpatient complication. Overall 30-day readmission was 11.8%(n = 42); an additional 4(1.1%) patients died at home within 30-days. Causes of readmission included abdominal infection(40.5%), liver dysfunction (7.1%), other gastrointestinal complication(31.0%), venous thromboembolism(7.1%), pulmonary complication(4.8%), and other(11.9%). Median length-ofstay for the readmission was 4 days. Factors associated with readmission included preoperative weight loss(RR-2.38, 95% CI 1.35–4.18), MELD score of >10(RR-2.26, 95% CI 1.10–4.64), ICU length-of-stay >1 day(RR-2.02, 95% CI 1.14–3.58), total length-of-stay >5 days(RR-2.60, 95% CI 1.39–4.83), and any inpatient complication(RR-1.94, 95% CI 1.07–3.52)(all P < 0.05). On multivariable analysis, the strongest predictor of readmission was length-of-stay >5 days (adjusted RR-2.30, 95% CI 1.24–4.28, p = 0.009) (Figure). Conclusion: Readmission after hepatic resection occurred in roughly 1 in 10 patients. Risk of readmission was associated with patient- (weight loss, MELD score), surgical(history of complication) and hospital- (length of stay) level factors.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

MO-A.06 QUALITY OF LIFE AFTER RESECTION OF INTRA- AND EXTRA-HEPATIC METASTASES FROM COLORECTAL ADENOCARCINOMA IN A MULTI-CENTER PROSPECTIVE PHASE II TRIAL P. E. SErrano1, N. G. Coburn2, K. S. Devitt1, C. Moulton1, S. P. Cleary1, C. Law2, P. D. Greig1, S. Gallinger1 and A. C. Wei1 1 Hepatobiliary And Surgical Oncology, Toronto, ON; 2 Hepatobiliary And Surgical Oncology, Toronto, ON The concept of combined resection of intra- and extrahepatic metastases (IHM and EHM) from colorectal cancer (CRC) is evolving. This study evaluated the quality of life (QOL) of patients with IHM and EHM from CRC undergoing complete metastasectomy. Methods: Participants of a Phase-II multi-institutional trial with any number of CRC IHM and up to 3 foci of EHM, resectable with RO intent completed QOL questionnaires (EORTC-QLQ-C30, EORTC-LMC 21 and FACT-Hep) at baseline, prior to surgery and at 4-8-12 months following complete metastasectomy. Mean scores were compared to baseline. Change >10% was considered a minimally important clinical difference (MICD). Results: There were 25 participants. Median disease-free survival was 6 (0–17) months. EORTC-QLQ-C30 globalQOL remained statistically and clinically unchanged compared to baseline at 4 and 8 months but had a clinically significant decline at 12 months (−15.3, P = 0.25). FACTHep Total (−8.3, P = 0.03) and Trial Outcome Index (−6, P = 0.03) score were statistically lower at 4 months post surgery, without reaching a MICD. At this time-point there was a transient increase in fatigue (+17.7, P = 0.004), appetite loss (+12.8, P = 0.02) and a decrease in social (−22.4, P = 0.004) and role (−12.6, P = 0.002) functioning. All subscales returned to baseline levels at 12 months. Conclusions: Despite early disease recurrence after aggressive metastasectomy of multi-site CRC, there was no significant deterioration of global QOL as a result of treatment. There was a transient increase in symptoms (fatigue and loss of appetite) and a decrease in functioning scores (role and social) at 4 months from surgery, improving by 12 months.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

MO-A.07 INTRAOPERATIVE ULTRASOUND AND SURGICAL STRATEGY IN HEPATIC RESECTION: WHAT DIFFERENCE DOES IT MAKE? R. Jrearz1, R. S. Hart1,2 and S. Jayaraman1,2 Division Of General Surgery, University Of Toronto, Toronto, ON; 2HPB Service, St. Joseph’s Health Centre, Toronto, ON 1

With modern advancements in preoperative imaging for liver surgery, intraoperative ultrasound (IOUS) may be perceived as superfluous. Our aim was to determine if IOUS provides new information that changes surgical strategy in hepatic resection. From a single institution, 111 consecutive patients who underwent 121 liver resections were analyzed. Preoperative CT and/or MRI determined the surgical strategy. Comparisons between the number, size, and location of lesions were compared between IOUS and preoperative imaging. Reasons for changes in surgical strategy were documented. Pathology reports were analyzed for margins. Continuous variables were compared using parametric t-tests and categorical variables were compared using the chi square test for independence. P-values of less than or equal to 0.05 were considered statistically significant. IOUS was used in 88 of 121 procedures. IOUS changed the surgical plan in 17% (n = 15) cases. Additional tumors were detected in 7% (6) of cases. A change in size and location were detected in, 9% (8) and 7% (6) of cases respectively. Surgical plans were changed in 8% (7) cases for reasons not related to IOUS. IOUS helped demarcate the resection plane in the remaining 75% (66) of cases. The average margin was 1.09 cm (1.18) for IOUS and 1.18 cm (1.05) for non-IOUS. There were no significant differences in margin length (p = NS) between the groups. Surgical strategy was directly altered due to IOUS in a substantial number of cases. IOUS guided resection planes resulted in R0 resections in nearly all procedures. The best operative plan in hepatic resection includes intraoperative ultrasound.

MO-A.08 RELIABILITY, SAFETY AND ONCOLOGICAL SUITABILITY OF INTRAOPERATIVE ULTRASOUND CRITERIA FOR HEPATIC VEINS SPARING ONCE IN CONTACT WITH COLORECTAL LIVER METASTASES AT THE CAVAL CONFLUENCE: VALIDATION ON AN INTENTION TO TREAT ANALYSES F. Procopio, M. Donadon, M. Cimino, D. Del Fabbro, G. Costa, A. Gatti, A. Palmisano and G. Torzilli University Of Milan-School Of Medicine, Humanitas Clinical And Research Center, Rozzano, MILAN Major-hepatectomies are generally selected for tumors in contact/involving the hepatic-vein(HV) at caval confluence (CC). Whenever a conservative approach is attempted, HV-reconstruction is recommended. The use of criterion based on intraoperative-ultrasonography (IOUS) may limit the need for extensive hepatectomy. An intention-to-treat prospective study to validate this hypothesis has been conducted in a cohort of patients with at least one colorectalHPB 2014, 16 (Suppl. 1), 1–99

85

metastasis (CLM) in contact with a HV at CC. Thirty-eight patients showing at preoperative imaging at least one CLM in contact with a HV at CC were enrolled.Based on IOUSfindings, 3 patterns of presentation were recognized and respective surgical policies were established : Type-A : HV in contact with CLM for <1/3 of its circumference, regardless of longitudinal extension : vein-sparing with tumordetachment was planned; Type-B : HV-involvement ranging 1/3–2/3 of its circumference, regardless of longitudinal extension : vein-sparing with HV wall-resection and reconstruction by direct-suture or patching was planned;TypeC : HV-involvement for ≥2/3 of its circumference with/ without interruption of vessel-wall at IOUS : vein-resection was planned. At IOUS-exploration of the 38 patients, 25 included at least a Type-A contact, 13 Type-B, while 9 Type-C invasion. No major-hepatectomy was required. Tumor-detachment from HV was feasible in 30 (79%) patients, HV-reconstruction with direct-suture in 13 (34%), while HV-resection was performed in 9 (23%). Postoperative mortality was nil.Overall and major morbidity rate was 39%, 5%. No local-recurrences have been observed at a median follow-up of 13 months (range 6–34). Preserving liverparenchyma in hepatectomy is the key-factor to ensure better postoperative results. This proposed tumor-vessel classification has been validated on an intention-to-treat perspective and in all patients was able to predict the proper surgical strategy. Adopting these criteria major-hepatectomies can be avoided in favor of a surgical policy, which is safe and featured by acceptable oncological radicality.

MO-A.09 PREOPERATIVE PREDICTIVE MODEL OF SURVIVAL AFTER RESECTION FOR INTRAHEPATIC CHOLANGIOCARCINOMA (IHCC) A. Schneck1, D. Subar2, R. Memeo1, F. Vittore1, P. Compagnon1, A. Laurent1 and D. Azoulay1 1 Service De Chirurgie Digestive, APHP Henri Mondor Hospital Créteil, VAL DE MARNE; 2Department Of General And HPB Surgery, Blackburn Royal Hospital, Blackburn, LANCASHIRE Background: Patient survival with intrahepatic cholangiocarcinoma (IHCC) after resection depends on histopathological factors. One study has reported on preoperative predictive factors. This study assesses a preoperative model in predicting survival of patients with IHCC who have undergone curative resection. Methods: This is a retrospective single institution study of patients who had curative hepatic resection between September 1990 and December 2012 for IHCC. Preoperative factors including patient sex, age, body mass index (BMI), performance status (PS), serum bilirubin, serum alkaline phosphatase (ALP), CA19.9, tumour size and number of tumour nodules on preoperative computed tomography (CT) were converted to categorical variables and analysed. Significant predictive factors on uni-and multivariate analysis was scored as 0 or 1 based on cut-off value for each significant variable and entered into a model. Results: 78 patients had hepatic resection for IHCC. The male to female ratio was 50 : 28 and mean age was 60 years. The median survival was 12.2 months. Factors predictive of survival in univariate and multivariate analyses were © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

86

Abstracts

BMI > 30 kg/m2, ALP >115 IU/L, tumour size > 6 cm, CA19.9 > 35 IU/ml and greater than 1 nodule present on preoperative CT. Classification of patients based on a predictive model using a score of 1 for low risk, 2–3 for intermediate risk and greater than 3 for high risk patients was predictive of survival (p < 0.05). Conclusion: Our model based on preoperative factors can predict long term survival. The model needs to be validated in a prospective trial and may help in patient selection for surgical intervention.

MO-A.10 OUTCOMES OF UNRESECTABLE SOLITARY HEPATOCELLULAR CARCINOMAS SIZED 5 TO 8 CM TREATED BY RADIOFREQUENCY ABLATION VERSUS TRANSARTERIAL CHEMOEMBOLIZATION W. Dai, T. Cheung and R. Poon Department Of Surgery, The University Of Hong Kong, Hong Kong, China, Pokfulam, NA Objective: To compare the efficacy of radiofrequency ablation (RFA) versus transarterial chemoembolization (TACE) in treating large (5–8 cm) unresectable solitary hepatocellular carcinomas (HCCs). Design: Retrospective study of prospectively collected data. Setting: RFA and TACE as feasible options for treating large solitary HCCs unresectable because of poor liver reserve or other patient factors. Patients and Intervention: Among 41 patients who had large unresectable solitary HCCs, 15 patients received RFA and 26 patients received TACE. Main Outcome Measures: The primary endpoint was overall survival and the secondary endpoints were tumor response, time to disease progression, and treatment-related morbidity and mortality. Results: The complication rate was 33% in the RFA group and 19.2% in the TACE group (P = 0.525). In the RFA group, the complete ablation rate was 80% and local recurrence rate 6.7%, and the median disease-free survival was 13 (2.8–38) months. In the TACE group, there was no complete response whereas the partial response rate was 15.4%, and the median time to disease progression was 8 (1–68) months. The RFA group had a median survival duration of 39.8 months and that in the TACE group was 19.75 months (P = 0.257). The 1-, 2and 5-year survival rates were 93.3%, 86.2% and 20.9% respectively in the RFA group and 73.1%, 40.6% and 18.3% respectively in the TACE group. Conclusion: RFA resulted in better initial tumor control and a higher short-term survival rate, but the two treatment modalities brought about similar long-term survival.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

MO-A.11 THE ASSOCIATION OF HYPOPHOSPHATEMIA WITH POST-HEPATETCTOMY LIVER FAILURE AND RECOVERY J. Hallet1, F. Zih2, J. Wong3, E. Cheng3, S. Hanna3, P. Karanicolas3, N. Coburn3 and C. Law3 1 Division Of Surgical Oncology, University Of Toronto, Toronto, ONTARIO; 2Department Of General Surgery, University Of Toronto, Toronto, ONTARIO; 3Division Of Surgical Oncology, Sunnybrook Health Sciences Centre, Toronto, ONTARIO Introduction: Hypophosphatemia is frequent following hepatectomy and thought to represent phosphate use during liver regeneration. Its association with postoperative liver function remains controversial. We sought to evaluate the association of post-hepatectomy hypophosphatemia with liver failure and recovery. Methods: Four hundred liver resections were retrospectively reviewed from 2003 to 2013 at a single institution. Primary outcome was occurrence of liver failure (“50/50” criteria). Secondary outcomes were liver recovery at discharge, 30-day morbidity according to Clavien-Dindo classification, and 30-day mortality. Outcomes of clinically significant posthepatectomy hypophosphatemia cases (HP [≤0.65 mmol/L]) were compared to non-hypophosphatemia controls (NHP). Results: 320 (80.0%) patients experienced HP. Liver failure occurred significantly more often with HP than NHP (19.7% Vs. 11.2%, p0.02). However, we observed a trend towards less liver failure at discharge in HP (1.7% Vs. 4.3%, p0.20). Length of stay, Clavien-Dindo complications, and mortality were similar. 82.0% of liver failures recovered before discharge. More patients who recovered experienced HP compared to those who did not (70.7% Vs. 11.1%, p < 0.01). Demographics, diagnosis and extent of surgery did not differ between hypophosphatemia or recovery groups. Conclusion: Post-hepatectomy HP is associated with increased incidence of liver failure, but not increased morbidity or mortality. HP occurred more often in patients who recovered, highlighting efficient liver regeneration. Further studies should explore the usefulness of the trend of phosphate drop for early identification of patients likely to recover fast from liver failure. HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

MO-A.12 VALIDATION OF OUR POLICY OF LONG-TERM DRAINS MAINTENANCE AFTER HEPATIC RESECTION: RESULTS OF A PROSPECTIVE COHORT ANALYSIS M. Donadon, G. Costa, F. Procopio, M. Cimino, D. Del Fabbro, A. Palmisano, A. Gatti and G. Torzilli Liver Surgery Unit, Department Of General Surgery, Humanitas Clinical and Research Center, Rozzano, MILAN Background: The utility of drains after hepatic resection is debated. The aim of this study was the validation of our definition, and management of biliary fistula after hepatic resection. We also reviewed the literature, in particular the definition by the International Study Group of Liver Surgery(ISGLS). Methods: Data on abdominal drains and on clinical, pathological and short-term outcome were reviewed in a prospective cohort of patients who underwent hepatic resection between 2004–2013. Drains were maintained at least 7 days, and the bilirubin levels were measured in POD3, 5, and 7. Drains were removed if the bilirubin level in POD7 was inferior than in POD5, and less than 10 mg/dl. Statistical analysis on prognostic factors for biliary fistula was performed. Results: Among 475 consecutive patients, 39 (8%) had biliary fistulas. Only 8 (1.7%) patients required interventional procedures. The area under the receiveroperating-characteristic curve on POD7 had the highest predictive value(0.81, p < 0.001). The multivariate analysis found a protective role for cirrhosis (OR = 0.56 = , p = 0.045), while a Pringle’s maneuver >90 minutes (OR = 3.45, p < 0.001), extended resections (OR = 6.45, p = 0.007), blood transfusions (OR = 2.4, p = 0.035), and resections including segment-I (OR = 1.94, p = 0.033), and segment-V (OR = 1.87, p = 0.024) were found to be associated with increased risk. Using the definition proposed by the ISGLS 66% of our patients would have received a diagnosis of biliary fistula. Conclusions: The maintenance of abdominal drains up to POD7 reduces the risk of biliary fistulas after hepatic resection. The greater is the complexity of hepatic resection the greater is the risk of biliary fistula, which may effectively managed with drains maintenance.

MO-A.13 LIVER CANCER IN TEXAS: ENVIRONMENTAL RISK FACTORS CONTRIBUTING TO HIGHER INCIDENCE RATE D. Zorzi1, G. Curcuru’2, C. Rhoads1, V. Gazis1, C. Rastellini1 and L. Cicalese1 1 UTMB, Galveston, TX; 2University Of Palermo, Palermo, PA Introduction: Incidence rate for primary liver cancer (LK) in Texas is rising and above US rate (9.3 vs. 6.7, respectively in 2010). Texas has the largest oil and gas industry and is the second largest agricultural producer. No population-based studies are available investigating Hazard Air Pollutants (HAPs) exposure as a contributing risk factor for LK. This study aims to evaluate the level and the distribution of HAPs and LK rate in Texas counties. HPB 2014, 16 (Suppl. 1), 1–99

87

Methods: Incidence rate per counties was provided by the Texas Cancer Registry. Total emission values (ev) in ton per year (tpy) for 188 HAPs were monitored by US Environmental Protection Agency. Counties with zero LK incidence (Gr0) versus counties with the highest LK incidence (over 10: Gr10) were compared. Results: Mean age-adjusted LK incidence rate in Texas (2002–2010) is 7.57 ± 2.58. Gr0 included 10 counties with the presence of 80% (151/188) HAPs. Gr0 HAPs mean ev was always below Gr10 mean ev (range 0–20 tpy). Gr10 included 15 counties with the presence of all 188 HAPs; Gr10 HAPs mean ev was consistently higher compared to Gr0 (over 100 tpy). In Gr10, ten counties were identified with HAPs total ev greater than state HAPs mean ev. Specifically 9 HAPs classified as cancerogenic, were identified for highest and significantly different ev (Table 1). Conclusions: In Texas counties at higher level of HAPs total ev corresponds a higher LK incidence rate. Our findings suggest HAPs exposure as a potential additional factor to be considered in the multifactorial liver carcinogenesis.

MO-A.14 (SIGMA-2/SMAC) IS A STRONG ENHANCER OF GEMCITABINE SENSITIVITY IN PANCREATIC ADENOCARCINOMA Y. Hashim1, D. Spitzer1, S. Vangveravong1, P. Goedegebuure1, M. Hornick1, R. Mach2, W. Hawkins1 1 Washington University School Of Medicine In St Louis, Department Of Surgery, St Louis, MO; 2University Of Pennsylvania, Department Of Radiochemistry, Philadelphia, PA Introduction: Gemcitabine remains a standard chemotherapy for pancreatic adenocarcinoma. Unfortunately the response rate is 9.4% and OS is extended by a few months. Second mitochondria-derived activator of caspases (SMAC) kills cells by reversing the activity of IAPs and was shown to increase pancreatic cancer sensitivity to Gemcitabine. Sigma-2 receptors are highly expressed in pancreas cancer cells. We developed a novel drug Sigma-SMAC (SW IV-134) that can be delivered selectively to pancreatic cancers using sigma-2 ligands and targeting IAPs pathway. This drug has high potency with minimal toxic effects. Here we investigated the combination of this targeted drug with Gemcitabine to improve therapeutic response and minimize toxicity. Methods: We tested sub-therapeutic doses of SW IV-134, Gemcitabine, and the combination on pancreatic cancer cells to test for synergy. Titer Glo viability, Caspases and annexin assays were performed to compare the potency between the drugs and their combination. Result: Viability assays showed 15% of the cells were alive after treatment with combination © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

88

Abstracts

therapy compared with 80% of cells treated by the single compounds. Caspase levels were 2.5 fold the baseline in combination therapy compared to 1.5 fold of single agent. FACS for Annexin staining showed 27%, 15% and 5% apoptotic cells treated with combination, Gemcitabin and SW IV-134 respectively. P values for these tests were < 0.05. Conclusion: We developed (SW IV-134) with promising efficacy and minimal toxicity. It has strong synergetic effect in combination with Gemcitabine. This strategy has a potential to improve chemotherapeutic efficacy, minimize toxicity and benefit patients with pancreatic cancer.

Conclusion: Our center listed and transplanted nearly a third of potentially eligible patients, based on stage. 23% of all HCC patients were transplanted. This is the highest utilization reported and a high benchmark, favoring HCC patients. The rising incidence of HCC, MELD scores, and a well organized liver cancer program contributed to this high utilization. Of immediate concern is the detrimental effect of this high utilization on patients with end stage liver disease awaiting LTx. This requires further study. On the other hand, the high utilization of HCC LTx in upcoming eras of hepatitis C cure may not be undesirable.

MO-A.17 REOPERATION FOR BLEEDING FOLLOWING LIVER TRANSPLANTATION D. A. DuBay, D. T. Redden, M. A. Thompson, L. Glueckert, A. B. Smith, J. H. Crawford, K. A. Jones, S. H. Gray, J. White and D. E. Eckhoff University Of Alabama At Birmingham, Birmingham, AL

MO-A.16 HIGH UTILIZATION OF LIVER TRANSPLANTATION FOR HEPATOCELLULAR CARCINOMA: IS IT CONCERNING? A. E. Alsina1, K. Vu4, J. Arrobas2, E. Franco1, N. Kemmer3, G. W. Neff3, A. Valencia2 and S. Houston2 1 Tampa General Hospital, Liver Transplantation Surgery, Tampa, FL; 2Tampa General Hospital, Office Of Clinical Research, Tampa, FL; 3Tampa General Hospital, Transplant Hepatology, Tampa, FL; 4University Of South Florida, Morsani College Of Medicine, Tampa, FL Data regarding the utilization of liver transplantation (LTx) for hepatocellular carcinoma (HCC) is limited by registry studies. Updated studies are needed from experienced LTx centers. We aimed to investigate the utilization of LTx for HCC in the last 4 years. We hypothesize that the utilization at this center is high, based on dedicated HCC program, MELD scores and other factors. Methods: The study comprised 353 consecutive HCC patients referred and evaluated at our transplant and cancer program between Jan. 2009 to Dec. 2012. Demographics, tumor characteristics, and Barcelona Clinic staging were calculated. Stage A and B patients were considered potentially eligible for LTx. Results: 23% of all HCC and 29% of eligible HCC received LTx. Median Wait List Time was 35 days. Main reasons for not referring were: tumor characteristics (76%) and progression (9%). Reasons for not listing were: comorbidities (27%), tumor characteristics ( 25%), and financial/insurance (12%). © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

Introduction: This study’s objective was to identify risk factors associated with liver transplantation (LTx) bleeding requiring reoperation as well as the impact on mortality and short-term clinical outcomes. Methods: A retrospective study was performed to identify risk factors associated with reoperation for bleeding following LTx performed at a single institution between 2000– 2012. Operative reports were used to identify patients who underwent reoperation for bleeding within 1 week following LTx (operations for non-bleeding etiologies were excluded). Results: Reoperation for bleeding was observed in 110/ 1136 (9.7%) of LTx patients. The following characteristics were associated with reoperation: recipient MELD score, number of platelets transfused, and non-utilization of Amicar (Table). Stratified by MELD score, the risk of reoperation for bleeding was 4.4% (MELD ≤10), 7.9% (MELD 10–19), 12.3% (MELD 20–29) and 19.0% (MELD ≥30). Stratified by units of platelets (plts) administered, the risk of reoperation for bleeding was 13.5% (no plts), 10.6% (1 u plts), 9.6% (2 u plts) and 6.5% (≥3 u plts). LTx patients who underwent reoperation for bleeding had a longer ICU stay (5 days ± 7 vs. 2 days ± 3, p < 0.001) and hospitalization (18 days ± 9 vs. 10 days ± 18, p < 0.001). The risk of death increased in patients who underwent reoperation for bleeding (HR 1.89, 95% CI 1.26, 2.85). Conclusion: Reoperation for bleeding following LTx was associated with increased resource utilization and recipient mortality. A lower threshold for intraoperative platelet transfusion and antifibrinolytics, especially in patients with high lab-MELD score, may decrease the incidence of reoperation for bleeding following LTx. HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

89

MO-A.19 SURVIVAL OUTCOMES IN LIVER TRANSPLANT RECIPIENTS WITH MELD OF 40 AND HIGHER A DECADE LONG EXPERIENCE H. J. Panchal1, J. Patterson2, J. Durinka3 and J. Ortiz3 Mount Sinai School Of Medicine, NY, NY; 2Lincoln Medical And Mental Health Center, NY, NY; 3Albert Einstein Healthcare Network, Philadelphia, PA 1

MO-A.18 THE IMPACT OF PRE-TRANSPLANT HOSPITALIZATION ON POST-OPERATIVE LENGTH OF STAY AND COST OF LIVER TRANSPLANTATION S. J. Knechtle, L. I. Brummett, C. Devroy, R. Subramanian, A. B. Adams, J. Magliocca and J. Spivey Emory University School Of Medicine, Atlanta, GA Background: Pre-operative acuity impacts outcomes of surgery and length of stay (LOS) although this has not been systematically evaluated for liver transplantation (LT). We used cost data to evaluate the impact of pre-transplant hospitalization on length of post-operative hospitalization and cost of initial hospital stay at this transplant center. Preoperative hospitalization was used as a surrogate for acuity. Methods: Between December, 2009 and July, 2013, 368 consecutive adult deceased donor LT were performed in 354 patients. 26 combined organ transplants were excluded, 14 retransplants were excluded, and 3 for incomplete cost data. For the remaining 325 LT, the data represent the total cost incurred for treatment of patients between the transplant and discharge from the hospital, excluding physician professional costs. The patients were divided into two groups: patients transplanted from home and patients hospitalized prior to LT. Results: 93 patients were hospitalized at our center prior to LT and 232 were admitted from home for their LT. The table shows total cost and LOS for the two groups. Conclusion: Pre-transplant hospital admission reflected greater acuity of illness as reflected by almost twice as long LOS and significantly higher charges. When bundled payment methodology and accountable care organizations consider LT reimbursement, acuity of illness substantially influences LOS and cost of LT and will need to be considered so as to not discourage care of sicker patients. The impact on cost of additional measures of acuity such as MELD, ICU stay pre-transplant, and BMI are also being evaluated at our center.

HPB 2014, 16 (Suppl. 1), 1–99

Introduction: The model for end-stage liver disease (MELD) score has been used as a prioritizing and prognostic tool since 2002, and has stabilized pre-transplant mortality. However, the number of waitlisted candidates withdrawn because they were too ill to be transplanted has doubled between 2009 and 2011. The increasing proportion of transplant candidates with a higher MELD score combined with the improvement in transplant outcomes mandate the need to study surgical outcomes in those patients with MELD score of 40 or higher. Methods: A retrospective analysis of all UNOS data on all liver transplantations performed between February 2002 and June 2011. Our primary outcomes of interest were graft and patient survival at 1, 3, 12, 36, and 60-months. A Kaplan Meier Product limit method was used to determine survival rate. Results: A total of 3848 transplant recipients were identified based on MELD score with a median age of 51 years, 63.6% of which were male. Graft survival at 1, 3, 12, 24, 36, and 60 months was 88.9%, 83.2%, 74.4%, 69.2%, 66.1%, and 61.4% respectively; patient survival was 90.6%, 85.6%, 77.9%, 73.1%, 70.2%, and 65.9% respectively. A multivariate Cox regression analysis was done to identify significant predictors of graft and patient survival. Donor and recipient age, Hispanic and Asian ethnicity, diabetes, and ICU hospitalization at transplant were all identified as predictors of graft and patient survival. Conclusion: The results indicate that liver transplantation in recipients with a MELD score of 40 or greater is promising with significant survival after transplantation.

MO-A.20 THE USE OF VENOUS JUMP GRAFTS IN PANCREAS TRANSPLANTATION IS NOT ASSOCIATED WITH INCREASED GRAFT FAILURE OR MORTALITY E. Siskind1, M. Akerman1, S. Ashburn1, E. Molmenti1, A. Basu1 and J. Ortiz2 1 North Shore Long Island Jewish Health System, Manhasset, NY; 2Albert Einstein Medical Center, Philadelphia, PA Introduction: Pancreas transplantation is an established treatment for patients with type 1 diabetes and renal failure or life threatening glucose unawareness. The technical components of the operation may vary by individual surgeon. Venous jump grafts are used to salvage a pancreas with a short portal vein or to facilitate an easier anastomosis. Early reports indicated that the use of venous jump grafts led to an increased risk of vascular thrombosis. There have been no large reports evaluating the safety of venous jump grafts in pancreas transplantation. We analyzed the UNOS database to determine whether venous jump grafts are associated with graft loss or patient death. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

90

Abstracts

Materials and methods: Data from UNOS on all adults pancreas or kidney pancreas transplant recipients between 1996 and 2012 were analyzed. Venous extension grafts were used in 2,657; they were not used in 18,124. Chi square test was used to compare many of the standard categorical variables. The MannWhitney test was used to compare the two groups for continuous variables. Kaplan-Meier/Product-Limit Estimates and their corresponding 95% confidence intervals were computed. One, 3, 5, 10 and 15 year unadjusted patient and graft survival rates were computed for both groups. Results: Kaplan -Meier analysis demonstrated similar patient survival (p < 0.641) and death censored graft survival (p < 0.351) at 1, 3, 5, 10 and 15 years between subjects with and without venous jump grafts. Conclusion: The use of venous jump grafts is NOT associated with a higher risk of graft loss or patient death.

SUNDAY, FEBRUARY 23, 2014, 10:30AM–12:30PM PARALLEL MINI ORAL B PANCREAS/BILIARY MO-B.01 IS CHROMOGRANIN A PROGNOSTIC FOR RESECTED PANCREATIC NEUROENDOCRINE TUMORS?

pancreatic neuroendocrine tumors (pNETs), but results to date have been controversial. Our purpose was to explore the clinical usefulness of preoperative CgA in predicting outcome in patients with resected pNETs. Methods: Patients who underwent resection of a pNET from 2002–2013 and had a CgA level tested within 3 months before surgery were identified from a prospective database. An elevated preoperative CgA was defined as a CgA lab value above the normal limit of the assay. Results: A total of 38 patients met inclusion criteria, of which 16 patients (42%) had elevated preoperative CgA. Of these, 45% were male and the median age was 57 years (range, 17–81). Operations included: 13 pancreaticoduodenectomies, 22 distal pancreatectomies, 1 central pancreatectomy and 2 enucleations. There were no differences in tumor size, node positivity, margin status, or tumor grade between the two groups on univariate analysis. Both diseasefree survival (p = 0.008) and overall survival (p = 0.05) were negatively impacted by an elevated preoperative CgA level (median FU for surviving patients, 2.5 years). The only independent predictor of DFS on multivariate analysis was elevated preoperative CgA (HR 9.36, 90% CI 1.60–54.7, p = 0.037). There were no independent predictors of OS. Conclusion: In patients with resected pNETs, an elevated preoperative CgA level was associated with decreased disease-free and overall survival, and remained the only independent predictor of DFS. Preoperative CgA appears to be a clinically useful prognostic marker following resection of pNET.

M. A. Shanahan, C. S. Cho, P. A. Rudnitzky, G. Leverson, E. R. Winslow and S. M. Weber University Of Wisconsin School Of Medicine & Public Health, Madison, WI Introduction: Previous studies have suggested prognostic potential for the tumor marker chromogranin A (CgA) for

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

MO-B.02 DETECTION OF SOMATIC MUTATIONS IN DIAGNOSTIC FINE NEEDLE ASPIRATES OF PANCREATIC CANCER WITH NEXT-GENERATION SEQUENCING V. Valero1, T. J. Saunders2, R. H. Hruban1, J. L. Cameron1, J. M. Herman3, C. A. Iacobuzio-Donahue1,2 and C. L. Wolfgang1,2 1 The Johns Hopkins University School Of Medicine, Department Of Surgery, Baltimore, MD; 2The Johns Hopkins University School Of Medicine, Department Of Pathology, Baltimore, MD; 3The Johns Hopkins University School Of Medicine, Department Of Radiation Oncology, Baltimore, MD A personalized approach to the treatment of pancreatic cancer that is based on genetic status depends on tissue for genetic analysis. Fine needle aspiration (FNA) is the most commonly used method for obtaining tissue prior to surgical resection. Potentially limiting factors of FNA include the paucity of malignant cells within a fibroblast-rich stroma and tumor heterogeneity. Therefore, we sought to determine: the ability to confidently detect mutant alleles in low cellularity samples, the concordance between the FNA and the primary tumor and the feasibility of identifying driver mutation in clinical FNAs. Methods: We determined the ability to detect cancerspecific mutations in a wild-type stromal background in an in vitro model of Panc1 cells diluted into the fibroblast cell line, CAF 35, and the genotypic concordance between FNAs and their matched resected human adenocarcinoma tissues using next-generation sequencing. Finally, the feasibility of driver mutation identification via FNA was explored by sequencing clinical FNA samples. Results: The limit of confident detection occurred at a 1 : 25 ratio of malignant to stromal cells for both the KRAS and TP53 mutant alleles corresponding to a hypothetical carcinoma with 4% neoplastic cellularity. The concordance between the FNA and tumor samples ranged from 83% to 100%, although there was 100% concordance for driver mutations in all samples. Sequencing of clinical FNAs revealed driver mutations of KRAS, TP53 and SMAD4 genes. Conclusion: Confident genotyping of low cellularity FNA samples is possible and accurately detects the driver mutations in primary tumors thus allowing the potential for personalized patient care.

91

Methods: A prospectively-maintained database of consecutive patients undergoing RPD from 2008–2013 was analyzed and compared to consecutive OPD cases performed from 2007–2008 (prior to robot utilization). Patients with borderline resectable disease were excluded. Additionally, the previously identified learning curve for RPD of 80 patients was excluded, ensuring the optimization of outcomes in both groups. Results: Demographics and outcomes of 58 RPD patients were compared with 98 patients undergoing OPD. There were no significant differences in age, gender, BMI, number of patients with prior abdominal surgery, Charlson Comorbidity Index, ASA, pathologic indication, malignancy, tumor size or Ca 19-9 between groups. Peri-operative outcomes are shown in Table 1. RPD was associated with significantly lower blood loss, higher lymph node yield, and lower incidence of pancreatic leak. There was a trend toward slightly longer operative time in the RPD group that did not reach statistical significance. Conclusion: Beyond its initial learning curve, RPD is associated with outcomes comparable to open resection. In this series, RPD was associated with improved operative blood loss, pancreatic leak rate, and lymph node harvest compared with ODP. Further study of RPD, potentially including a multi-institutional comparison to OPD, is warranted.

MO-B.04 COMPARISON OF ROBOT VERSUS OPEN PANCREATICODUODENECTOMY B. A. Boone1, S. Downs-Canner1, J. Steve1, M. S. Zenati1, M. E. Hogg1, A. J. Moser2, D. L. Bartlett1, H. A. Choudry1, H. J. Zeh1, A. H. Zureikat1 1 Department Of Surgery, University Of Pittsburgh, Pittsburgh, PA; 2Department Of Surgery, Beth Israel Deaconess Medical Center, Boston, MA Introduction: The safety, feasibility and learning curve of robotic pancreaticoduodenectomy (RPD) has been demonstrated, however data comparing large series of RPD to open pancreaticoduodenectomy (OPD) is lacking. Therefore, we sought to compare outcomes of RPD to OPD. HPB 2014, 16 (Suppl. 1), 1–99

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

92

Abstracts

MO-B.05 GENERAL SURGERY RESIDENT PARTICIPATION IN COMPLEX PANCREATIC PROCEDURES Y. Tilahun1, P. Veldhuis1, S. Eubanks1, P. Arnoletti1,2 and S. G. De La Fuente1,2 1 Florida Hospital Orlando, Orlando, FL; 2University Of Central Florida, Orlando, FL Background: Growing concerns exist regarding competence of general surgery (GS) residents in complex pancreatic procedures as more HPB-focused fellowships become available. This limited experience may be associated with worse patient outcomes. We analyzed postoperative outcomes in patients that underwent complex pancreatic surgeries with the participation of GS residents. Methods: Data from the NSQIP-PUF from 2011 was used to assess association between resident participation and postoperative outcomes. The primary variables measured were 30-day postoperative mortality, overall complication rate, serious complication rate, and length of stay (LOS). Results: Of the 2,156 surgeries identified, GS residents participated in 85% (n = 1,839) of cases. Standard pancreaticoduodenectomy was the most common procedure (n = 682, 37%), followed by distal pancreatectomy (n = 458, 25%), and pylorus-preserving pancreaticoduodenectomy (n = 335, 18%). Distribution of cases was similar in the no-resident group. Patient demographics were similar between groups, although more patients in the resident group had CHF, PVD, TIAs, and bleeding disorders. Both anesthesia (no-resident 385 ± 186 min vs. resident 406 ± 163 min, p = 0.019) and operative times (no-resident 311 ± 172 min vs. resident 331 ± 151 min, p = 0.019) were longer in resident-assisted cases. No differences in overall complication rate, serious complications or mortality rates were found between groups, although LOS was longer for those patients without GS residents. Conclusion: Participation of GS residents in complex pancreatic procedures is associated with longer anesthesia and operative times. This increment in times does not result in higher early postoperative morbidity or mortality.

MO-B.06 DIAGNOSTIC LAPAROSCOPY SHOULD BE PERFORMED BEFORE DEFINITIVE RESECTION FOR PANCREATIC CANCER: A FINANCIAL ARGUMENT T. T. Jayakrishnan1, H. Nadeem1, R. T. Groeschl1, B. George1, J. P. Thomas1, P. S. Ritch1, K. K. Christians1, S. Tsai1, D. B. Evans1, S. G. Pappas2, T. C. Gamblin1 and K. K. Turaga1 1 Medical College Of Wisconsin, Milwaukee, WISCONSIN; 2 Loyola University Medical Center, Maywood, ILLINOIS Objective: To study the cost-effectiveness of routine diagnostic laparoscopy (DL) at the time of definitive resection for pancreatic cancer (PC) in the surgery-first (SF-PC) and neoadjuvant therapy paradigms (NAT). Background: Laparoscopy is recommended to detect radiographically-occult extra-pancreatic disease in patients with localized pancreatic cancer to prevent non-therapeutic laparotomy. There is concern that the selection process of © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

neoadjuvant therapy and repeated imaging may make routine laparoscopy unnecessary. We hypothesized that DL remains cost effective in patients who undergo pancreatectomy treated with the neoadjuvant approach. Methods: Decision tree modeling compared routine DL versus exploratory laparotomy (ExLap) at the time of curative resection for – (i) SF-PC setting, and (ii) borderline resectable PC (BR-PC) in the NAT setting. Costs (US$) from payers’ perspective, quality-adjusted-life-months (QALMs) and incremental-cost-effectiveness-ratios (ICER) were calculated. Base case estimates and multi-way sensitivityanalyses were performed. The willingness-to-pay (WTP) was $4,166/QALM (or $50,000/QALY). Results: The base-case costs were $34,921 for ExLap vs. $33,441.74 for DL in the SF-PC setting; and $39,633 ExLap vs. $39,713 DL in BR-PC. Routine laparoscopy is the dominant (preferred) strategy in the surgery first approach (cost reduction of $10,695/QALM) and in borderline resectable patients treated with neoadjuvant therapy (additional cost of $4,158/QALM which is lower than the willingness-to-pay of $4,166/QALM). Conclusion: Our analysis supports the cost-effectiveness of routine diagnostic laparoscopy before curative resection of pancreatic cancer patients treated either with surgery first or neoadjuvant approach.

MO-B.07 THE EFFECT OF MARGIN STATUS ON SURVIVAL FOLLOWING PANCREATICODUODENECTOMY D. J. Niemeyer, R. M. Seshadri, R. Z. Swan, D. Sindram, R. Nazemzadeh, J. B. Martinie and D. A. Iannitti Carolinas Medical Center, Charlotte, NC Introduction: Pancreaticoduodenectomy (PD) is a key component of the multidisciplinary treatment of pancreatic adenocarcinoma. Margin status of the resection remains a controversial subject, with some small studies demonstrating no difference in survival between margin negative (R0) resections and microscopically positive (R1) resections. The current study aims to compare survival of patients with stage II and stage III pancreatic cancers depending on resection margin status. Methods: The Commission on Cancer’s National Cancer Database (NCDB) is a registry of 342,267 patients diagnosed with pancreatic cancer from 1998–2011. Demographic, cancer related, and treatment related data are available for all patients. Survival data is available before 2007. Results: A total of 27,818 PD were performed for adenocarcinoma between 1998–2011. Survival data was available for 16,925, of which 2,582 were identified as clinical stage II or stage III disease. A large number of patients did not have pre-op staging information. Overall survival for stage II disease was significantly higher with R0 compared with R1 resections (median 18.99 vs. 12.98 months, p < 0.0001). There was no statistically significant survival difference between R0 and R1 resections for Stage III disease (median 15.44 vs 14.59 months, p = 0.1017). Conclusion: The NCDB provides the opportunity to examine a large number of PD patients in a multiinstitutional retrospective manner. Survival of stage II pancreatic adenocarcinoma after a PD is negatively affected by margin status. Survival of stage III disease, defined by HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

93

involvement of large vessels, is unaffected by microscopically positive surgical margins.

MO-B.08 NEOADJUVANT RADIATION IS EQUIVALENT TO ADJUVANT RADIATION FOLLOWING PANCREATICODUODENECTOMY FOR STAGE II AND III ADENOCARCINOMA D. J. Niemeyer, R. M. Seshadri, R. Z. Swan, D. Sindram, A. J. Crimaldi, J. B. Martinie and D. A. Iannitti Carolinas Medical Center, Charlotte, NC Introduction: The use of radiation therapy for treatment of pancreatic adenocarcinoma is well established in the US. The timing of treatment relative to pancreaticoduodenectomy (PD), either as neoadjuvant (NR) or adjuvant radiation(AR) is debated. The current study aims to compare survival of patients with stage II and stage III adenocarcinoma depending on the use and timing of radiation therapy. Methods: The Commission on Cancer’s National Cancer Database (NCDB) is a registry of 342,267 patients diagnosed with pancreatic cancer from 1998–2011. Demographic, cancer specific, and treatment related data are available for all patients. Survival data is available before 2007. Results: 27,818 PD were performed for adenocarcinoma from 1998–2011. Survival data is available for 16,925, of which 2,582 were identified as clinical stage II or III, with many having no pre-op staging information. 130 stage II patients underwent NR, 674 received AR. 81 stage III patients received NR, while 266 received AR. For stage II and III disease, any form of radiation improved survival significantly compared with no radiation (Stage II no rad/any rad 13.8/20.47 months p < 0.0001, stage III no rad/any rad 9.79/17.05 months p < 0.0001). Comparison of NR to AR was insignificant for stage II and III disease (stage II NR/AR 24.44/20.24 months p = 0.0689, stage III NR/AR 19.09/ 16.46 months p = 0.4030). Conclusion: The use of radiation as an adjunct to surgical resection significantly improves overall survival in pancreatic cancer. Analysis of the large number of patients in the NCDB demonstrate no significant survival benefit of NR when compared to AR in stage II and stage III disease.

HPB 2014, 16 (Suppl. 1), 1–99

MO-B.09 RISK FACTORS FOR PANCREATIC FISTULA AFTER ROBOTIC PANCREATICODUODENECTOMY P. M. Polanco, M. E. Hogg, M. Shakir, M. S. Zenati, M. Choudry, B. A. Boone, D. L. Bartlett, H. Zeh and A. H. Zureikat Division Of Surgical Oncology, University Of Pittsburgh Medical Center, Pittsburgh, PA Background: Prognostic scores predicting post-operative pancreatic fistulas (POPF) have been reported and validated for open pancreaticoduodenectomy (OPD). However, no reports have outlined the risk factors of POPF in robotic pancreaticoduodenectomy (RPD). Our objectives were to identify independent risk factors for POPF in patients undergoing RPD and to validate the prognostic scores previously described for OPD. Methods: A prospectively-maintained single institutional database of patients that underwent RPD (2008–2013) was analyzed. POPF was classified according to ISGPF criteria. Demographics and peri-operative outcomes where compared in patients with and without POPF. Univariate and multivariate analysis (UVA/MVA) with logistic regression modeling was used to identify independent variables for POPF. Braga and Callery scores were calculated and tested with logistic regression analysis. Results: 150 consecutive RPDs were analyzed. POPF occurred in 26 (17.3%); 13 (8.6%) of which were ISGPF category B and C. On UVA, patients with POPF had larger body mass index (BMI), smaller duct, smaller tumor size, longer OR time, larger estimated blood loss (EBL) and RBC transfusion (all p < 0.05). On MVA, EBL, small duct size, larger BMI and small tumor size remained the best independent risk factors for POPF after RDP (all p < 0.001). Increased Callery (OR: 1.520, 95% CI: 1.065–1.608, p = 0.01) and Braga (OR: 1.2, 95% CI: 1.064–1.430, p = 0.005) scores predicted an increased risk of POPF following RPD. Conclusion: In this series of RPD, incidence of POPF was similar to that reported for historical OPD. Larger BMI, higher EBL, smaller tumor size, and smaller duct are the main predictors of POPF in RPD. Callery and Braga scores have predictive value for POPF in RPD.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

94

Abstracts

MO-B.11 HOSPITAL READMISSION AFTER PANCREATICODUODENECTOMY IN A HIGH VOLUME CENTER MO-B.10 IN-HOSPITAL MORTALITY FOLLOWING PANCREATICODUODENECTOMY J. He, M. Weiss, J. Cameron, N. Ahuja, K. Hirose, T. Pawlik and C. Wolfgang Department Of Surgery, Johns Hopkins Medical Institutions, Baltimore, MD Introduction: This aim of this study is to investigate causes of in-hospital mortality after pancreaticoduodenectomy in a single institution. Methods: We reviewed clinical data of 3890 consecutive patients who underwent pancreaticoduodenectomy from 1990 to 2011. In-hospital mortality was defined as death during the index admissions for pancreaticoduodenectomy. Patients who survived the index admission were used as comparison. Results: The demographics and the median Charlson AgeComorbidity Index (CACI) were summarized in Table 1. The distribution of pathology was similar in two cohorts. The overall in-hospital mortality was 1.6% (61 of 3890). Two patients died during the pancreaticoduodenectomy due to severe bleeding and cardiac arrest, while the other 59 died at a median of 18 days (range 0–120). In-hospital mortality cohort is older than other cohort (p < 0.001). It had higher rate of complications such as pancreatic fistula, postop bleeding, sepsis related liver failure, bowel ischemia, vascular occlusion related liver failure, abdominal compartment syndrome and myocardium ischemia (p < 0.001). The percentage of patients underwent vascular resections or repair or redo exploratory laparotomy was also significantly higher (p < 0.001). Incidence of in-hosptial mortality in 1990s was similar to that in 2000s (1.8% vs 1.4%, p = 0.36). Conclusion: Although causes for in-hospital mortality are multifactorial, pancreatic fistula and vascular complication related ischemia are highly associated with in-hospital mortality.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

J. Rose, F. B. Rocha, A. Alseidi, T. B. Biehl and S. Helton Virginia Mason Medical Center, Seattle, WASHINGTON Background: Thirty day hospital readmission is a new quality metric that will be used to reduce hospital reimbursement in the future. The aims of this study were to identify predictors of readmission following pancreaticoduodenectomy and to estimate health care resource dollars saved from avoidable readmission. Methods: The electronic medical record and hospital financial database were utilized to retrospectively identify all patients undergoing pancreaticoduodenectomy between 2008 and 2012 in a high-volume tertiary hospital. Patient demographics, hospital charges, perioperative outcomes, and discharge disposition were determined. Readmission rate, corresponding etiologies, and associated charges were established. Predictors of readmission were determined by multivariate logistic regression. Results: Of the 380 patients that underwent pancreaticoduodenectomy between 2008 and 2012, 69 (18%) were readmitted within 30 days (median 7 days). Readmissions accounted for an additional 458 hospital days (median 5 days per readmission). Age, gender, Charlson comorbidity index, operative indication, length-of-operation, length-of-stay, discharge disposition, payer status, or hospital charges were not predictive of readmission. However, having a complication prior to discharge was a risk factor [OR 2.5 (1.5–4.3); p = <0.001]. Hospital mortality and complication rates were 0.5% and 32% respectively. The most common causes for readmission were: infection (52%), gastrointestinal-related (16%), dehydration (10%), and pain (3%). Twelve readmissions (17%) were determined to be avoidable, accounting for 65 additional hospital days and associated charges totaling $231,000. Discussion: Early identification of patients at risk for avoidable readmission following pancreaticoduodenectomy has great potential to reduce health care expenditures in high volume hospitals performing these complex operations.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

95

MO-B.12 QUANTITY AND QUALITY: A HIGH VOLUME PANCREATIC CENTERS ADHERENCE TO HIGHLY VALIDATED QUALITY MEASURES OF PANCREATIC CANCER CARE M. Ongchin, J. Steve, D. Bartlett, M. Choudry, K. Lee, W. Marsh, A. Tsung, A. Zureikat, H. Zeh and M. Hogg University Of Pittsburgh, Pittsburgh, PA Background: Despite extensive data demonstrating improved outcomes at high volume pancreatic centers, little data examines the heterogeneity amongst institutions. Bilimoria et al developed 50 quality indicators (QI) addressing 5 domains in pancreatic cancer care (PCC). Despite the availability of these validated metrics since 2009, few centers report their performance. We examined our institution’s compliance with these PCC QI, hypothesizing that tracking these measures would be feasible and assure high compliance. Methods: We analyzed a prospectively maintained Pancreatic Quality Improvement database specifically used to track these QI between 4/2011–7/2013. Results: 570 patients were captured, 322 had pancreatic adenocarcinoma: 125 resectable, 25 locally advanced, 172 metastatic. We were >90% compliant with 42/43 of the validated QI. Compliance on QI #26 was 86% (table). In patients with >60days to treatment, 55% were metastatic contemplating therapy. Median time from diagnosis to surgery was 24days. 67% received neoadjuvant therapy; median time to treatment (MTTT) was 22days. 74% received adjuvant therapy; MTTT was 51days. Surgical outcomes were 2.2% peri-operative mortality, 85% R0 resection, 20.5 median lymph nodes, 350 cc median EBL, 459 min median OR time, and 34% readmission. 26% enrolled in one of 11 clinical trials. Our composite score was 9/10 (median = 4 in Bilimoria’s study). The QI database is maintained by 1/5 FTE (8 hrs/wk). Conclusion: Our high volume Pancreatic Center complied with 98% of highly validated QI. Time maintaining this database is feasible, prospectively tracking PCC performance. Institutions performing pancreatic surgery should track and report these indicators as part of their quality measures of outcomes.

MO-B.13 TUMOR STAGE AT PRESENTATION AND ACCESS TO SURGERY IS NOT INFLUENCED BY RACE IN PANCREATIC CANCER PATIENTS Y. Tilahum1, X. Zhu1, S. Litherland1, P. Veldhuis1, D. A. Decker1, S. Eubanks1, A. Simmonds1,2, P. Arnoletti1,2 and S. G. De La Fuente1,2 1 Florida Hospital Orlando, Orlando, FL; 2University Of Central Florida, Orlando, FL Background: Previous studies have shown race-dependent disparities in oncologic outcomes among patients with pancreatic cancer. Proposed explanations include differences in access to care, socioeconomic status, education level, HPB 2014, 16 (Suppl. 1), 1–99

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

96

Abstracts

religious factors, and availability of social support. We analyzed the influence of race in tumor stage at presentation and access to surgery in a large ethnically diverse patient population from the state of Florida. Methods: A prospectively maintained tumor registry of all patients with pancreatic cancer treated at the Florida Adventist Health System from 2001 to 2012 was queried. Outcome variables included tumor stage at presentation and percentage of patients that underwent surgical treatment during the study period. Comparison groups were divided by race and gender. Results: A total 812 patients were seen with pancreatic cancer within the studied period. The majority of patients were Caucasian (Table). There were no significant differences in race (p = 0.81) or gender (p = 0.16) associated tumor stage at presentation. Additionally, a similar number of patients underwent surgical treatment within the different ethnicities (Caucasian 22.3%; Hispanics 22.8%; AfricanAmerican 18.3%, p = 0.70). Conclusion: Despite previous reports showing differences in overall survival rates, stage at presentation and access to surgery appear to be similar between races in a large ethnically diverse -pancreatic cancer patient population.

Results: There were 32800 patients who were on dialysis before surgery. A total of 63 pancreatectomies were performed for patients on dialysis (pancreaticoduodenectomy (n = 34), distal or subtotal pancreatectomy (n = 25) or total pancreatectomy (n = 4)). Median age was 63. Indications were pancreatic cancer (n = 17), peri-ampullary tumor (n = 8), neuroendocrine tumor (n = 5), other malignancy (n = 6), benign pancreatic tumor (n = 6) and other benign diseases (n = 21). Median length of hospital stay was 9 days. Overall morbidity and mortality were 42.9 and 4.8%, respectively. Morbidity rates were 52.9, 36.0 and 0% for pancreaticoduodenectomy, distal or subtotal pancreatectomy and total pancreatectomy, respectively. Major post-operative complications included sepsis (20.6%), blood transfusion (15.9%), deep organ space infection (14.3%), re-operation (6.3%), and cardiac arrest (3.2%). In addition, failure to wean ventilation (20.6%) and re-intubation (11.1%) rates were significant. Conclusion: This is the largest reported series of patients on chronic dialysis undergoing pancreatectomy. In carefully selected cases, pancreatectomy can be performed with less than 5% mortality rates. However, particular attention needs to be paid for post-operative respiratory function.

MO-B.16 ADJUVANT CHEMOTHERAPY PROVIDES BETTER OVERALL SURVIVAL AFTER RADICAL RESECTION FOR PANCREATIC ADENOCARCINOMA S. Hurton, F. MacDonald, S. Ayloo and M. Molinari Dalhousie University , Halifax, NS

MO-B.15 PANCREATECTOMY IN PATIENTS ON CHRONIC DIALYSIS: IS IT TOO RISKY? S. W. Cho and J. S. Cardinal Department Of Surgical Oncology, West Virginia University Hospitals, Morgantown, WV Introduction: Need for dialysis treatment is a significant co-morbidity that may exclude patients from pancreatectomy. However, its impact on surgical morbidity and mortality has not been well defined. We sought to quantify complication rates of pancreatectomy in patients on chronic dialysis. Method: Database from the American College of Surgeons National Surgical Quality Improvement Project (ACSNSQIP) (2005–2011) was queried to identify patients on dialysis. Those with acute renal failure or undergoing emergent surgery were excluded. CPT codes were used to select pancreatectomy cases. © 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

Objectives: The expected overall survival (OS) of patients undergoing radical resection for pancreatic adenocarcinoma (PC) is in the range of 10–20%. Several randomized controlled trials comparing adjuvant chemotherapy (AC) versus surgery alone (SA) have shown conflicting results. This study aimed to conduct a systematic review of the literature and perform a meta-analysis (if appropriate) of the effect of AC vs. SA on the OS of patients undergoing radical pancreatic resections. Methods: Seven electronic databases were searched using standard technique to August 2013 without restriction on publication status or language. The primary aim was to assess the pooled effect of AC on the OS after pancreatic resections. A random-effects model was used for the metaanalysis as significant heterogeneity among the included studies was identified (I2 53.2, P = 0.029). Results: Nine Phase III RCTs were included (publication years: 1993–2012). Analysis for publication bias did not show any significant unbalance (Funnel Plot; Figure 1). A total of 665 patients were randomized to AC after surgery while 652 patients were randomized to SA (Table1). Pooled data showed that AC provides a significant OS benefit in comparison to SA: Odds Ratio 0.591 (95% CI 0.388–0.900; P = 0.014) (Forest Plot, Figure 2). Conclusion: The administration of AC provides a significant improvement for the OS of patients undergoing radical PC resections and should be recommended to all patients irrespective of TNM status of their primary disease. Figure 1: Funnel plot Figure 2: Forest plot Table 1: Summary of all the randomized controlled studies included in the metanalysis.

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts

HPB 2014, 16 (Suppl. 1), 1–99

97

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

98

Abstracts

MO-B.18 OBESITY IS AN INDEPENDENT PREDICTOR OF BILE DUCT INJURIES IN PATIENTS UNDERGOING CHOLECYSTECTOMY H. Aziz, A. Amini, V. Pandit, T. Jie and E. Ong University Of Arizona, Tucson, AZ

MO-B.17 LEARNING CURVE FOR ROBOTIC DISTAL PANCREATECTOMY M. Shakir, M. Khreiss, P. M. Polanco, B. A. Boone, M. Zenati , M. Hogg, M. H. Choudry, D. L. Bartlett, H. Zeh and A. H. Zureikat University Of Pittsburgh Medical Center, Pittsburgh, PA Background: Robotic Distal Pancreatectomy (RDP) has been reported as equivalent or superior to laparoscopic distal pancreatectomy and is being performed with increasing frequency, however no data on how many cases are required to attain clinical proficiency has been reported. Understanding the “learning curve” would facilitate dissemination and adoption of this innovative surgical approach. We hypothesized that we could identify inflexion points and plateaus in important metrics of surgical outcomes that correspond to maximal efficiency. Methods: We evaluated metrics of peri-operative safety and efficiency in our prospectively maintained database of RDP. Patients with involvement of the spleno-portal junction or concomitant involvement of other organs were excluded. Patients were divided chronologically into 8 groups of 12 patients each for analysis. ANOVA, Kruskal Wallis, and Fisher Exact tests were used in comparing the groups. Results: 100 patients underwent RADP from 1/2008 to 8/2013 with no 30-day mortality. Average age was 59.7 years. The 8 groups demonstrated equivalent demographics, BMI, Charlson Co-morbidity index, tumor size and pathology. 68 patients (70.8%) underwent resection for malignancy. There were 2 conversions to an open operation (2 %). Operative time reduced significantly after the first 36 patients from 303 minutes to 215 minutes (p < 0.001). The rate of administration of blood transfusion declined after 24 patients (p < 0.05). No significant trend was noted for conversion rate, operative blood loss, margin positivity and length of stay in the hospital. Conclusions: RDP is safe and feasible, with low 90-day morbidity and no 30-day mortality. Reduction in operative times plateau after 36 cases.

Introduction: Iatrogenic bile duct injury is a serious complication of cholecystectomy. The aim of this study was to assess predictors of bile duct injury using a national database. Methods: The Nationwide Inpatient Sample (2009) was queried for cholecystectomy. We used a) diagnoses for bile duct injury and b) bile duct injury repair procedure codes (51.36, 51.37, 51.39, 51.71, 51.72, and 51.79) as a surrogate marker for bile duct injuries. Results: A total of 56 patients had bile duct injury. The mean age was 58.2 ± 19.7 years, 53.5% were males, and median Charlson co-morbidity score was 2 [2–3]. Multivariate analysis revealed Obesity (2.8[2.1–4.3]; p-0.03) and emergent cholecystectomy (1.2 [1.1–1.6]; p-0.01) as the independent predictors for bile duct injury in patients undergoing cholecystectomy. Sub-analysis, revealed that obese patients undergoing emergent cholecystectomy were 1.8 folds more likely to have bile duct injury as compared to obese patients with elective cholecystectomy (1.8 [1.45– 2.1]). Results: Our study finds a new association between the obesity and bile duct injuries which has never been reported in literature before. The effect of obesity on outcomes in biliary surgery needs prospective evaluation.

MO-B.19 A SYSTEMATIC REVIEW AND META-ANALYSES OF HEPATIC ARTERY BASED THERAPIES FOR UNRESECTABLE INTRAHEPATIC CHOLANGIOCARCINOMA L. Boehm, T. Thayyil Jayakrishnan, J. Miura, K. K. Turaga, F. M. Johnston, S. Tsai and T. C. Gamblin Medical College Of Wisconsin Department Of Surgery, Milwaukee, WI Introduction: Hepatic artery based therapies (HAT) are offered for patients with intrahepatic cholangiocarcinoma

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association

HPB 2014, 16 (Suppl. 1), 1–99

Abstracts (ICC) that often present as unresectable. The present study is a systematic review and meta-analyses of the comparative effectiveness of HAT – hepatic arterial infusion (HAI), transcatheter arterial chemoembolization (TACE), drugeluting bead TACE (DEB-TACE), and Yttrium90 radioembolization (Y90) for unresectable ICC. Methods: A prospectively registered study protocol (PROSPERO ID – CRD42013004830) utilized specific key words to find peer-reviewed, English articles from PubMed (1990–2013). Eligible articles based on predetermined inclusion-exclusion criteria were used for data extraction. Primary outcome was median overall survival (OS), and secondary outcome was tumor response to therapy (RECIST). Results: Of the 793 articles identified by the search, 20 were selected for data extraction (n = 627 patients). Complete tumor response was reported in two patients (one TACE and one DEB-TACE). Partial response to therapy was highest with HAI (41.3%, range 12–100%, n = 63) vs. DEB-TACE (32.4%, range 4.2–90%, n = 34) vs. Y90 (26.4%, range 10.5– 35.3%, n = 121) vs. TACE (12.6%, range 0–45%, n = 333). Incidence of disease progression was lowest after HAI (12.7%, 3.8–20%) vs. Y90 (14.9%, 4.5–21.7%) vs. DEBTACE (20.6%, 0–29.1%) vs. TACE (25.5%, 0–75%). Highest Median OS was observed for HAI (21, 11–31) months vs. Y90 (14, 9–22) months vs. TACE (13, 6–16) months vs. DEB-TACE (12, 12–13) months. Conclusion: For patients with unresectable ICC treated with hepatic artery based therapies, hepatic arterial infusion (HAI) offers best outcomes in terms of tumor response and overall survival. Cost-effectiveness and optimization by patient-disease characteristics are a focus for future research.

MO-B.20 HOSPITAL READMISSIONS RATES FOLLOWING OPEN AND LAPAROSCOPIC CHOLECYSTECTOMY L. Watkins1, P. Veldhuis 1, S. Eubanks1, P. Arnoletti1,2 and S. G. De La Fuente1,2 1 Florida Hospital Orlando, Orlando, FL; 2University Of Central Florida, Orlando, FL Background: The introduction of laparoscopic techniques has radically changed the management of patients with benign biliary conditions over the past two decades. Welldocumented benefits of minimally invasive approaches include decreased postoperative pain, early discharge from

HPB 2014, 16 (Suppl. 1), 1–99

99

the hospital and quicker initiation of normal activity. Limited data however exists regarding readmission rates following laparoscopic cholecystectomy. In this study, we determined readmission rates over a 3-year period using a large national database. Methods: The National Impatient Sample was queried for patients requiring readmissions 30-days after discharged for the years 2009, 2010 and 2011. Patients were identified based on diagnostic related groups (DRG) associated with both open and laparoscopic cholecystectomy (414–419). Readmission percentages are defined as the percentage of index stays with at least one readmission within 30 days of discharge. Costs reflect the actual costs of care and not the charges the hospital billed for a case. Results: A total 99,842 patients that underwent open cholecystectomy were discharged during the study period, while 889,914 were discharged after undergoing a laparoscopic cholecystectomy. Thirty-day readmission rates were significantly higher in the open group compared to the laparoscopic group (open n = 10,870 [10.8%] vs. laparoscopic n = 65,917 [7.4%], p < 0.0001). The average cost per readmission for patients that underwent an open cholecystectomy was $12,789 compared to $11,073 for patients within the laparoscopic group (p < 0.0001). Readmission rates within the analyzed years remained the same. Conclusion: Although lower than open cholecystectomy, readmission rates following laparoscopic cholecystectomy are relative common. Readmissions following cholecystectomy are associated with substantial costs.

© 2014 The Authors HPB © 2014 Americas Hepato-Pancreato-Biliary Association