Friday, March 13, 2015, 6:30pm–7:30pm Cocktail Video Presentation

ABSTRACTS FRIDAY, MARCH 13, 2015, 7:00AM–8:00AM PRESIDENTIAL PLENARY PP.01 DEFINING THE PRACTICE OF PANCREATODUODENECTOMY AROUND THE WORLD M. T. McMillan1, M. H. Sprys1, G. Malleo2, C. Bassi2, C. M. Vollmer1 1 University Of Pennsylvania Perelman School Of Medicine, Philadelphia, PA; 2University Of Verona, Verona, VERONA Introduction: Pancreatoduodenectomy (PD) is a technically challenging operation characterized by numerous management decisions. We hypothesize that there is significant variation in the contemporary global practice of pancreatoduodenectomy. Methods: A survey with native-language translation was distributed to members of six international GI surgical societies (including AHPBA and IHPBA). Practice patterns and surgical decision-making for PD were assessed. To evaluate global variance, regions were clustered: North America, South America/Mexico, Asia/Australia, and Europe/Africa. Results: Surveys were completed by 864 surgeons, representing six continents and seven languages. Median age and experience were 46 and 14 years. Surgeons performed a median of 13 PDs in the past calendar year, and reported a median career experience of 90, with only 54% surpassing the published learning curve (>60). Significant regional differences were observed for annual and career PD volumes (P < 0.001). Only 4% of respondents practice pancreas surgery exclusively, but 57% perform HPB surgery only – greatest in Asia/Australia (70%, P < 0.0001). Worldwide, the preferred form of anastomotic reconstruction was pancreaticojejunostomy (88%); however, this choice was more common in North America compared with Europe/ Africa (97 vs. 81%, P < 0.0001). Regional variability was also evident in terms of suture technique, stent use/type, drain use/type/number, as well as the use of octreotide, sealants, and autologous patches (P < 0.02 for all). In particular, there were stark differences in practice between North and South American surgeons (Table). Conclusion: Globally, there is significant variability in the practice of pancreatoduodenectomy. Many of these choices contrast with established randomized evidence and may contribute to variance in outcomes.

HPB 2015, 17 (Suppl. 1), 1–81

PP.02 COLORECTAL LIVER METASTASES: DISAPPEARING LESIONS IN THE ERA OF EOVIST HEPATOBILIARY MAGNETIC RESONANCE IMAGING J. W. Owen, K. J. Fowler, M. B. Doyle, N. E. Saad, D. C. Linehan, W. C. Chapman Washington University In Saint Louis, Saint Louis, MO Purpose: Prior studies show that the disappearance of colorectal liver metastases on CT and PET does not imply complete pathological response. Eovist MRI is known to be the most sensitive imaging modality for liver lesions. Our study evaluates “Disappearing Lesions” on Eovist MRI to determine if disappearance predicts pathologic response. Materials and Methods: Retrospective review of hepatic resections for colorectal metastases between 01/2008 and 01/2014 was performed. Patients with pre-neoadjuvant imaging and pre-operative Eovist MR were included. “Disappearing lesions” were lesions on baseline imaging that were not identifiable on pre-operative Eovist MRI. Complete pathologic response was defined on pathology or by no re-development of a lesion with 1 year follow-up if the site was not resected. Persistent disease was defined as viable lesions on pathology, or imaging evidence of recurrence within 1 year. Results: Baseline imaging and Eovist MRI was available for 24 patients. In 24 patients, 198 colorectal metastases were identified on baseline imaging. On pre-operative Eovist MR 82 of the 194 lesions (42%) were “disappearing” lesions. At © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

2

Abstracts

surgical pathology or 1 year follow-up imaging, 37 of 82 lesions (45%) demonstrated viable tumor (16) or recurrence (21). Thirty-five of 82 lesions (43%) were nonviable at pathology (9) or without evidence of recurrence at 1 year (26). 10 lesions were resected, but difficult to correlate with surgical pathology retrospectively and were indeterminate. Conclusion: Despite the sensitivity of Eovist MR imaging, disappearing lesions should be considered for surgical resection as 45% of disappearing lesions were viable on surgical pathology or recurred when not resected.

those with MVI and CD44 or CD133 is 30% and 27.7%; respectively (p < 0.001). Thirteen patients developed tumor recurrence. On multivariable analysis, CD44 and CD133 expression in combination with MVI, were found to be independent predictors of tumor recurrence and overall patient survival (recurrence p < 0.003, OR = 8.05; p = 0.001, OR = 9.5, survival p = 0.001, HR 3.7, p = 0.004, HR 3.2 respectively). Conclusion: The expression of CD44 or CD133, combined with MVI, were independent factors associated with poor outcomes in patients undergoing transplantation for HCC.

PP.03 CANCER STEM CELL MARKER EXPRESSION IN COMBINATION WITH MICROVASCULAR INVASION PREDICT POOR SURVIVAL IN PATIENTS UNDERGOING LIVER TRANSPLANTATION WITH HEPATOCELLULAR CARCINOMA V. Vilchez1, L. Turcios 1, Y. Zaytseva2, E. Maynard1, M. Shah1, M. F. Daily1, C. Tzeng1, D. Davenport1, A. Castellanos1, S. Krohmer 3, R. Gedaly1 1 Department Of Surgery/Transplant Division – University Of Kentucky, Lexington, KENTUCKY; 2Markey Cancer Center – Core Support, Lexington, KENTUCKY; 3 Department Of Radiology – University Of Kentucky, Lexington, KENTUKCY Objective: We investigated the expression of Liver Cancer Stem Cells (LCSC) markers, CD44 and CD133, combined with the presence of microvascular invasion (MVI); as predictors of outcomes in patients undergoing Liver Transplantation (LT) for HCC. Methods: Explanted livers from 95 patients with HCC who underwent LT at the University of Kentucky Transplant Center were analyzed. The expression of CD44 and CD133 was evaluated using immunofluorescence. Results: Median age was 56 + 7 years, and 77 (81%) were male. The most common causes of end-stage liver disease were hepatitis C (50%) and alcoholic liver disease (41%). Forty-one patients had lab MELD score >15. Median follow-up was 64 months. Twenty two (23%) tumors had MVI. Forty three (45%) tumors were positive for CD44, and 33 (34%) for CD133. Overall 1, 3, and 5 year survival rates were 86%, 75%, and 64%, respectively. The combination of MVI and either CD44 or CD133 positivity was associated with significantly worse overall outcomes. The 5-year survival rate of patients with MVI alone is 51% compared to © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

PP.04 INTRAHEPATIC CHOLANGIOCARCINOMA AND GALLBLADDER CANCER: DISTINGUISHING MOLECULAR PROFILES TO GUIDE POTENTIAL THERAPY M. Potkonjak, J. Miura, K. K. Turaga, F. M. Johnston, S. Tsai, K. Christians, T. C. Gamblin Division Of Surgical Oncology, Department Of Surgery, Medical College Of Wisconsin, Milwaukee, WI Chemotherapy regimens according to National Comprehensive Cancer Network (NCCN) guidelines for intrahepatic cholangiocarcinoma (IHC) and gallbladder adenocarcinoma HPB 2015, 17 (Suppl. 1), 1–81

Abstracts (GC) are interchangeable; however, the molecular differences driving tumorigenesis for these cancers remain poorly defined. The present study utilized biomarker analysis of actionable targets for IHC and GC to distinguish them and potentially refine current treatment strategies. 217 IHC and 28 GC specimens referred to Caris Life Sciences between 2009 thru 2012 were evaluated. Specific testing by immunohistochemical analysis for 17 different biomarkers was performed. In the collective cohort (n = 245), actionable targets included: 95% low TS, 82% low RMM1, and 74% low ERCC1, indicating potential susceptibility to fluoropyrimidines/capecitabine, gemcitabine, and platinum agents, respectively. Additional non-NCCN compendium targets included TOPO1 (53.3% high, irinotecan), MGMT (50.3% low, temozolomide), TOP2A (33% high, anthracyclines), and PGP (30.1% low, taxanes). Subgroup analysis by tumor origin demonstrated a differential biomarker expression pattern with a higher frequency of IHC tumors showing low levels of TS (99% vs. 72%, p < 0.0001), and RRM1 (85% vs. 64%, p = 0.021) when compared to GC. Conversely a greater frequency of GC demonstrated high levels of TOPO1 (76% vs. 50%, p = 0.018) versus IHC, indicating a potential increased benefit from irinotecan. Biomarker analysis possesses the capacity to identify additional targets for which established agents are available. Differences in molecular profiles of IHC and GC provide evidence that the two are distinct diseases and require different treatments.

FRIDAY, MARCH 13, 2015, 12:30PM–1:00PM LUNCH VIDEO PRESENTATION VL.01 TRANS-THORACIC MINIMALLY INVASIVE SEGMENT 8 LIVER RESECTION GUIDED BY AUGMENTED REALITY J. Hallet1,2,4, L. Soler1,2, M. Diana2, D. Mutter1,2,3, T. Baumert2, F. Habersetzer2, J. Marescaux1,2, P. Pessaux1,2,3 1 Institut De Recherche Sur Les Cancers De L’Appareil Digestif (IRCAD), Strasbourg, ALSACE; 2Institut Hospitalier Universitaire De Strasbourg, Universitaire Strasbourg, Strasbourg, ALSACE; 3Service De Chirurgie Digestive, Nouvel Hal Civil, Strasbourg, ALSACE; 4 Sunnybrook Health Sciences Centre – Odette Cancer Centre, Toronto, ONTARIO Background: Liver dome tumors are not traditionally amenable to minimally invasive hepatectomy (MIH) due to superior/central location. In order to increase the number of lesions amenable to MIH, new approaches are needed. Additional challenges in MIH include loss of 3D visualization and tactile perception for intraoperative guidance within the intra-hepatic anatomy. Methods: This video includes the use of a pre-operative 3D virtual model and intra-operative augmented reality (AR) navigation to facilitate trans-thoracic MIH of the liver dome.

HPB 2015, 17 (Suppl. 1), 1–81

3

Results: We present a 52 year-old gentleman with a 3 cm isolated hepatocellular carcinoma in segment 8. A transthoracic approach was chosen to allow for MIH. The video begins with presentation of 3D reconstruction and virtual resection planning. Principles of AR are detailed. The surgical steps include positioning and triangulation of thoracic ports under AR guidance, followed by trans-diaphragmatic tumor localization, identification of the phrenotomy site, and planning of margins using intra-operative ultrasound combined with AR. The parenchyma is transected with bipolar radiofrequency ablation and ultrasonic scalpel. After closure of diaphragm, the specimen is extracted through an enlarged thoracic port, and a chest tube is placed. The procedure was well tolerated. The chest tube was pulled on day 3 and discharge occurred on day 4. Conclusion: This video of liver resection for challenging tumor localization illustrates a different, safe, and valuable approach to MIH. It highlights how 3D virtual resection planning and AR can enhance and facilitate complex MIH, thereby easing the transition into the minimally invasive era for liver surgery.

FRIDAY, MARCH 13, 2015, 3:00PM–4:30PM LONG ORAL A – PANCREAS ONCOLOGY LO-A.01 UTILITY OF ESTABLISHING A PANCREAS CANCER SCREENING PROGRAM WITHIN A HIGH VOLUME PANCREATIC CANCER PROGRAM B. A. Krzywda, S. M. Lahiff, D. M. McDowell, B. George, P. S. Ritch, B. A. Erickson, F. M. Johnston, K. K. Christians, D. B. Evans, S. Tsai Medical College Of Wisconsin, Milwaukee, WI Background: Approximately 10% of pancreatic cancer (PC) may be hereditary and screening of high risk individuals has been recommended. Herein we describe the establishment of a comprehensive multidisciplinary screening program. Methods: Screening criteria included the presence of PC in: 2+ first-degree relatives (FDR), or 3+ any degree relatives (ADR), or any known hereditary cancer syndrome with increased PC risk. Lifetime PC risk was calculated using the CancerGene PancPro software. The clinic provided genetic counseling and nutrition/wellness education. MRI imaging was selectively recommended based the lifetime PC risk. Results: Forty-three patients were screened; 65% were female and the median age was 54 (IQR:11). Family history was significant for two FDR in 12 (28%) patients and three ADR in 11 (26%). Median age of the earliest affected family member with PC was 59 (IQR:15). Hereditary cancer syndromes were present in 18 (41%) patients: BRCA1(3), BRCA2(6), MLH1(1), PMS2(1), PALB2(1), ATM(1), CDKN2A(4) and STK11(1). Median PancPro estimated lifetime risk of screened patients was 7% (IQR:6). Twenty-two (51%) of 43 patients had a lifetime risk over 10%. Elevated © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

4

Abstracts

CA19-9 or HbA1c was detected in 2 (5%) and 4 (9%) patients, respectively. Screening MRI was obtained in 37 (86%) of 43 patients and 10 (27%) of the 37 had pancreatic cystic lesions. No patient has undergone surgical resection of a pancreatic lesion. Conclusions: Initiation of a high risk PC screening clinic identifies patients with radiographic or biochemical abnormalities for which surveillance is necessary. Guidelines for the frequency of surveillance and indications for surgery are needed.

LO-A.02 HAS SURVIVAL IMPROVED FOLLOWING RESECTION FOR PANCREATIC ADENOCARCINOMA? A. S. Rosemurgy, R. Klein, C. Ryan, P. Sukharmwala, B. Sadowitz, K. Luberice, S. B. Ross Florida Hospital Tampa, Tampa, FL Introduction: Billions of dollars have been spent on the research and treatment of pancreatic cancer. This study was undertaken to determine if survival after resection of pancreatic adenocarcinoma has been extended over the past two decades. Methods: The SEER database was queried for patients who underwent pancreatectomy for pancreatic adenocarcinoma from 1992 through 2010. AJCC Stage and survival were determined for each patient. Data were analyzed using Mantel-Cox test and linear regression. Significance was accepted at p < 0.05. Results: 15,604 patients underwent pancreatectomy from 1992 through 2010. Survival improved from 1992 through 2010 (p < 0.0001), as denoted in Figure 1 with the patients divided into three cohorts for illustrative purposes (1992–97, N = 1,846; 1998–2003, N = 4,528; 2004–10, N = 9,230). Similarly, median survival increased 1992 through 2010 (14 vs. 15 vs. 18 months for the cohorts, p < 0.0001). However, 5-year survival rates did not change 1992 through 2010 (14.4% vs. 15.2% vs. 17.0% for the cohorts; p = 0.07). More patients (p = 0.007) and relatively more patients (p = 0.004) underwent resections of Stage I and Stage II cancers 2004 through 2010 with commensurately smaller tumors (p = 0.01). Conclusions: From 1992 through 2010, progressively more patients underwent pancreatectomy for pancreatic adenocarcinoma with progressively smaller tumors and earlier stages. These patients lived more years (e.g., improved survival curves and median survival) but without improved 5-year survival, denoting better early and intermediate survival. Early detection, better perioperative care, more efficacious noncurative chemotherapy undoubtedly play a role, but better solutions for long-term survival must be sought.

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

LO-A.03 HIGH-GRADE INTRADUCTAL PAPILLARY MUCINOUS NEOPLASM IS NOT MALIGNANCY N. Rezaee1, J. He1, B. Salman1, R. H. Hruban2,3, J. L. Cameron1, N. Ahuja1,2, A. Lennon1,4, M. J. Weiss1,2, L. D. Wood3, C. L. Wolfgang1,2,3 1 Department Of Surgery, Johns Hopkins Medical Institutions, Baltimore, MD; 2Department Of Oncology, Johns Hopkins Medical Institutions, Baltimore, MD; 3 Department Of Pathology, Johns Hopkins Medical Institutions, Baltimore, MD; 4Department Of Gastroenterology, Johns Hopkins Medical Institutions, Baltimore, MD Background: Since identification of intraductal papillary mucinous neoplasm (IPMN) in 1996, high-grade dysplasia and IPMN-associated invasive carcinoma was used frequently under the umbrella term “malignancy”. We aimed to compare the pathological features and survival outcomes of high-grade IPMN to invasive carcinoma. Patients and Methods: From 1996 to 2013 data of 616 patients who underwent pancreatic resection for an IPMN were reviewed. IPMNs were classified as low/intermediate-, high-grade dysplasia (HGD), and invasive carcinoma. Results: A total of 293 (48%) patients diagnosed with low/ intermediate-grade dysplasia, 140 (23%) with HGD, and 183 (30%) with invasive carcinoma. Actual 5-year survival was 55% for the entire cohort. The median overall survival was 94 months for HGD, which was similar to low/intermediategrade IPMN (118 months, p = 0.07), and superior to invasive carcinoma (29 months, p < 0.001) (figure). Invasive carcinoma was associated with regional lymph node metastasis in 34%, perineural invasion in 38%, and vascular invasion in 38%. In contrast no lymph node metastasis, perineural or vascular invasion was observed after resection of HGD. Compared to invasive carcinoma, HGD was associated with a lower rate of positive margin (38% vs. 24%, p = 0.007). Among patients who had more than 6 months follow-up, the recurrence rate after resection of HGD (16%) was similar to low/intermediate dysplasia (19%, p = 0.50); and was lower compared to invasive IPMN (29%, p = 0.03). Conclusion: IPMN with high-grade dysplasia has a favorable survival outcome and a lower rate of recurrence after resection compared to IPMN-associated invasive carcinoma, and thus should not be considered a malignant entity. HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

5

IRE near the portal vein, and plastic stenting should be considered when performing IRE near the common bile duct. IRE is a potentially crucial tool in the arsenal of surgeons treating otherwise inoperable pancreatic cancer.

LO-A.04 IRREVERSIBLE ELECTROPORATION (NANOKNIFE) FOR PANCREATIC CANCER: A SINGLE INSTITUTION SERIES OF 50 CONSECUTIVE PATIENTS K. Mahendraraj, I. Epelboym, B. Schrope, J. A. Chabot, M. D. Kluger Department Of Surgery, College Of Physicians And Surgeons, Columbia University Medical Center, New York, NEW YORK Introduction: The NanoKnife® irreversible electroporation system (IRE) uses electrical energy to destroy neoplastic tissue invading surrounding neurovascular structures. Large scale IRE for pancreatic cancer has yet to be reported. This study examines a large cohort of IRE-treated pancreatic cancer patients to evaluate the safety of this novel surgical approach. Methods: Data was abstracted on all T3 and T4 pancreatic cancer patients who underwent IRE at a tertiary hepatobiliary unit from 2012–2014. Standard statistical methodology was used. Results: 50 consecutive patients were treated with IRE by 3 pancreatic surgeons, with 36(72%) cases performed by a single surgeon. Mean patient age was 65.8 ± 7.8 years, with 31(62%) male patients. There were 45(90%) adenocarcinoma cases, most commonly involving the pancreatic head (n = 16;32%) or body (n = 16;32%). IRE was used for primary local control in 25(50%) cases and margin ablation in 21(42%). Median survival was 11.8 ± 6.2 months. Median follow-up was 7.8 ± 9.6 months, with length of stay 7.34 ± 5.6 days and readmission rate of 20%(n = 10). 30- and 90-day complication rates were 36%(n = 18) and 6%(n = 3), most commonly portal vein thrombosis(n = 4;8%), intraabdominal collection(n = 3;6%), and anemia requiring transfusion(n = 3;6%). Overall mortality attributable to IRE was 6%(n = 3). 3 additional mortalities were related to disease progression. Conclusions: IRE offers a feasible technique to manage advanced pancreatic cancer. To reduce morbidity and mortality, anticoagulation should be considered when performing HPB 2015, 17 (Suppl. 1), 1–81

LO-A.05 PRETREATMENT SERUM CA 19-9 LEVELS IN PATIENTS WITH LOCALIZED PANCREATIC CANCER TREATED WITH NEOADJUVANT THERAPY M. Aldakkak, K. K. Christians, A. N. Krepline, B. George, P. S. Ritch, B. A. Erickson, F. M. Johnston, D. B. Evans, S. Tsai Medical College Of Wisconsin, Milwaukee, WI Background: Among pancreatic cancer (PC) pts treated with a surgery-first approach, normal CA19-9 levels have been associated with improved survival. The impact of neoadjuvant therapy on this association is unknown. Methods: Localized PC pts with a CA19-9 level prior to neoadjuvant therapy were dichotomized into two groups; low CA19-9 and elevated CA19-9 based on a cutoff of 36 U/mL. Results: CA19-9 was evaluable in 230 pts prior to any treatment; 57 (25%) were low and 173 (75%) were elevated. The median CA19-9 level at diagnosis in low and elevated CA19-9 pts was 14 (IQR:23) and 267 (IQR:594) respectively. Neoadjuvant therapy including successful surgery was completed in 164 (71%) of the 230 patients; 41 (72%) of 57 low CA19-9 and 123 (71%) of 173 elevated CA19-9 pts (p = 0.90). Median survival of all 230 pts was 23.8 months; 36.7 months for the 164 pts who completed all therapy © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

6

Abstracts

including surgery vs. 11.7 months for the 66 pts not resected. Among the 164 pts who completed all therapy, no difference in median survival was observed between low CA199 and elevated CA19-9 pts; 36.7 months vs. 33.1 months, p = 0.89. Conclusions: An elevated CA19-9 at diagnosis did not predict a failure to complete neoadjuvant therapy and was not associated with inferior survival. These data suggest two cautionary notes: an elevated CA19-9 at diagnosis should not be considered synonymous with advanced (non-surgical) disease; and, a low/normal CA19-9 should not be interpreted as a predictor of favorable outcome (and used to justify a surgery first strategy).

Conclusion: This study illustrates criteria for the highest yield of genetic evaluation for high-risk of PDAC. Insurance coverage for unaffected relatives is lacking. Identification of a causative mutation in an affected family member allows for cost-effective targeted testing in at-risk relatives. Individuals with apparently idiopathic pancreatitis, onset of pancreatitis < 30 years, and those with a family history of pancreatitis or PDAC are candidates for genetic evaluation.

FRIDAY, MARCH 13, 2015, 3:00PM–4:30PM LONG ORAL B – LIVER HCC LO-B.04 SURGICAL RESECTION VERSUS ABLATION FOR HEPATOCELLULAR CARCINOMA LESS THAN 3CM: A POPULATION BASED ANALYSIS J. T. Miura, R. T. Groeschl, F. M. Johnston, S. Tsai, K. K. Christians, K. K. Turaga, T. C. Gamblin Medical College Of Wisconsin, Department Of Surgery, Milwaukee, WI

LO-A.06 ONE YEAR EXPERIENCE OF CHARACTERIZATION OF GENETIC RISK IN A HIGH-RISK PANCREATIC CLINIC K. Flores, K. Dinh, E. Rouleau, W. Wassef, J. LaFemina University Of Massachusetts, Worcester, MA Background: Despite the established benefit of early detection in pancreatic cancer (PDAC) prognosis, high-risk pancreatic clinics are less common than their breast or colon counterparts. One of the difficulties of establishing a highrisk clinic is delineating which individuals are “high-risk.” Methods: We retrospectively examined patients referred for genetic counseling for PDAC from January 2009-June 2014. Patients were referred for a personal and/or family history of PDAC or a potential diagnosis of hereditary pancreatitis (HP). Results: 75 patients were referred for genetic counseling; 36 underwent testing. Twelve (33%) mutation carriers were identified, demonstrating a positivity rate higher than in highrisk clinics for other malignancies. The most common reason to decline testing was lack of insurance. 11% of patients with a family history of PDAC were found to carry a mutation. 20% of those a personal history of PDAC were found to carry a mutation. Ten of 43 patients with a personal history of chronic pancreatitis were found to carry ≥1 mutations. Of these, 8 were heterozygous for CFTR mutations, 1 was CFTR homozygous, and 1 was homozygous for SPINK1 mutations. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

Background: Ablation for ≤3 cm hepatocellular carcinoma (HCC) has been demonstrated to be an effective treatment strategy. Whether ablation achieves a similar survival benefit as compared to surgical resection for early stage HCC remains ill defined. The present study sought to examine the outcomes of patients with ≤3 cm HCC following ablation versus resection. Methods: Patients treated by ablation or surgical resection for ≤3 cm T1 HCC were identified from the National Cancer Database (2002–2011). Cox proportional hazards models were used to assess overall survival (OS) between treatment types (ablation vs resection) following adjustment for age, gender, alpha-fetoprotein (AFP), Charlson Comorbidity Score, and cirrhosis. Results: A total of 2,855 patients underwent ablation (n = 1,984) or resection (n = 871) for solitary HCC ≤3 cm. The median age of the collective cohort was 61 (IQR: 55–70) with the majority being male (n = 2,007, 70.1%). Patients treated with ablation as compared to resection had a higher frequency in AFP elevation (46.5% vs 39%, p < 0.01) and presence of cirrhosis (22.2% vs 14.9%, p < 0.01). Unadjusted OS at 3 and 5 years was greater following resection (67%, 55%) versus ablation (52%, 36%, p < 0.01). In multivariable models, resection was independently associated with improved OS (HR: 0.65, 95% CI: 0.51–0.83; p < 0.01). Conclusion: While more invasive, resection of HCC ≤3 cm results in better long-term survival as compared to ablation. Treatment strategies for small solitary HCC should emphasize a resection first approach, with ablation being reserved for patients precluded from surgery. [Correction added on 24 February 2015. Moved “LO-B.04 “SURGICAL RESECTION VERSUS ABLATION FOR HEPATOCELLULAR CARCINOMA LESS THAN 3CM: A POPULATION BASED ANALYSIS” before “LO-B.01”.]

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

LO-B.01 POST-EMBOLIZATION SYNDROME AS AN EARLY PREDICTOR OF LONG-TERM OUTCOME AFTER TRANSARTERIAL CHEMOEMBOLIZATION FOR HEPATOCELLULAR CARCINOMA M. C. Mason1,2, N. N. Massarweh2,3, A. Salami2, M. A. Sultenfuss4, D. A. Anaya2,3 1 Michael E DeBakey Department Of Surgery, Baylor College Of Medicine, Houston, TX; 2VA HSR&D Center For Innovations In Quality, Effectiveness And Safety, Houston, TX; 3Michael E DeBakey Department Of Surgery – Division Of Surgical Oncology, Houston, TX; 4 Department Of Radiology, Baylor College Of Medicine, Houston, TX Introduction: Transarterial chemoembolization (TACE) is commonly used in the management of hepatocellular carcinoma (HCC) patients not candidates for curative resection or as a bridge to transplantation. A common post-TACE complication is post-embolization syndrome (PES). The goal of this work was to evaluate PES as an early, post-procedural predictor of treatment outcome. Methods: A retrospective cohort study of HCC patients who underwent TACE at a tertiary referral center was performed (2009–2014). Patients were excluded if they had other therapies in combination with TACE. Patients were categorized based on whether they presented with PES, defined as fever with or without abdominal pain within 2 weeks of TACE. The primary outcome was overall survival (OS) evaluated using multivariate Cox regression while adjusting for relevant demographic and clinical characteristics. Results: Among 144 patients treated with TACE alone (70.1% ≥65 years and 44.4% Child’s B or C cirrhosis), 36.1% experienced PES. Median follow-up for the cohort was 11.4 (0.6–49.9) months. Median and 3-year OS were 16 months and 18.0% in the PES group versus 25 months and 41.0% in the no PES group (log-rank, p = 0.027). After multivariate modeling, patients with PES had a higher risk of death compared to non-PES patients (Hazard Ratio 1.94; 95% Confidence Interval 1.12–3.34, P = 0.016). Conclusion: Patients who suffer from PES have significantly worse overall survival even after adjusting for important factors. Future investigation into the pathophysiologic mechanism underlying PES may help identify patients at increased risk of death and better select treatment strategies for HCC patients with different biologic behavior.

7

LO-B.02 DISPARITIES IN CARE FOR PATIENTS WITH CURABLE HEPATOCELLULAR CARCINOMA R. S. Hoehn, D. J. Hanseman, K. Wima, A. Ertel, P. L. Jernigan, D. E. Abbott, S. A. Shah Department Of Surgery, University Of Cincinnati School Of Medicine, Cincinnati, OH USA Objectives: The incidence of hepatocellular carcinoma (HCC) is increasing in the United States. The aim of this study was to evaluate the use and outcomes of surgical management of early stage HCC. Methods: The National Cancer Database (NCDB) was queried for all patients diagnosed with curable HCC (Stage I/II) from 1998-2011 (n=43,859). Demographic information, tumor characteristics, treatment trends, and outcomes were assessed. Multivariate logistic regression was used to determine predictors of having surgery and of long-term survival. Median follow-up was 2.5 years. Results: Patients who underwent surgery (n=17,406; 39.7%) received surgical resection (34.6%), liver transplant (28.7%), radiofrequency ablation (27.1%), or other therapies. While surgery was associated with improved survival (HR 0.76, 95% CI 0.72-0.79), it was only performed on 42% of stage I patients and 50% of patients with tumors smaller than 2cm. Asian and white patients were more likely than black patients to undergo surgery, as were patients from zip codes in the highest education and income quartiles compared to the lowest (p<0.001). Predictors for not having surgery included black race, government insurance, and treatment at a non-academic center (Table 1). Postoperative mortality and long-term survival were better for patients with private insurance or those treated at an academic center (Table 1). Conclusion: In this analysis of the NCDB, fewer than half of patients with curable HCC underwent surgery, and this appears to be associated with race, socioeconomic status and access to care. Further research is necessary to understand barriers and disparities in care for curable HCC patients.

LO-B.03 THE MODIFIED RESPONSE EVALUATION CRITERIA IN SOLID TUMORS (MRECIST) PREDICTS SURVIVAL FOLLOWING TRANSARTERIAL CHEMOEMBOLIZATION (TACE) FOR HEPATOCELLULAR CARCINOMA N. S. Haywood, N. S. Haywood, S. Saddekni, K. H. Gennaro, D. T. Redden, J. A. White, S. H. Gray, A. K. Aal, D. E. Eckhoff, D. A. DuBay University Of Alabama At Birmingham, Birmingham, AL Background: Transarterial chemoembolization (TACE) is the most common oncologic treatment administered for the HPB 2015, 17 (Suppl. 1), 1–81

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

8

Abstracts

treatment of hepatocellular carcinoma (HCC) in the US. The impact of TACE-induced HCC tumor necrosis on patient survival is poorly defined. Hypothesis: We hypothesize that survival will be superior in HCC patients with increased TACE-induced tumor necrosis. Methods: All first TACE interventions for HCC performed at a single institution from 2008–2013 were retrospectively reviewed (n = 344, Table 1). HCC tumor response to TACE was quantified via the modified response evaluation criteria in solid tumors (mRECIST) criteria. Differences in survival were compared using the log-rank test. A multivariable analysis of survival predictors was completed with a Cox proportional hazard model. Results: The median survival following TACE treatment for HCC varied according to the mRECIST response (p = 0.012), with the longest survival observed in patients with a complete response and shortest survival in patients with progressive disease (Table 1) Patients with a complete response had the lowest frequency of repeat TACE, and highest probability of receiving a liver transplant (Table 1). The mRECIST score remained significantly associated with survival (p = 0.0005) in a multivariable model of survival predictors controlling for age, gender, race, tumor size and number, and Child’s score. Conclusions: The mRECIST response to TACE in patients with HCC was predictive of survival, the need for repeat TACE, and the probability of receiving a liver transplant. However, the absolute differences in median survival between mRECIST categories were not as large as predicted.

LO-B.05 ROLE OF SURGICAL RESECTION AND LOCOREGIONAL THERAPY IN PATIENTS WITH STAGE 3 HEPATOCELLULAR CARCINOMA A RETROSPECTIVE REVIEW FROM THE NATIONAL CANCER DATABASE R. Seshadri, E. H. Baker, S. W. Ross, M. Templin, R. Z. Swan, J. B. Martinie, D. A. Iannitti Carolinas Medical Center, Charlotte, NC Introduction: In advanced stages, hepatocellular carcinoma (HCC) is often associated with major vascular involvement (cava, portal vein). Our aim was to analyze the role of surgical resection (SR) and locoregional therapy (LRT) in these advanced stage patients to determine if there was a survival benefit. Methods: The study is a retrospective analysis from the Commission on Cancer’s National Cancer Data Base (NCDB) between 1998–2011. 148,882 patients with liver cancer were identified, of which 126,858 had HCC. Of these, © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

64,227 patients (1998–2006) had 5-year survival data available and 16,036 patients had Stage 3A disease based on AJCC classification. Of these patients, 1,346 had SR, 1,003 had LRT and 11,942 patients had neither intervention. Kaplan-Meier curves and log rank tests were used for statistical analysis. Results: 14,291 patients met analysis criteria. Mean age (years) in the SR, LRT and no intervention group were 62.5 64.3 and 64.2 respectively. Most patients were males in all 3 groups (77.5%, 74.5% and 68.1%). Mean tumor size (cm) in the 3 groups was 9.8, 6.4 and 8.4 respectively. SR and LRT were primarily performed in major academic and comprehensive cancer programs compared to community cancer programs and other centers (SR: 93% vs 7%; LRT: 94.6% vs 5.4%). The median 5 year survival (months) was 26.55 in SR; 16.36 in LRT and 4.8 in no intervention group. (p < 0.0001). Conclusion: Surgical resection and locoregional therapy offer a survival benefit in select patients diagnosed with Stage 3 HCC.

LO-B.06 CLINICAL OUTCOMES OF RESECTION VS. ABLATION OF SINGLE NODULE HEPATOCELLULAR CARCINOMA IN THE ERA OF ORGAN SHORTAGE. SHOULD WE EXTEND THE LIMITS? O. C. Kutlu1, M. Wachtel2, S. Dissanaike1 TTU HSC Dept Of Surgery, Lubbock, TX; 2TTU HSC Dept Of Pathology, Lubbock, TX 1

Introduction: Donor shortage is a limitation for transplantation, leaving resection as the modality of choice for HCC. Radiofrequency ablation(RFA) is an option for early HCC patients or for bad surgical candidates. Studies are being published investigating the use of RFA in larger tumors with smaller ablative margins. We aimed to compare RFA to other modalities in a large epidemiological database. Methods: SEER database was queried for patients at least 20 years age, diagnosed between 2004–8, no metastatic disease, with tumors less than 3.5 cm. Table 1 displays data of importance. Binomial logit regression calculated estimates and standard errors of odds ratios of death at 1 y and 3 y, using R3.1.1. Null hypotheses were rejected if P < 0.05/ 2 = 0.025. HPB 2015, 17 (Suppl. 1), 1–81

Abstracts Results: Figure 1 displays the results of binary logit regression. While survival was similar in RFA and Resection at 1 y, at 3 y RFA fared worse. Elevated AFP and age were not significant at 1 y, but were associated with increased odds of death at 3 y. No assumptions of general linear models were found to be violated. No influential outlier groups were seen. Conclusion: Researchers are seeking ways to facilitate resection and RFA in HCC. Studies have been published discussing RFA and margins <1 cm for HCC up to 5 cm’s in size. RFA probes have been shown to produce a cavity size of 4.5 cm’s in diameter. Although 3.5 cm was the largest tumor size in our study, three year outcomes of RFA were significantly inferior to resection and transplantation, reaching almost twice the odds of death.

FRIDAY, MARCH 13, 2015, 3:00PM–4:30PM LONG ORAL C – TRANSPLANT I LO-C.01 SPLIT-LIVER TRANSPLANTATION: REPORT OF RIGHT AND LEFT GRAFTS OUTCOMES FROM A MULTICENTER ARGENTINEAN GROUP E. Halac1, M. Dip1, F. Alvarez2, J. Leiva Espinoza2, E. Quinonez3,4, P. Romero 7, F. Nievas5, R. Maurette6, C. Luque8, D. Matus1, P. Surraco1, M. Fauda5,6, G. Gondolesi7, L. McCormack3, J. Mattera2,4, O. Imventarza1 1 Hospital Nacional De PediatrJuan P Garrahan, CABA, CABA; 2Hospital Italiano De Buenos Aires, CABA, CABA; 3 Hospital Aleman, CABA, CABA; 4Hospital El Cruce Dr. Nestor Carlos Kirchner, Florencio Varela, BUENOS AIRES; 5Hospital Universitario Austral, Derqui, BUENOS AIRES; 6Hospital Britanico, CABA, CABA; 7Fundacion Favaloro, CABA, CABA; 8Hospital General De Ninos Dr Ricardo Gutierrez, CABA, CABA Grafts from split livers constitute an accepted approach to expand the donor pool. There was a significant interest from most of the Argentinean centers to increase in the use of this technique over the last 5 years. Aims: Describe and analyze the outcomes of right (RSG) and left (LSG) grafts from a multicenter study. HPB 2015, 17 (Suppl. 1), 1–81

9

Methods: Multicenter retrospective study including data from 111 recipients of split liver grafts between 1/1/2009– 31/12/2013. Incidence of surgical complications, patient and graft survival, and factors that affected RSG and LSG survival were analyzed. Results: Grafts types were 57 LSG and 54 RSG. Median follow up times for LSG and RSG were 46 and 42 months respectively. The 36-month patient and graft survivals for LSG were 83% and 79%, and for RSG was 78% and 69% respectively. Retrasplantation rates for LSG and RSG were 3.5% and 11% respectively. Artherial complications were the most common cause of early retransplantation (less than 12 months). Cold ischemia time (CIT) longer than 10 hours and the use of high risk donors (older than 40 years or BMI ≥30 or ≥5 days ICU) were independent factors for diminished graft survival in RSG. No analyzed variables were associated with worse graft survival in LSG. Biliary complications were the most frequent in both groups(57% in LSG and 33% in RSG). Conclusions: Partial grafts obtained from liver splitting are an excellent option for patients in need of liver transplantation, and have the potential to alleviate the organ shortage. Adequate donor selection and reducing CIT are crucial for optimizing results.

LO-C.02 INTRAOPERATIVE HYPOTENSION DURING LIVER TRANSPLANTATION IS ASSOCIATED WITH DECREASED ONE YEAR PATIENT AND GRAFT SURVIVAL P. F. Sauer, D. A. DuBay, P. A. MacLennan, J. H. Crawford, J. A. White, S. H. Gray, D. E. Eckhoff University Of Alabama At Birmingham, Birmingham, AL Introduction: Patients with end-stage liver disease experience peripheral vasodilatation and have lower mean arterial blood pressure (MAP) at baseline. However, it is unclear how intraoperative hypotension effects post-liver transplant patient and graft survival. We hypothesized that increased duration of MAP <60 mmHg (MAP <60) was associated with decreased patient survival. Methods: A retrospective study of adult liver transplant patients from a single center was performed. Anesthesia records were used to quantify the intraoperative duration of MAP <60. Patients were stratified into 3 groups based on cumulative duration MAP <60: <30 minutes, 30–59 minutes and 60+ minutes. Blood product usage, length of hospitalization, vasoactive drug administration, and 1-year patient and graft survival were measured. MAP <60 duration groups were compared using ANOVA, Chi-square tests, and LogRank tests for continuous and categorical variables, and survival curves, respectively. Results: Overall 565 patients were included: <30 minutes (N = 461, 81.6%), 30–59 minutes (N = 42, 7.4%), and 60+ minutes (N = 62, 11.0%). Patients in the 60+ minutes group had the highest MELD score (27.7, p = 0.0051), received the most units of pRBCs (5.5, p < 0.0001), FFP (2.96, p < 0.0001), platelets (1.5, p < 0.0001) and intravenous phenylephrine (14487 μg, p < 0.0001), and had the lowest 1-year patient (72.2%, p = 0.0060) and graft survival (72.2%, p = 0.0063)(Table). © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

10

Abstracts

Conclusion: Intraoperative hypotension, i.e., MAP <60, for 60+ minutes was associated with increased intraoperative phenylephrine, blood product usage, and decreased 1-year patient and graft survival. Surgical technique, anesthetic management, and the avoidance of hypotension are critically important for patient outcomes.

mission rates. These data suggest that with the proposed policy of broader sharing, patient outcomes will be impaired, though require more hospital resources.

LO-C.04 SOCIETAL REINTEGRATION FOLLOWING CADAVERIC LIVER TRANSPLANTATION R. P. Kelly, M. Molinari Dalhousie University, Halifax, NS

LO-C.03 A GLIMPSE INTO THE FUTURE: DO WE KNOW WHAT INCREASING NATIONALLY SHARED LIVER ALLOGRAFTS WILL MEAN? A. Ertel, K. Wima, R. Hoehn, D. Abbott, S. Shah University Of Cincinnati Department Of Surgery, Cincinnati, OHIO Purpose: Due to the current geographic disparities in liver allocation, a new policy endorsing broader sharing of liver allografts has been proposed. In this model there will be a substantial increase in the number of nationally shared, deceased donor livers. We aimed to investigate the potential ramifications this new policy may have on patient outcomes and resource utilization following liver transplantation (LT). Methods: The Scientific Registry of Transplant Recipients (SRTR) was queried for all nationally shared LT from 2007– 2011 and linked to the University Health Consortium (UHC) database (n = 12,445). Univariate analysis was used to determine how donor, recipient characteristics, clinical outcomes and hospital resource utilization were associated with national sharing as compared to local and regional sharing. Results: Nationally shared livers have a higher donor risk index (>1.8), and are more likely to be classified as expanded criteria donors compared to regional and locally shared livers. Recipients of nationally shared livers were more likely to be independent, have lower MELD scores, and were less likely to be hospitalized or severely limited at the time of transplant. Nationally shared LTs were more likely to be performed at high volume centers, resulted in higher transplant hospitalization costs, in hospital mortality, and 30-day readmissions (Table 1). Conclusion: Nationally shared donor livers are of lesser quality, and are being transplanted into healthier recipients. Despite this, they have increased mortality, cost, and read© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

Background: Data related to social reintegration following Orthotopic liver transplantation (OLT) is lacking. For the purpose of this study “social re-integration” was defined as: the ability of an individual to return to the same marital, educational, and financial status they held pre-OLT. The primary aim of this study was to assess the rate of social re-integration following OLT in Atlantic Canada. Methods: Between September 2006 and January 2008 patients that were ≥3 years post OLT were included in this prospective cohort study. Externally validated Questionnaires were self administered to those who participated. All data was then prospectively collected. Categorical variables were analyzed by Chi-Square or Fisher’s exact test and continuous variables by Student’s t-test. Results: Among 157 eligible patients 110 (70%) participated. The mean age was 57 years (SD + 11.4). Prior to OLT, 75% were married and 6% were divorced. Following OLT there was no significant difference in marital status. Employment rate fell from 72% pre-OLT to 30% post-OLT. Lowskill employees were less likely to return to work post-OLT compared to than those who were self-employed. Personal income fell an average of $4,363 (SD + 20,733) (p = 0.03). Nevertheless, 38% of patients reported improved earning potential post-OLT. Overall, 80% of participants expressed satisfaction for their role in society post-OLT. Conclusion: Following OLT, few patients are able to fully reintegrate into the positions and status held pre-OLT. Employment status is most likely to be negatively affected following OLT. These findings suggest that focused strategies for preserving social status be implemented before and after OLT.

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

LO-C.05 DEFINING THE INFLUENCE OF GENDER DISCORDANCE ON SURVIVAL FOLLOWING LIVING DONOR LIVER TRANSPLANTATION J. T. Miura, A. Fathi, F. M. Johnston, S. Tsai, K. K. Christians, K. K. Turaga, T. C. Gamblin Medical College Of Wisconsin, Department Of Surgery, Milwaukee, WI Background: Donor-recipient gender discordance has been suggested to be an independent predictor of poor outcomes following deceased donor liver transplantation (DDLT). Whether a similar trend occurs with living donor transplants (LDLT) remains ill defined. The present study sought to evaluate the association between gender mismatch and survival following LDLT. Study Design: Patients that underwent LDLT were identified from the Organ Procurement and Transplant Network database (2000–2012), and categorized by the following donor-recipient gender groups: Female (F)/F, Male (M)/F, M/M, and F/M. Log rank test and multivariable Cox proportional hazards models were used to assess graft survival (GS) and overall survival (OS) of the study groups. Results: A total of 3,143 LDLT were identified: 731 F/F, 794 M/F, 1,015 M/M, and 603 F/M. Ten year GS and OS were highest in the F/F group, at 62% and 71% respectively, as compared to M/F (53%, 66%), M/M (55%,62%), and F/M (55%,66%). Multivariate analysis demonstrated that gender mismatch was not associated with poorer GS (MM: Reference; MF: HR 1.22, p = 0.06; FF: HR 1.02, p = 0.86; FM: HR 1.18, p = 0.18). Independent factors associated with poorer GS included older recipient age (HR: 1.01, 95% CI: 1.00–1.02; p = 0.01), increasing donor age (HR: 1.01, 95%CI: 1.00–1.02; p = 0.01), higher Model for End-stage Liver Disease score (HR: 1.02, 95%CI: 1.01–1.04; p = 0.003), and left lobe allografts (HR: 2.08, 95%CI: 1.55– 2.79; p < 0.001). Conclusions: Donor-recipient gender discordance does not impact both GS and OS following LDLT. Living donor organ allocation in an era of supply limitation should not be influenced by donor gender.

LO-C.06 RECCURRENT HEPATOCELLULAR CANCER AFTER LIVER TRANSPLANTION: THE ROLE OF LIVER-DIRECTED THERAPY V. Donchev, G. Voidonikolas, M. Sheckley, A. Annamalai, I. Kim, A. Klein, A. Wachsman, M. Friedman, S. Colquhoun, N. Nissen Cedars-Sinai Medical Center, Los Angeles, CA Introduction: Recurrence of hepatocellular cancer (HCC) after liver transplant (LT) generally carries a poor prognosis. We reviewed our experience to determine the role of liverdirected treatments. Methods: Retrospective review of 10-year single center experience. All patients were within radiographic Milan criteria for HCC at the time of LT. Of 201 patients undergoing LT for HCC, 29 (14%) were identified with HCC recurrence. Results: Mean and median time to recurrence after LT was 24 +/− 4 mo. and 15 mo. (range 3–160 mo.) respectively. Explant pathology showed that most patients (26/29) had pathologic staging exceeding Milan criteria. The initial HPB 2015, 17 (Suppl. 1), 1–81

11

recurrence was liver-only in 7 patients (24%), liver-dominant in 5 patients (17%) and systemic in 17 patients (59%). Time to recurrence was greatest in the liver-only group at mean 34 mo., followed by mean 24 mo. in the liver-dominant group and mean 14 mo. in systemic group. Patients with liver-only or liver-dominant tumor underwent a total of 57 loco-regional treatments (range 1–14). There were 3 significant treatment-related complications. Survival: Mean survival after tumor recurrence in all patients was 15 +/− 4 mo. The 1 and 3 year survival after recurrence was 62% and 21% respectively. Patients undergoing liver-directed therapy had improved survival compared to the remainder of patients (25 mo. vs 14 mo., p < .05, figure). Conclusions: HCC recurrence after LT is heterogeneous. Almost all recurrences are in patients with poor explant pathology. Aggressive liver-directed therapies are safe and effective in patients with liver-dominant recurrence and can lead to improved survival.

FRIDAY, MARCH 13, 2015, 4:30PM–6:30PM LONG ORAL D – PANCREAS PERIOPERATIVE/TECHNIQUES LO-D.01 IMPROVED PERIOPERATIVE OUTCOMES WITH EPIDURAL ANALGESIA IN PATIENTS UNDERGOING PANCREATECTOMY: A NATIONWIDE ANALYSIS D. E. Sanford, W. G. Hawkins, R. C. Fields Washington University, St. Louis, MISSOURI Background: Despite scant evidence demonstrating benefit, epidural analgesia (EA) is often used for patients undergoing pancreatectomy. We sought to examine the impact of epidural analgesia on postoperative outcomes after pancreatectomy. Methods: We used the 2008–2011 Healthcare Cost and Utilization Project Nationwide Inpatient Sample to examine the effect of EA on perioperative outcomes. Student’s t tests and © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

12

Abstracts

chi-square tests were used for univariate comparisons. Multivariable logistic and linear regression with propensity score matching were utilized for risk-adjusted comparisons. Results: 12,440 patients underwent pancreatectomy. Of these, 1,130 (9.1%) patients received epidural analgesia. Patients who received EA were significantly more likely to be male, undergo pancreaticoduodenectomy, have cancer, and undergo surgery at higher volume centers. By univariate comparison, patients who received EA had significantly lower rates of pneumonia, blood transfusions, and acute renal failure, and this was associated with decreased postoperative length of stay (LOS), lower hospital charges, and decreased postoperative mortality. In multivariate analyses, EA was independently associated with decreased postoperative LOS (adjusted mean difference = −1.2 days, p < 0.01), decreased hospital charges (adjusted mean difference = −$16,814, p < 0.01), and decreased postoperative mortality (adjusted OR = 0.42, p < 0.01). Using 1 : 1 propensity score matching, patients who received EA (n = 1,070) had significantly decreased postoperative LOS (11.0 days vs 12.1 days, p = 0.01), lower hospital charges ($112,086 vs $128,939, p < 0.01), and decreased postoperative mortality (1.5% vs 3.6%, p < 0.01) compared to matched controls without EA (n = 1,070) (Table). Conclusion: EA is associated with improved perioperative outcomes and reduced hospital charges following pancreatectomy. Additional studies are required to fully understand if this relationship is causative.

LO-D.02 IMPACT OF MARGIN STATUS ON SURVIVAL IN MAIN DUCT INVOLVED INTRADUCTAL PAPILLARY MUCINOUS NEOPLASM M. T. Welsh1, A. M. Roch1, J. L. Cioffi1, J. J. Easler2, J. M. DeWitt2, E. P. Ceppa1, M. G. House1, N. J. Zyromski1, A. Nakeeb1, C. Schmidt1 1 Department Of Surgery, Indiana University School Of Medecine, Indianapolis, IN; 2Department Of Medicine, Division Of Gastroenterology, Indiana University School Of Medicine, Indianapolis, IN Background: The natural history of Main Duct (MD)involved Intraductal Papillary Mucinous Neoplasm (IPMN) is poorly understood. We examined the impact of pancreatic margin status on recurrence and survival, and whether there is benefit of total pancreatectomy (TP) over partial pancreatectomy (PP). Methods: A retrospective review of a prospectively maintained database of patients who underwent resection for IPMN at a single academic center (2000–2013) was performed. Results: 152 patients with MD-involved IPMN were included in this study. Of them, 15 underwent TP and 137 PP. Median follow-up was 41 months. There was no difference in surgical (postoperative complications, hospital stay) and long-term outcomes (malignant recurrence, overall survival OS and disease-free survival DFS) between TP and PP groups. In patients who underwent PP, there was no difference in OS according to the presence of IPMN at margin (96 vs 123 months, p = 0.18). However, when present, degree of IPMN dysplasia at the margin (low vs moderate vs high vs invasive) impacted OS (84 vs 8 vs 13 vs 9 months, respectively p = 0.002). In addition, topography of positive margin in the ductal system (main vs branch vs mixed) impacted DFS (19 vs 65 vs 34 months, respectively p = 0.009). Conclusion: Positive margin status in main-duct involved IPMN may have effects on patient survival. Involvement of the main duct at the margin, and higher degrees of IPMN dysplasia at the margin predict a worse survival. Total pancreatectomy may need careful consideration in select patients in these groups.

LO-D.03 INCREASED MORBIDITY AND MORTALITY OF CONCOMITANT COLECTOMY DURING PANCREATICODUODENECTOMY: A NSQIP PROPENSITY SCORE MATCHED ANALYSIS J. W. Harris, J. T. Martin, E. C. Maynard, P. C. McGrath, C. D. Tzeng University Of Kentucky, Lexington, KY Introduction: Select patients with periampullary cancers require concomitant colon resection (CR) during pancreaticoduodenectomy (PD) for margin-negative resections. This study analyzed the impact of concomitant CR on post-PD major morbidity and mortality using a large national dataset. Methods: National Surgical Quality Improvement Program (NSQIP) patients undergoing PD for periampullary cancers (with/without CR) from 2005–2012, were screened. A 4 : 1 © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts propensity score matched analysis was constructed to isolate the impact of CR upon PD. Risk factors for 30-day major morbidity and mortality were analyzed to determine the postoperative sequelae of PD + CR. Results: Of 10,965 PD and 159 PD + CR patients in total, 624 and 156, respectively, were selected for the 4 : 1 matched analysis. PD + CR resulted in significantly higher major morbidity and mortality (50.0% and 9.0%) vs. PD alone (28.8% and 2.9%, respectively, p < 0.001). Multivariate analysis identified the following risk factors for major morbidity after PD: concomitant CR (OR-3.19, p < 0.001), smoking history (OR-1.92, p = 0.005), lack of functional independence (OR-3.29, p = 0.018), cardiac disease (OR2.39, p = 0.011), decreased albumin (per g/dL, OR-1.38, p = 0.033), and longer operative time (vs. median time, OR-1.56, p = 0.029). Independent predictors of mortality included concomitant CR (OR-3.16, p = 0.010), ventilator dependence (OR-13.87, p < 0.001), and septic shock (OR6.02, p < 0.001). Conclusion: Contrary to previous single-institution studies, this propensity score matched analysis using the NSQIP dataset showed that adding CR to PD significantly increased the magnitude of surgery and was an independent predictor of both major morbidity and mortality. Using high-resolution imaging, patients who may need PD + CR should be identified preoperatively, maximally optimized, and referred to expert centers.

13

evaluated. We hypothesize that increased peri-operative risks associated with open pancreaticoduodenectomy (OPD) in elderly patients may be mitigated with TLPD. Methods: A review of prospectively collected data on all pancreaticoduodenectomy (PD) was conducted from August 2008 to January 2014 (N = 756). Elderly patients (Age ≥70 years, N = 281) were compared to non-elderly patients (Age <70 years, N = 475) with respect to risk-adjusted perioperative morbidity and mortality. Differences in outcomes between TLPD (N = 106) versus OPD (N = 175) were evaluated in the elderly subgroup in an intention-to-treat analysis. Results: Elderly patients have increased risk of ICU admission, any cardiac event, and pneumonia following PD compared to non-elderly patients. TLPD in elderly patients is associated with decreased risk of surgical site infection (SSI) and delayed gastric emptying (DGE) as well as decreased estimated blood loss (EBL) and transfusion. Only factors that were statistically significant on multivariate analysis are reported in the results table. All other outcomes (pancreatic fistula, hemorrhage, length of stay, etc.) were not statistically significant on multivariate analysis. Conclusions: Elderly patients undergoing TLPD experience a similar risk of 90-day mortality, ICU admission, and cardiopulmonary events compared to patients undergoing OPD. TLPD does appear to offer benefits to the elderly with respect to decreased EBL, DGE, and SSI.

LO-D.05 FACTORS INFLUENCING FAILURE TO RESCUE AFTER PANCREATICODUODENECTOMY: A NSQIP PERSPECTIVE LO-D.04 LAPAROSCOPIC PANCREATICODUODENECTOMY DOES NOT COMPLETELY MITIGATE INCREASED PERI-OPERATIVE RISKS IN ELDERLY PATIENTS SEEN WITH OPEN PANCREATICODUODENECTOMY M. C. Tee, K. P. Croome, C. R. Shubert, M. B. Farnell, M. J. Truty, F. G. Que, K. Reid-Lombardo, R. L. Smoot, D. M. Nagorney, M. L. Kendrick Mayo Clinic, Rochester, MN Background: The effect of totally laparoscopic pancreaticoduodenectomy (TLPD) in elderly patients has not been HPB 2015, 17 (Suppl. 1), 1–81

P. Varley, A. Tsung University Of Pittsburgh, PITTSBURGH, PA Background: Previous studies have identified structural factors that may impact the failure to rescue rate after pancreaticoduodenectomy (PD). The goal of this study is to identify patient-level factors associated with failure to rescue in order to guide interventions that prevent progression to mortality. Methods: Patients undergoing PD as the primary procedure were identified from the 2005–2012 National Surgical Quality Improvement Project (NSQIP) Participant Data Use (PUF) files. Since NSQIP only measures mortality for 30 days post-operatively, we treated failure to rescue as time to event data and analyzed it using Cox proportional hazards methods. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

14

Abstracts

Results: A total of 14,546 patients were available for analysis. Of these, 1137 (7.8%) experienced only a minor complication while 5321 (36.6%) experienced at least one severe complication. Failure to rescue rate was 0% vs. 5.9%, respectively (p < 0.001). Failure to rescue rates were 1.2%, 4.2% and 18.6% in patients experiencing a total of 1, 2 or 3+ serious complications (p < 0.001). Results from univariable Cox regression were used to build a multivariable Cox model which was refined by AIC criteria. Factors significantly influencing failure to rescue after serious complication included number of complications, resident participation, age, reoperation, and dyspnea (Table 1). Conclusions: Essentially all patients who experience postoperative mortality after PD first had a serious complication. Interestingly, our analysis shows that it is not the first postoperative complication, but instead the accumulation of multiple events that results in failure to rescue. Preventing this progression should be the focus of future quality improvement efforts.

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

LO-D.06 THE IMPACT OF RECENT HOSPITALIZATION ON SURGICAL SITE INFECTION FOLLOWING PANCREATECTOMY D. E. Sanford, R. C. Fields, W. G. Hawkins Washington University In St. Louis, St. Louis, MO Background: Surgical site infections (SSI) are a major cause of increased morbidity and cost after pancreatectomy. Patients undergoing pancreatectomy frequently have had recent inpatient hospital admissions prior to their surgical admission (recent presurgical admission, RPSA), which could increase the risk of SSI. We sought to examine the impact of RPSA on SSI following pancreatectomy. Methods: We used the 2009–2011 Healthcare Cost Utilization Project California State Inpatient Database. RPSA was defined as hospital stays >48 hrs with a discharge date within 30 days prior to admission for pancreatectomy. We used Chi-square tests, Student’s t tests, and multivariable logistic regression. Results: 3,376 patients underwent pancreatectomy, and 444 (13.2%) had RPSA. The most common RPSA diagnoses were pancreatitis (n = 164, 36.9%) and biliary obstruction (n = 255, 57.4%), and 235 (52.9%) underwent an endoscopic procedure during RPSA. 180 (40.5%) RPSAs were to different hospitals other than where patients’ pancreatectomy took place. In univariate analysis, patients with RPSA had a significantly higher rate of postoperative SSIs, and this was associated with longer length of postoperative stay, higher postoperative hospital costs, and increased postoperative 30-day readmission rates (Table). In Multivariate analysis, RPSA was an independent predictor of postoperative SSI (OR = 1.68, p = 0.013), and the risk of SSI increased with increasing RPSA length of stay (OR = 1.07 per day, p = 0.001). Conclusions: RPSA is an important risk factor for SSI after pancreatectomy. Many patients with RPSA are not admitted preoperatively to the same hospital where pancreatectomy occurs; in such circumstances, SSI rates may not be a sole reflection of the care provided by operating hospitals.

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

15

LO-D.07 PANCREATECTOMY WITH VEIN RESECTION/RECONSTRUCTION: TECHNIQUE MATTERS

LO-D.09 DISTAL PANCREATECTOMY WITH CELIAC AXIS RESECTION: WHAT ARE THE ADDED RISKS?

M. M. Dua1, T. B. Tran1, J. Q. Klausner2, K. J. Hwa1, G. A. Poultsides1, J. A. Norton1, B. C. Visser1 1 Stanford University School Of Medicine, Stanford, CA; 2 UCLA Medical Center, Los Angeles, CA

J. D. Beane1, M. G. House1, S. C. Pitt3, E. M. Kilbane1, B. L. Hall3, A. Parmar4, T. S. Riall4, H. A. Pitt2 1 Indiana University School Of Medicine, Indianapolis, IN; 2 Temple University School Of Medicine, Philadelphia, PA; 3 Washington University School Of Medicine, St. Louis, MO; 4University Of Texas Medical Branch, Galveston, TX

Introduction: A variety of techniques have been described for portal vein (PV) and/or superior mesenteric vein (SMV) resection/reconstruction during pancreatectomy. The ideal strategy remains unclear. Methods: Using a prospective database, we identified all patients between 2005–2014 who underwent PV/SMV resection/reconstruction during pancreatectomy (subtotal/ total/whipple) for all diagnoses. We reviewed the medical records and imaging for operative details and outcomes, with special emphasis on patency. Results: Ninety patients underwent vein resection/ reconstruction with one of five techniques:1) longitudinal venorrhaphy (n = 17, 19%); 2) transverse venorrhaphy (n = 9, 10%); 3) primary end-to-end (n = 28, 31%); 4) patch venoplasty (n = 17, 19%); and 5) interposition graft (n = 19, 21%). With median follow-up (last available imaging to assess patency) of 282 dys, thrombosis was observed in 16/90 (18%). The rate of thrombosis varied according to technique. All patients with primary end-to-end or transverse venorrhaphy remained patent. Longitudinal venorrhaphy, patch closure, and interposition graft were all associated with significant rates of thrombosis (25%, 31%, 44%, respectively, p < 0.01 vs no thrombosis). Comparing those thrombosed to those that remained patent, there were no differences with respect to pancreatectomy type, preoperative knowledge of vein involvement, and neoadjuvant therapy. Patients with thrombosis had significantly longer operative times (480 vs 401 min,p < 0.01) and increased blood loss (1150 vs 600 mL,p < 0.05). Post-operative heparin drip was used in only 7%. Prophylactic aspirin was used in 69% of the total cohort (66% of patent, 81% of thrombosed) and showed no protective benefit. Conclusions: Primary end-to-end and transverse venorrhaphy have better patency than the alternatives after PV/SMV resection and should be the preferred techniques for short (<3 cm reconstructions).

Background: Surgeons have become aggressive at operating on tumors of the body of the pancreas which require resection of the celiac axis (Appleby procedure). Reported series are small and not adequately controlled. The aim of this analysis was to report a large series of Appleby procedures with a comparison group to determine the relative risk. Methods: Data were gathered through the American College of Surgeons-National Surgical Quality Improvement Program, Pancreatectomy Demonstration Project. Over 14 months, 822 patients underwent a distal pancreatectomy (DP) at 43 institutions. Twenty of these patients (2.4%) also underwent celiac axis resection (CAR). Appleby procedure patients were then matched by age, gender, BMI, serum albumin, ASA class, gland texture, duct size and pathology to 180 patients undergoing DP without CAR. Operative and postoperative outcomes were compared by Fisher’s Exact and Wilcoxon tests. Results: The median age of the DP and DP + CAR patients was 65 and 64 years, respectively. Most patients were female (67 and 70%). The mean BMI of the two groups was identical (27.1 kg/m2). The majority of patients had adenocarcinomas (63 and 60%) or neuroendocrine tumors (13 and 15%). Operating Room (OR) and postoperative outcomes are presented in the table. Conclusions: Distal pancreatectomy with celiac axis resection is associated with increased operative time, renal failure and a 10% operative mortality. The decision to offer an Appleby procedure should be made with full disclosure of the increased risks.

LO-D.10 SURGICAL MANAGEMENT OF CHRONIC PANCREATITIS: A THERAPY IN DECLINE? L. A. Bliss, C. J. Yang, M. Eskander, S. DeGeus, M. P. Callery, T. S. Kent, A. J. Moser, S. D. Freedman, J. F. Tseng Beth Israel Deaconess Medical Center, Boston, MA Background: Surgical intervention is uncommon in chronic pancreatitis management. Literature largely describes single institution or international experiences. This study describes US-based chronic pancreatitis surgical management and its impact on readmission rates. HPB 2015, 17 (Suppl. 1), 1–81

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

16

Abstracts

Methods: Retrospective analysis of chronic pancreatitis patients in Healthcare Cost and Utilization Project Florida State Inpatient Database 2007–2011 using revisit variables. Patients with malignancy or congenital abnormalities excluded. Surgical interventions and complications of chronic pancreatitis identified using ICD-9 codes. Univariate analysis of sex, Elixhauser score, race, insurance, complications, receipt of surgery by chi-square. Number of readmissions, time to surgery by Wilcoxon rank sum. Multivariate analysis of operative management by logistic regression. Results: 21,448 chronic pancreatitis patients. 6.18% (1,325) underwent surgery including 629 with drainage procedures, 275 with pancreatectomies, 735 with cholecystectomies. Procedures decreased from 8.65% in 2007 to 3.10% in 2011 (p < 0.0001). 12.95% (2,778) developed pancreatitis-related complications: pancreatic cysts or pseudocysts (4.40%), diabetes (10.02%). Pancreatic exocrine insufficiency in <11 patients. Median number of readmissions 1 (IQR 0–4) and 5 (IQR 2–9) among non-surgical and surgical patients, respectively (p < 0.001). Median number of admissions prior to pancreatectomy was 2 (IQR 1–4) and drainage procedure was 2 (IQR 1–6). Predictors of surgical intervention displayed in table. Conclusions: Chronic pancreatitis leads to numerous inpatient readmissions and difficult to manage complications. Surgical intervention occurs in a declining minority of cases. Complicated patients are more likely to undergo surgery; operative patients experience more admissions than nonoperative patients. The complexities of chronic pancreatitis management warrant early multidisciplinary evaluation and ongoing consideration of surgical and nonsurgical options.

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

FRIDAY, MARCH 13, 2015, 4:30PM–6:30PM LONG ORAL E – LIVER ONCOLOGY LO-E.01 NEOADJUVANT CHEMOTHERAPY DOES NOT IMPAIR LIVER REGENERATION FOLLOWING MAJOR HEPATECTOMY OR PORTAL VEIN EMBOLIZATION FOR COLORECTAL LIVER METASTASES E. Simoneau1, N. Molla1, R. Alanazi2, J. Alshenaifi2, M. Aljiffry3, A. Medkhaly2, L. Boucher1, P. Metrakos1, M. Hassanain1,2 1 Department Of Surgery, McGill University, Montreal, QC; 2 Department Of Radiology, McGill University, Montreal, QC; 3Department Of Surgery, College Of Medicine, King Saud University, Riyadh, SA; 4Department Of Surgery, Faculty Of Medicine, King Abdulaziz University, Jeddah, SA; 5Department Of Oncology, McGill University, Montreal, QC Introduction: Treatment strategies for colorectal liver metastasis (CRCLM) such as major hepatectomy and portal vein embolization (PVE) rely on the regenerative capacity of the liver. Neoadjuvant chemotherapy is most often used with patients undergoing these procedures. We aimed to investigate the effect of neoadjuvant chemotherapy on liver regeneration after PVE and after major hepatectomy. Methods: All CRCLM patients undergoing PVE or major resection (without PVE) with 3D liver volumetry measurements were included. Liver regeneration (expressed as future liver remnant (FLR) and percentage of liver regeneration(% LR)), total liver volume (TLV) and clinical characteristics were collected from our CRCLM database. Results: Between 2003–2013, 226 patients were included (85 major resections, 141 PVE). Mean age was 63 ± 12 years old and median number of cycles was 6(5–8). In each group, overall adequate regeneration was observed (+96.5% in FLR (p < 0.001) post PVE and +45.8% in FLR (p < 0.001) post resection). In the PVE group, chemotherapy variables did not show significant association with the amount of liver regeneration (number of cycles (p = 0.435), timing (p = 0.563), chemotherapy agent (p = 0.116)). Similarly in the major hepatectomy group, neoadjuvant chemotherapy administration did not show a significant association with %LR (p = 0.592) or with other treatment variables (number of cycles, p = 0.114; agent, p = 0.061, timing, p = 0.126). In both groups, the predicted FLR was inversely correlated with the % in liver regeneration only (p < 0.001). Conclusion: Neoadjuvant chemotherapy does not seem to affect the liver regeneration. The predicted FLR only is inversely correlating with the amount of LR occurring after major resection or after PVE.

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

LO-E.02 HIGH-DOSE STEREOTACTIC BODY RADIOTHERAPY (SBRT) FOR PRIMARY AND METASTATIC LIVER TUMORS A. Kirichenko, D. Parda, K. Tom, P. Abrams, J. Oliva, M. Szramowski, N. Thai Allegheny Health Network, Pittsburgh, PA Purpose: SBRT has emerged as an effective non-invasive radiation therapy to precisely target liver tumors with ablative doses of radiation while avoiding surrounding liver tissue. We report our single center experience on the efficacy and tolerability of high-dose SBRT in the management of inoperable primary and metastatic liver tumors from 2007– 2013. Patients and Methods: Patients with 1–4 hepatic lesions and tumor diameter ≤8 cm received liver SBRT of 40–60 Gy delivered in 4–6 fractions. The primary end point was local control with at least 8 months of radiographic follow-up, and secondary end points were toxicity and survival. Results: 68 patients (79 lesions) completed high-dose liver SBRT for HCC (31) or oligo-metastases (37). 22 patients underwent hepatic resections or liver transplant in combination with SBRT. With median followup 22.5 months (range, 8–67 months) overall survival was 78% for patients with hepatic oligometastases and 60% in HCC patients (including 11 patients who completed SBRT prior to liver transplant). No incidence of >grade 2 treatment toxicity or accelerated MELD score migration was observed. Overall local control within radiation field at two years after SBRT was 94% and for lesions with diameter of ≤4 cm was 100%. Conclusion: In this retrospective analysis we demonstrate that liver SBRT is safe and effective for the treatment of hepatic malignancies providing local control rates similar to hepatic resection.

17

ratio(PLR) are adverse prognostic markers for survival in various cancers including colorectal carcinoma. Methods: Data from a multicenter, prospective-randomized DEBIRI(Irinotecan-drug eluting beads) study comparing LC Bead, loaded with irinotecan plus chemotherapy and bevacizumab versus chemotherapy with bevacizumab in unresectable metastatic colorectal cancer was analyzed. Results: There were 30 in the control arm and 41 patients (112 treatments) in the DEBIRI test arm. The median NLR and PLR for the whole cohort was 3.41 (range: 1.1–13.9, IQR: 2.9,4.2) and 187.5 (range: 77–792, IQR: 146.6,227) respectively. NLR and PLR correlated well with adjusted R2 of 0.825 (p-0.000, F = 128). On univariate analysis, an NLR >5(median OS 31.9 vs.14.7 months, p = 0.067) and a PLR >150 (median 31.6 vs. 24 months, p = 0.046) was associated with decreased overall survival (OS). Other factors associated with decreased OS were presence of response (CR/PR or SD) to treatment (16.9 vs. 33.3 months, p0.001), ECOG score of ≥1 (33.7 vs. 17.4 months, p = 0.006), lung metastasis (18.5 vs. 33.6 months, p = 0.016), history of heart disease (16.4 vs. 28.7, p = 0.033), albumin <3.5 (10.5 vs. 31.9, p = 0.001). On multivariate analysis with Cox proportionalhazards model, only response to treatment was an independent predictor of decreased OS (hazard ratio = 0.25, 95%CI: 0.06–0.9, p = 0.044). Conclusion: In this study, response to treatment was the most important favorable prognostic marker in metastatic colorectal carcinoma with liver dominant disease. Patients with high NLR and PLR showed a trend towards worse prognosis, albeit not independently and needs to be investigated in larger studies.

LO-E.03 THE ROLE OF LYMPHOCYTE TO NEUTROPHIL RATIO (LNR) AND PLATELET TO LYMPHOCYTE RATIO (PLR) AS PROGNOSTIC MARKERS IN METASTATIC COLORECTAL CARCINOMA: A REVIEW OF DATA FROM A RANDOMIZED CONTROLLED STUDY P. Philips1, C. R. Scoggins1, R. D. Tomalty2, M. T. Schreeder3, J. Kaugh4, K. Kim4, W. R. Rilling5, C. J. Laing6, C. M. Tatum7, L. R. Kelly7, R. D. Garcia-Monaco8, V. R. Sharma1, R. A. Redman1, T. S. Crocenzi1, S. M. Strasberg1, R. C. Martin1 1 University Of Louisville And James Graham Brown Cancer Center, Louisville, KY; 2Huntsville Hospital Interventional Radiology, Huntsville, AL; 3Clearview Cancer Institute, Huntsville, AL, USA, Huntsville, AL; 4 Winship Cancer Institute – Emory University, Atlanta, GA; 5 Froedtert Medical College, Milwaukee, WI; 6Radiological Associates Of Sacramento (RAS), Sacramento, CA; 7Norton Radiology Associates, Louisville, KY; 8Hospital Italiano, Buenos Aires, ARGENTIAN Background: Recent evidence suggests that elevated neutrophil-lymphocyte ratio(NLR) and platelet-lymphocyte HPB 2015, 17 (Suppl. 1), 1–81

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

18

Abstracts

LO-E.04 MANAGEMENT AND OUTCOME OF COLORECTAL CANCER (CRC) LIVER METASTASES IN THE ELDERLY: A POPULATION-BASED STUDY S. Nanji1,2, W. J. Mackillop2,3,4, X. Wei3, C. M. Booth2,3,4 Department Of Surgery, Queen’s University, Kingston, ON; 2Department Of Oncology, Queen’s University, Kingston, ON; 3Department Of Public Health Sciences, Queen’s University, Kingston, ON; 4Division Of Cancer Care And Epidemiology, Queens University Cancer Research Institute, Kingston, ON

1

Background: Surgical resection is standard treatment for patients with CRC liver metastases (LM). Limited data describe practice and outcomes among elderly patients. We report management and outcomes of surgical resection of CRC LM in the elderly in routine practice. Methods: All cases of CRC in Ontario who underwent surgical resection of LM from 1994–2009 were identified using the population-based Ontario Cancer Registry. We linked electronic records of treatment to the registry to identify surgical procedures and utilization of chemotherapy. Pathology reports provided details regarding extent of disease and surgical procedure. Patients were classified as <65, 65–69, 70–74, and ≥75 years of age. Results: We identified 1310 patients: 710 (54%) <65; 220 (17%) 65–69; 194 (15%) 70–74; and 186 (14%) ≥75 years of age. Mean number of lesions (2.3, 2.1, 1.8, 1.6, p < 0.0001) and mean size of the largest lesion (4.0, 4.3, 4.4, 4.5 cm, p = 0.031) varied across age groups. Elderly patients were less likely to undergo a major liver resection (≥3 segments): 55%, 17%, 15%, 12%, p = 0.10. Peri-operative chemotherapy was used less frequently in the elderly (71%, 63%, 51%, 41%, p < 0.0001). 90-day mortality (2%, 6%, 4%, 8%, p < 0.001) was greatest among patients ≥75 years of age. Overall survival at 5 years across the age groups was 49%, 40%, 47%, 28% (p < 0.0001). Conclusions: Resection of CRC liver metastases is associated with greater risk of post-operative mortality among elderly patients despite less aggressive treatment. Although the long-term outcomes are inferior to younger patients, a substantial proportion of elderly patients will have long-term survival.

LO-E.05 FIBROSES PREDICTS SURVIVAL IN PATIENTS WITH COLORECTAL LIVER METASTASES RESECTED AFTER PREOPERATIVE CHEMOTHERAPY

analyze the incidence of these pathological findings in patients exposed or not to preoperative treatment and, in the group of preoperative treatment, identify predictive factors of pathological response and their impact on survival results. From 2009 to 2012, ninety patients were analyzed, and twenty of them did not have preoperative chemotherapy. Survival analysis only included patients who had preoperative chemotherapy. These patients had higher percentages of ≥25% fibrosis (p = 0.007) and <50% viable cells (p = 0.005). With a median follow-up period of 26.5 months, patients who had preoperative treatment and <25% of fibrosis had a statistically significant worse 3-year overall survival (96.4 × 59.9%, p = 0.019). In multivariate analyses, the only independent prognostic factor for overall survival was ≥25% of fibrosis (HR = 5.5, p = 0.045, 95% CI 1.0– 29.3) and it was correlated with Kras wild type status (HR = 4.1, p = 0.032, 95% CI 1.1–15.4). No pathologic findings could predict disease-free survival results. Percentage of fibroses seems to be the best predictor of survival among response variables.

LO-E.06 CYTOREDUCTIVE SURGERY AND HYPERTHERMIC INTRAPERITONEAL CHEMOTHERAPY IN PATIENTS WITH LIVER INVOLVEMENT

M. C. Marques, H. S. Ribeiro, R. S. De Souza, W. L. Da Costa, A. L. Diniz, A. L. De Godoy, I. C. De Farias, M. F. Begnami, R. Chojniak, V. H. Fonseca, R. S. Bonachi, F. J. Coimbra A.C. Camargo Cancer Center, Sao Paulo, Brasil

Y. Berger1, S. Aycart1, P. Tabrizian1, J. Mandeli2, S. Hiotis1, U. Sarpel1, D. Labow1 1 Department Of Surgery, Division Of Surgical Oncology, Mount Sinai Medical Center, New York, NEW YORK; 2 Department Of Preventive Medicine, Mount Sinai School Of Medicine, New York, NEW YORK

Pathological response to preoperative chemotherapy in patients with resected colorectal liver metastases has been identified as one of the most powerful predictors of outcomes. Fibrosis, necrosis and percentage of viable cells are recognized as markers of response, but these changes are also present in liver specimens from patients who didn’t have preoperative treatment. The aim of this study was to

Background: The aim of this study was to examine the perioperative and long-term results in patients with liver involvement undergoing cytoreductive surgery and hyperthermic intraperitoneal chemotherapy (CRS/HIPEC) at a single tertiary institution. Methods: Data were obtained from a prospectively collected database maintained between March 2007 and July

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts 2014. All patients undergoing CRS/HIPEC were divided into patients who underwent synchronous liver resection (group 1) and those who did not (group 2). Perioperative and longterm results were compared between the groups. Results: Out of 269 CRS/HIPEC procedures, group 1 included 103 procedures (38%) performed in 101 patients with primary diagnoses of colorectal cancer (CRC, 27%), appendiceal cancer (33%), pseudomyxoma peritonei (13%) and other malignancies. Most patients (77%) in group 1 underwent stripping of the liver capsule or minor parenchymal resection, whereas 24 patients (23%) underwent parenchymal resection of at least one liver segment. Group 1 had longer duration of surgery (p < 0.001), increased intraoperative blood loss (p = 0.004), higher number of organs resected (p < 0.001), longer hospital stay (p = 0.002) and more ICU admissions (p = 0.02). Major complication and 90-day mortality rates were not significantly different between the groups. Median overall survival for CRC patients in group 1 was poorer (21.8 months, vs. 43.8 months in group 2, p = 0.01). Conclusions: CRS/HIPEC procedures with synchronous liver resection are prolonged, challenging and often require multivisceral resections in order to achieve optimal cytoreduction, but are safe even when substantial parenchymal resection is being performed. Liver involvement at the time of CRS/HIPEC is a marker of poor survival in CRC patients.

19

reported underutilization of operative treatment in patients with colorectal metastases to the liver. The current study examines trends in hepatic resection and survival among Medicare recipients with hepatic metastases. Methods: Medicare recipients with incident colorectal cancer diagnosed 1991–2009 were identified from the linked SEER-Medicare dataset. Patients were stratified into historical control (1991–2001) and current cohort (2002–2009). Analyses compared rates of hepatectomy and peri-operative morbidity and mortality. Cox proportional hazards model tested effects of clinically relevant variables on overall survival. Results: Of 31,574 patients with metastatic colorectal cancer to the liver, 14,925 were in the current cohort treated after 2002 and 16,649 comprised the historical control group. The rate of hepatic resection increased from 6.1% pre-2002 to 7.9% currently (p < 0.001). The proportion of patients treated with major (>3 segments) or minor hepatectomies did not change (p = 0.345). Peri-operative morbidity including hemorrhage, infections, or gastrointestinal complications did not differ (all p > 0.089). 30-day mortality did not differ between historical controls (4.5%) and current cohort (3.7%), p = 0.329. After adjusting for significant effects of age and Charlson comorbidity index (both p < 0.001), both use of hepatic resection (HR = 0.40, 95%CI: 0.38–0.42, p < 0.001) and current treatment after 2002 (HR = 0.93, 95%CI: 0.91–0.96, p < 0.001) were associated with improved overall survival. Conclusions: Despite statistical significance, clinical improvement in the rate of hepatic resection from 6.1% to 7.9% is minor. As survival continues to improve, ongoing efforts should focus on improving access to operative treatment for patients with hepatic metastases.

LO-E.08 SURVIVAL OF PATIENTS AFTER PORTAL EMBOLIZATION FOR LIVER METASTASES OF COLORECTAL CANCER Y. Collin, R. Huang, M. Plasse, R. Letourneau, M. Dagenais, S. Turcotte, A. Roy, R. Lapointe, F. Vandenbroucke-Menu Centre Hospitalier Universitaire de Montreal, Service de Chirurgie HBP, Montreal, QC

LO-E.07 ACCESS AND SURVIVAL: A POPULATION-BASED STUDY OF CURRENT OUTCOMES AFTER HEPATIC RESECTION IN PATIENTS WITH METASTATIC COLORECTAL CANCER V. M. Zaydfudim1, T. L. McMurry2, A. M. Harrigan1, C. M. Friel1, G. J. Stukenborg2, T. W. Bauer1, R. B. Adams1, T. L. Hedrick1 1 University Of Virginia, Charlottesville, VIRGINIA; 2 University Of Virginia, Charlottesville, VIRGINIA Background: Despite compelling institutional data supporting hepatic resection, population-based studies historically HPB 2015, 17 (Suppl. 1), 1–81

Introduction: Portal vein embolization (PVE) is an efficient method used to induce liver hypertrophy in cases of otherwise unresectable liver tumors, but its impact on survival is debated. This study compares survival and peri-operative data between patients resected with prior PVE (group PVE) and without (group NoPVE). Methods: This is a prospective study 128 patients with colorectal liver metastases (CRLM): 71 with PVE and 57 without. Results: The groups were without significant difference for age, sex and prior medical history. The PVE group had more lesions (3.30 VS 2.84; p < 0.001). Within the PVE group, operative time was longer (221.31 VS 186.54 min; p = 0.01). There was no statistical difference between the groups for blood loss (756.25 VS 604.02 mL; p = 0.078). There was no © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

20

Abstracts

difference in complication rates between the groups (p = 0.465) and there was no peri-operative mortality. Hospital stay was shorter in the PVE group (7.84 VS 8.95 d, p = 0.038). 5-year overall survival rate and disease free survival rate were similar between groups (53.2% VS 54.4%; p = 0.999 and 38.3% VS 40.6%; p = 0.823) Median overall survival and disease free survival were 65.4 months and 33.6 months respectively, with a median overall follow-up of 39.3 months. Conclusion: Our results showed similar survival whether PVE was used or not. Hence, PVE offers a chance for cure for patients who could not be operated upfront. Moreover, PVE patients seem to have the same prognosis even with a larger extent of disease.

IRB approved and distributed to 42 high volume centers in Canada, US, Mexico, and South America. Data were collected, collated and analyzed. Results: Complete surveys were returned from 23 institutions, including representatives from each region (Canada, US, Mexico, and South America). All of the institutions responding performed PVE with 5 centers (21.7%) also performing ALPPS procedures. In the previous year, the average PVE and ALPPS procedures performed were 15.75 and 6.2 per instutition respectively. Only 18 (78.3%) reported the capability to extend PVE to segment 4, and 12 (52.2%) reported embolization utilizing embolic microspheres. Twenty respondents (87%) rated PVE the safest option for liver hypertrophy; however, 12 respondents (52.2%) believe the ALPPS procedure is most likely to result in adequate hypertrophy. Conclusions: There exists extreme variability in utilization and attitudes toward the available techniques for FLR volume augmentation. Penetration of best practice techniques for PVE is lacking and may contribue toward the attraction of the riskier ALPPS procedure.

FRIDAY, MARCH 13, 2015, 5:00PM–6:30PM ORAL POSTER I (BILIARY, EDUCATION, LIVER) OP-I.01 T2 GALLBLADDER CANCER – STILL A NODAL DISEASE M. R. Sheikh, H. Osman, S. Cheek, S. Hunter, D. R. Jeyarajah Methodist Dallas Medical Center, Dallas, TX

LO-E.09 EVALUATING ATTITUDES TOWARD AND APPLICATION OF MECHANISMS TO AUGMENT THE LIVER IN NORTH AND SOUTH AMERICA (THE MALINSA SURVEY) R. W. Day, C. Conrad, J. Vauthey, T. A. Aloia Department Of Surgical Oncology, The University Of Texas MD Anderson Cancer Center, Houston, TEXAS Introduction: Various techniques, including PVE, PVL, and ALPPS, are being used to augment the future liver remnant volume in preparation for major hepatectomy. Given the significant variation between and within these techniques, there is no scientific way to compare their safety and efficacy. Purpose: The aim of this study was to survey and document the availability, variation, utilization and attitudes toward each of these techniques across high volume HPB centers in North and South America. Method: A descriptive 20 question survey was developed and internally validated with expert review. The survey was © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

Introduction: Treatment of gall bladder cancer (GBC) has traditionally been viewed with pessimism and lymph node positivity has been associated with worse prognosis. The aim of this study is to analyze lymph node positivity in patients with T2 tumors. Methods: All patients who underwent surgery for GBC between September 2005 and June 2014 have been identified retrospectively in our database. Data collected included clinical presentation, operative findings, and histopathological data. Results: Charts of 36 patients were reviewed. 26 patients had incidental GBC diagnosis following cholecystectomy. 10 patients were T2 on initial cholecystectomy and all underwent subsequent radical resection. 2 patients from this group were N1 on initial cholecystectomy and 4 more patients became N1 on second surgery. Overall 60% patients with T2 disease had node positivity. 2 patients were found to have residual disease at the liver margin and were upgraded to T3 following resection, one of them also had N1 disease. Overall 50% patients with T2 disease had stage upgrade after radical resection. 10 patients were diagnosed on imaging. 3 of these patients were unresectable and 6 were either stage T3 or higher or node positive. Conclusions: 60% of T2 GBC were node positive in our experience. 50% T2 patients found on initial HPB 2015, 17 (Suppl. 1), 1–81

Abstracts cholecystectomy have stage upgrade as a result of radical surgery. These findings support the call for radical resection in patients with incidental diagnosis of T2 tumor on cholecystectomy. This study also shows preoperative imaging diagnosis is associated with higher stage and overall worse prognosis.

21

OP-I.04 MAJOR BILE DUCT INJURY AFTER LAPAROSCOPIC CHOLECYSTECTOMY: EXPERIENCE FROM A TERTIARY REFERRAL CENTER IN A RURAL STATE A. Greenbaum, E. Alkhalili, I. Nir University Of New Mexico, Albuquerque, NEW MEXICO

OP-I.03 SUBTOTAL CHOLECYSTECTOMY FOR THE HOSTILE GALLBLADDER M. E. Lidsky, A. W. Castleberry, A. Perez, T. N. Pappas Department Of Surgery, Duke University Medical Center, Durham, NC Background: Outcomes following the inability to safely control the cystic duct in the setting of a hostile triangle of Calot during cholecystectomy remain unknown. The purpose of this study was to analyze the safety and efficacy of subtotal cholecystectomy, with specific attention to the necessity and timing of secondary procedures. Methods: Medical records of 16,585 cholecystectomies from January 2002-August 2014 were reviewed, with identification of patients managed with subtotal cholecystectomy, defined as the inability to isolate and transect the cystic duct. We investigated surgical indications, intraoperative variables, and 30-day postoperative mortality and morbidity. We also analyzed the necessity for ERCP, percutaneous drainage procedures, and completion cholecystectomy. Results: 69 (0.4%) patients underwent subtotal cholecystectomy, of which 57 (82.6%) were laparoscopic; 30 (43.5%) required conversion to laparotomy. 1 (1.4%) patient died postoperatively, and 26 (37.7%) patients suffered 35 complications, most frequently infectious (14 wound/surgical site infections, 4 UTIs). Indication for cholecystectomy included acute cholecystitis (69.6%), 10 (14.5%) of which had a cholecystostomy tube, symptomatic cholelithiasis (23.2%), chronic cholecystitis (13%), and biliary pancreatitis (10.1%). Secondary interventions were required in the form of 49 ERCPs in 20 (29%) patients, percutaneous drainage for biloma or abscess in 6 (8.7%), and completion cholecystectomy on average 13.75 months (5–27 months) after the index operation in 4 (5.8%). Conclusions: The hostile gallbladder represents a complicated disease process for which patient safety is of paramount. Postoperative morbidity after subtotal cholecystectomy is significant. Most do not require completion cholecystectomy; however, these patients demand close observation and, frequently, secondary interventions.

HPB 2015, 17 (Suppl. 1), 1–81

Background: Risk factors for iatrogenic bile duct injuries (BDI) after laparoscopic cholecystectomy (LC) remain a topic of controversy. Few studies have examined ethnicity or body mass index as potential risk factors resulting in the need for complex biliary reconstruction. We hypothesize that LC performed in a rural setting may present a higher risk of BDI in certain patient populations. Methods: This retrospective cohort study includes all patients referred to our tertiary center from 2010–2014 for biliary reconstruction secondary to major BDI during LC. Results: A total of 21 patients were analyzed. These patients were predominately female (76% vs. 24% male). A disproportionate number of patients were Native American (47.6% vs. 10.4% comprising the New Mexico state population). The mean BMI of all patients was 33.5, though the Native American population average BMI was 39.2. The majority of referrals came from rural centers (62%). The mean time to BDI recognition after LC was highest at rural facilities (16.4 days vs. 3.3 days at metropolitan centers) as well as the timing of referral (4.7 days vs. 0.8 days). The overall rate of morbidity was 42.8% and one patient died of sepsis prior to reconstruction. Conclusion: Major BDI after LC is a highly morbid event. This retrospective cohort study suggests Native American patients, specifically those who are morbidly obese, may be at higher risk of BDI during laparoscopic cholecystectomy in a rural setting. We prompt rural surgeons to consider early referral of high risk patients to a tertiary setting for gallbladder surgery.

OP-I.05 RACIAL DISPARITIES IN PATIENTS WITH GALLBLADDER CANCER S. Zenoni1, X. Zhu1, P. Velduis1, S. Eubanks1,2, P. Arnoletti1,2, S. De La Fuente1,2 1 Florida Hospital Orlando, Orlando, FL; 2University Of Central Florida, Orlando, FL Background: Epidemiological studies have shown widely variable geographic patterns in patients with gallbladder cancer. The incidence rates are extraordinarily high in Latin America, Asia and some countries in eastern and central Europe but relatively low in the United States. In contrast to other biliary malignancies, there is limited data regarding racial disparities and oncologic outcomes in patients with gallbladder cancer. In this study, a prospectively maintained tumor registry was used to determine overall survival, timeto-initial treatment, and time-to-surgery according to race in patients with gallbladder cancer. Methods: Patients with gallbladder cancer registered at the Florida Hospital tumor database from the year 2001–2013 were included in the analysis. Analyzed variables included © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

22

Abstracts

basic demographics as well as grade and pathological ACCJ stage at diagnosis, date at diagnosis, initial treatment date, surgery date, last date of contact, patient’s status (alive/dead) at the last contact, and overall survival according to race. Results: A total 133 patients with gallbladder cancer were identified from the database; 91 of which were identified as non-Hispanic, 22 as Hispanics, and 20 as African Americans. There were no differences in age, gender distribution, grade and pathological ACCJ stage at presentations between the different races. Five-years overall survival (figure), time-toinitial treatment, and time-to-surgery were not significantly different between groups. Conclusions: This study shows similar survival rates and time to therapy between Caucasian and minorities. Furthermore, as opposed to what it has been observed with other cancers, no differences were noted in time-to-treatment in gallbladder cancer according to race.

had shorter DSS compared with patients not requiring bile duct resection (9.3 vs 39.9 mo; p = 0.002; Figure). When accounting for differences between the two groups, the need for bile duct resection was independently associated with reduced DSS (HR: 3.06;95%CI: 1.12–8.34;p = 0.029). Conclusion: Major hepatectomy with concomitant bile duct resection is seldom performed in patients undergoing resection of colorectal cancer liver metastases. Bile duct resection is associated with higher major morbidity and reduced disease-specific survival. Stringent selection criteria should be applied when patients may need bile duct resection during hepatectomy for colorectal cancer liver metastases.

OP-I.06 THE IMPACT OF CONCOMITANT BILE DUCT RESECTION WITH MAJOR HEPATECTOMY ON SURVIVAL OUTCOMES OF PATIENTS UNDERGOING TREATMENT OF METASTATIC COLORECTAL CANCER: A MULTI-INSTITUTIONAL ANALYSIS OF 429 PATIENTS L. M. Postlewait1, M. H. Squires1, D. A. Kooby1, S. M. Weber2, C. R. Scoggins3, K. Cardona1, C. S. Cho2, R. C. Martin3, E. Winslow2, S. K. Maithel1 1 Division Of Surgical Oncology, Emory University, Atlanta, GA; 2Division Of Surgical Oncology, University Of Wisconsin, Madison, WI; 3Division Of Surgical Oncology, University Of Louisville, Louisville, KY Background: Data are lacking on the long-term outcomes of patients undergoing major hepatectomy requiring a bile duct resection for the treatment of colorectal cancer metastases. Methods: All patients who underwent major hepatectomy (≥3 segments) for metastatic colorectal cancer from 2000 to 2010 at three US academic institutions were included. Patients who died from unknown cause were excluded. Primary outcome was disease-specific survival (DSS). Results: Of 456 patients, 429 met inclusion criteria. Median follow-up was 38.7 mos. Bile duct resection was performed in 9 patients (2.1%) and was associated with preoperative portal vein embolization (25.0%vs4.3%;p = 0.049). There were no significant differences in other clinicopathologic factors between the two groups (age, ASA class, margin status, number of lesions, tumor size, cirrhosis, perineural invasion, and lymphovascular invasion). Bile duct resection was associated with increased postoperative major complications (Clavien III-V) in univariate (66.7%vs20.7%;p = 0.004) and multivariate analyses (HR: 6.22;95%CI: 1.44–26.97;p = 0.015). There were no differences in length of stay, reoperation, readmission, or 30-day mortality rates. Patients who underwent bile duct resection © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

OP-I.08 INCORPORATING TRANSITION TO PRACTICE INTO A HEPATOPANCREATICOBILIARY FELLOWSHIP: A MODEL FOR GRADUATED AUTONOMY C. Scally, S. G. Warner, R. M. Minter University Of Michigan, Ann Arbor, MICHIGAN Introduction: There is significant concern for graduating trainees’ readiness for independent practice. The American College of Surgeons (ACS) has developed pilot fellowships to facilitate “Transition to Practice.” However, these programs currently serve as an alternative to advanced subspecialty training. We sought to evaluate our institutional experience incorporating a novel transition to practice element into an advanced HPB fellowship. Methods: We obtained complete operative records from Fiscal Years 2012–14 through present for our HPB fellows (2011–13 fellow A, 2013–15 fellow B). We then identified all cases in which the fellows billed as the attending surgeon, as well as all cases in which they participated as a trainee. We also analyzed relative value units (RVUs), and total charges billed by the fellows per financial year. HPB 2015, 17 (Suppl. 1), 1–81

Abstracts Results: The independent practice element expanded in each year of the fellowship (Table). Independent cases represented 36.6% of the fellows’ total operative volume from 2012–14. In the second year of the fellowship, the fellows have increased the complexity of diagnoses seen in their personal clinic with explicit recognition of their ability to progress in this manner. The fellows do not participate in either emergency or service call as an attending. Conclusions: Our HPB fellows’ training included a significant independent practice component, with an expanding scope of practice in each year of training. This method of graduated autonomy in an advanced HPB fellowship may represent a feasible blended model for advanced surgical training, meeting both the need for specialty expertise and preparation for independent practice.

HPB 2015, 17 (Suppl. 1), 1–81

23

OP-I.09 STROKE VOLUME VARIATION (SVV) CONTINUOUS MONITORING FOR INTRAOPERATIVE INTRAVASCULAR FLUID MONITORING IN HPB SURGERY E. H. Baker, J. Drummond, A. Cochran, R. Seshadri, J. Martinie, D. Iannitti, R. Swan Carolinas HealthCare System, Charlotte, NORTH CAROLINA Introduction: Previous studies have demonstrated that stroke volume variation (SVV) represents a more accurate determinant of intravascular volume status and fluid responsiveness. This has been particularly true for patients who are intubated in which previously used measurements of intravascular status such as mean arterial pressure (MAP) or central venous pressure (CVP) may be misrepresentative. Methods: Continuous intraoperative monitoring was performed for a consecutive series of patients who underwent laparoscopic and open liver and pancreas surgeries. Data points collected for analysis included SVV, MAP, CVP, stroke volume (SV), cardiac index (CI) every 5 minutes. Estimated blood loss (EBL), hemoglobin (Hgb), pH and surgeon and anesthesia estimates of patient volume status were determined every hour. Results: 23 patients were enrolled and underwent continuous monitoring during the following procedures: 5 open liver, 6 laparoscopic liver, 6 open pancreas, 6 laparoscopic pancreas. A weak inverse relationship was seen between SVV and total fluid status in 14 out of 23 total cases (60.9%) as indicated by a negative linear regression slope. The correlations were low for all groups, with an average R2 of 0.10 – or 10% of the variability in the model explained by the SVV and total fluid variables. Conclusions: SVV correlates inversely with total fluid status and may be a used as a non-invasive tool in determining fluid status in HPB procedures. While the sample size was small, we plan on using the data to develop a standardized fluid management protocol for HPB surgeries and examine surgical outcomes compared to matched, retrospective cases.

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

24

Abstracts

OP-I.10 RECURRENCE PATTERN AND SURVIVAL IN PATIENTS UNDERGOING SIMULTANEOUS RESECTION FOR SYNCHRONOUS LIVER METASTASES FROM PRIMARY COLORECTAL CANCER : RETROSPECTIVE STUDY OF 286 PATIENTS FROM A SINGLE HIGH VOLUME HPB CENTRE R. Kumar1, S. Park2 Sir H.N. Reliance Foundation Hospital, Mumbai, MAHARASHTRA; 2National Cancer Center, Goyang, GYEONGGI-DO 1

Background: The optimal combination of available therapies for patients with resectable synchronous liver metastases from colorectal cancer (SLMCC) is unknown, and the pattern of recurrence after simultaneous resection has been poorly investigated. In this study, the authors examined recurrence patterns and survival after simultaneous resection for SLMCC. Methods: Consecutive patients with SLMRC who underwent complete simultaneous resection of both the rectal primary and liver metastases with curative intend between May 2001 and December 2010 were identified from a prospective database. Fifteen patients were excluded due to follow-up loss. Clinicopathological factors were retrospectively analyzed to investigate initial recurrence pattern affecting survival. Results: In total, 286 patients underwent simultaneous resection of both primary colorectal cancer and liver metastases with curative intend. The 3, 5 and 10-years recurrencefree survival rate were 35.5%, 31.5% and 27.2%, respectively, for the entire cohort with a median follow-up of 60.8 months for survivors. 196 patients(68.5%) developed a recurrence. Most common site of the initial recurrence was liver(48.5%), followed by the lung (25%). Initial recurrence pattern correlated with survival (P < 0.001). Analysis demonstrated that a loco-regional recurrence was significant risk factor for survival. Conclusions: Of the patients with SLMCC who developed recurrent disease, systemic sites were overwhelmingly more common than pelvic recurrences. The current results indicated that initial recurrence patterns in patients undergoing simultaneous resection with curative intent was important to predict survival, especially in patients with loco-regional recurrence. Also, simultaneous resection seems safe and feasible despite lack of evidence, provided an institution specific protocol is followed.

OP-I.11 IS THERE A ROLE FOR NEAR INFRARED SPECTOMETRY (NIRS) OXYMETRY DURING LIVER SURGERY? Y. Collin1, T. Hu1, R. Allard1, A. Cloutier1, F. Payette1, A. Denault2,3, R. Lapointe1, F. Vandenbroucke-Menu1 1 Centre Hospitalier Universitaire de Montreal, Service de Chirurgie HBP, Montreal, QC; 2Centre Hospitalier Universitaire De Montreal, Soins Intensifs, Montreal, QC; 3 Institut De Cardiologie De Montreal, Soins Intensifs, Montreal, QC Introduction: Peri-operative cerebral and somatic oxymetry is used for monitoring during cardiac surgery, but it © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

has never been used in liver surgery. Our goal was to define the place of NIRS oxymetry in liver resections. Methods: 90 patients undergoing major hepatectomy were included (45 men and 45 women). Oxymetry was obtained by NIRS at 4 sites (cerebral right and left, arm and thigh) before and during surgery. Baseline oxymetry (BaseO) and desaturation (Desat) (Threshold = 80% baseline) values were compared to peri-operative data. Results: Median ICU stay was 2d and median hospital stay (LOS) was 7d. Cerebral BaseO correlated significantly with duration of stay in the intensive care unit (ICU) stay (p = 0.04), in the hospital length of stay (LOS) (p = 0.01) and respiratory insufficiency (p = 0.002). Arm BaseO correlated with blood loss (p = 0.05), blood transfusion (p = 0.03), ICU stay (p = 0.01) and surgical complications (p = 0.049). Also, thigh BaseO correlated with surgical complications (p = 0.0035) and LOS (p = 0.01). Cerebral Desat did not correlate with any complication but thigh Desat did with blood loss (p = 0.03), LOS (p = 0.05) and surgical complications (0.0132). Arm Desat inversely correlated with Pringle duration (p = 0.01). Conclusion: NIRS is a very simple approach for oxymetry evaluation during major liver surgery. Baseline and systemic values are correlated with some operative data and complications. These preliminary results should lead to a more extensive study to determine whether correction of impaired level of oxymetry will improve patients’ outcome.

OP-I.12 USING THE CLINICAL RISK SCORE TO PREDICT WHICH PATIENTS WILL BENEFIT FROM PRE-OPERATIVE CHEMOTHERAPY FOR COLORECTAL LIVER METASTASES N. Sela1,2, K. A. Bertens1,2, S. Welch1,2, J. Chung2, C. S. Yoshy1, R. Hernandez-Alejandro1,2 1 London Health Sciences Centre, London, ONTARIO; 2 Western University, London, ONTARIO Background: Pre-operative chemotherapy has become a mainstay in the treatment of colorectal liver metastases (CRLM) in patients with both resectable and unresectable disease. We aim to evaluate whether the Clinical Risk Score (CRS) can be used to predict patients with resectable disease who will benefit from metastectomy before chemotherapy. Methods: All patients with CRLM who underwent 4 or more cycles of chemotherapy at our institution between 2000 and 2013 were retrospectively analyzed. Patients had to have cross-sectional computed tomography (CT) imaging before and after the treatment to meet inclusion criteria. Tumor response was determined using RECIST 1.1 criteria. The primary outcome was tumor progression, defined as RECIST 1.1 progressive disease. No progression was defined as RECIST 1.1 stable disease, complete response, or partial response. Multivariate regression was used to assess if the CRS, as well as its individual components, were predictive of disease progression. Results: Seventy-seven patients were identified (mean follow-up of 2.19 ± 1.50 years). Twenty-seven patients had progressive disease (35.1%). On multivariate analysis, a lower CRS was predictive of disease progression on chemotherapy (OR = 0.370, p = 0.018). Furthermore, patients with HPB 2015, 17 (Suppl. 1), 1–81

Abstracts 5 or more metastases (OR 0.122, p = 0.023), and increased size of the largest lesion (OR 0.647, p = 0.047) were less likely to progress. Conclusion: Patients with more aggressive tumors, as reflected by a higher CRS, are less likely to have tumor progression on chemotherapy. Conversely, patients with low CRS are more likely to progress on chemotherapy, and therefore those with resectable disease upfront would benefit from metastectomy (without pre-operative chemotherapy).

OP-I.13 INTRAOPERATIVE RADIOFREQUENCY ABLATION VERSUS SURGICAL RESECTION IN SOLITARY SMALL HCC A. M. Elgendi, M. Elshafey, E. Bdeawey Faculty Of Medicine, Alexandria University, ALEXANDRIA, SELECT A STATE/PROVINCE Background: Percutaneous radiofrequency ablation (RFA) is used for treatment of small HCC however surgeons are frequently using intraoperative RFA for tumors at locations difficult for the percutaneous procedure. The aim was to evaluate the results of intraoperative RFA for small HCCs (<2 cm) at locations difficult for percutaneous route. Methods: 420 patients with small solitary HCC (<2 cm) were treated; 328 via percutaneous RFA while 92 patients presented at sites not amenable for percutaneous route. 48 out of 92 patients underwent surgical resection, while 44/92 patients underwent intraoperative RFA. Results: The location and depth of the HCC from the liver capsule was the only significant factors in the choice of the surgeon between resection and RFA. RFA group acheived complete ablation rate of 100% compared to the surgery group, where all patients achieved R0 resection. Complication rate was comparable (p = 1.0). After a median follow-up of 46 months (range, 16–65 months), no tumors showed neither local progression nor local recurrence and no significant difference was observed between two groups as regards early recurrence and number of de novo lesions (p = 0.49). One-year and 3-year survival rates were 92% and 83%, respectively, in the resection group comparable to the corresponding rates of 91% and 76% in the RFA group (p = 0.8). Conclusion: For small HCC in locations difficult for a percutaneous approach, intraoperative RFA can be an alternative option for deep seated tumors necessitating more than one segmentectomy achieving similar tumor control, overall and disease-free survival.

OP-I.14 HEALTH-RELATED QUALITY OF LIFE FOLLOWING TREATMENT OF NEUROENDOCRINE LIVER METASTASIS G. Spolverato, H. Maqsood, Y. Kim, T. Luo, R. Gupta, T. M. Pawlik Johns Hopkins Hospital, Baltimore, MARYLAND Background: A large subset of patients with neuroendocrine liver metastasis(NELM) is symptomatic at the time of presentation. In addition to improving survival, treatment of NELM seeks to provide palliation of symptoms. HPB 2015, 17 (Suppl. 1), 1–81

25

Data on health-related quality of life(QoL) are uncommon. We sought to define patient-reported QoL following treatment of NELM. Methods: Patients who underwent treatment of NELM at Johns Hopkins Hospital between 1998–2013 and who were alive as of March 2014 were identified(n = 276). These patients were invited to complete a QOL survey designed using validated assessment tools. Results: The response rate was 27.5%(n = 76); questionnaires were completed at a median of 49.1 months(range, 23.3–117.8) following initial treatment. Median patient age was 55 years and the majority was male(n = 43, 56.6%). Most patients had a pancreatic(n = 21, 27.6%) or a small bowel(n = 30, 39.5%) primary tumor; the overwhelming majority had multiple NELM(88.2%). Prior to the initiation of any therapy, 84.2% patients reported symptoms, with the most common symptoms being fatigue(78.9%), diarrhea(67.1%), and flushing(44.7%). Initial treatment of NELM consisted of resection ± ablation(64.5%) or intraarterial therapy(IAT)(35.5%). Many patients reported overall improvement in physical health(41.4%) and mental health(34.3%). After treatment, the proportion of patients with severe symptoms decreased from 36.8% to 21.0%(P = 0.03); symptoms such as diarrhea and flushing improved at the end of the treatment course(both p < 0.05), whereas fatigue remained unchanged(p = 0.45). Conclusions: Surgery and IAT management of NELM provides a reasonable improvement in patient-reported symptoms and QoL. Liver-directed therapies should be considered in those patients with symptomatic, high-volume disease even if complete treatment of the NELM is not feasible.

OP-I.15 IMAGING SURVEILLANCE OF HEPATOCELLULAR ADENOMAS Y. Chun, R. Parker, S. Reddy, E. Ehrenwald, M. Hill, S. Inampudi, T. Sielaff Virginia Piper Cancer Institute, Minneapolis, MN Background: A consensus surveillance protocol is lacking for patients with hepatocellular adenomas. Methods: Patients with hypervascular hepatic lesions ≤5 cm that did not meet criteria for focal nodular hyperplasia or hepatocellular carcinoma were entered into a surveillance schedule with contrast-enhanced MRI 6, 12, and 24 months after baseline imaging. Patients with risk factors, including male gender and active oral contraceptive pill (OCP) use, were excluded. If lesions remained stable or decreased in size, then surveillance imaging was discontinued. Results: Between 2011–2014, 109 consecutive patients with benign hypervascular lesions were evaluated at our multidisciplinary liver conference. Median follow-up from date of diagnosis was 24 months (range, 0–168 months). By imaging criteria or biopsy, 37 patients were diagnosed with focal nodular hyperplasia, and 44 patients with hepatocellular adenoma, including 18 with adenomatosis. Twenty-eight patients had indeterminate lesions. Clinically significant hemorrhage +/− rupture occurred in 6 patients without prior imaging and 2 patients with known hypervascular hepatic lesions. All patients who suffered hemorrhage had adenomas >5 cm and other risk factors, including OCP use (n = 7) and anticoagulation (n = 1). In 43 patients eligible for our © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

26

Abstracts

surveillance schedule, all lesions remained stable or decreased in size, and no patient developed complications. Conclusions: Patients with hepatocellular adenomas without risk factors, such as size >5 cm or OCP use, can safely be observed with serial imaging 6, 12, and 24 months after initial diagnosis. If lesions remain stable or decrease in size, then longer-term surveillance is unlikely to identify patients at risk for complications.

OP-I.17 NATIONAL TRENDS WITH LAPAROSCOPIC LIVER RESECTION: RESULTS FROM A POPULATION-BASED ANALYSIS J. He, N. Amini, G. Spolverato, K. Hirose, M. A. Makary, C. L. Wolfgang, M. J. Weiss, T. M. Pawlik The Johns Hopkins Hospital, Baltimore, MD Background: Interest in laparoscopic liver resection (LLR) has grown since the International “Louisville Statement” was published in 2009. However, limited population-based data on LLR utilization patterns and outcomes are available. Methods: LLR data from the Nationwide Inpatient Sample (NIS, 2000–2012) and the National Surgical Quality Improvement Project (NSQIP, 2005–2012) were divided into 2 cohorts, before and after the Louisville Statement. Patient demographics, indications, trends in LLR utilization, and perioperative outcomes were compared before and after the 2009. Results: Patients undergoing open versus LLR were comparable with regard to age, sex, and comorbidity status (Table). 1,131 and 642 LLR were identified from NIS and NSQIP, respectively. The majority of patients underwent LLR for a malignant indication (NIS: primary malignancy, 29.9% vs. metastasis, 43.2%; NSQIP: primary malignancy, 25.7% vs. metastasis, 42.2%). The mean annual volume of LLR increased from 2000–2008 vs. 2009–2012 (NIS: 63 vs. 168; NSQIP: 52 vs. 127; both P < 0.01). The perioperative mortality associated with LLR was low (NIS: 2.8% vs. NSQIP: 0.9%), while the morbidity was higher (NIS: 38.1% vs. NSQIP: 30.7%); mortality and morbidity did not change over time (both P > 0.05). Since 2009, LLR was associated with a shorter length of stay (NIS: 5 vs. 6 days, P < 0.01) and more likely to be performed in teaching hospitals (NIS: 93% vs. 87%, P = 0.02). Conclusions: Since the Louisville Statement in 2009, utilization of LLR has increased. LLR appears to be safe with low mortality and reasonable morbidity, as well as be associated with a modest decrease in LOS.

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

OP-I.18 POST-HEPATECTOMY HYPERBILIRUBINEMIA: THE POINT OF NO RETURN J. Y. Liu1, L. M. Postlewait1, J. W. Etra1, M. H. Squires1, K. Cardona1, J. H. Winer1, J. M. Sarmiento2, C. A. Staley 1, S. K. Maithel1, D. A. Kooby1, M. C. Russell1 1 Emory Winship Cancer Institute, Atlanta, GA; 2Emory Department Of Surgery, Atlanta, GA Background: Post-hepatectomy hyperbilirubinemia is associated with liver insufficiency and failure. The threshold of the highest survivable total bilirubin (tbili) is not defined. Our aim was to identify the peak postoperative tbili beyond which is survival is improbable. Methods: An institutional database of patients undergoing major hepatectomy (≥3 segments), excluding biliary resections, from 2000–2012 was reviewed. A peak bilirubin of ≥18 mg/dL in the first 45 days post op was associated with increasing 90-day mortality (90DM). Clinicopathologic factors were assessed for association with 90DM. We also examined predictors of elevated postoperative tbili. Results: 607 patients were identified with a 90DM of 4.4%. 90DM for a peak tbili ≥18 (n = 16) was 81%, compared to 2.4% for a bilirubin <18 mg/dL (graph). All patients with a tbili ≥30 died (n = 7). On multivariate analysis (MVA) for 90DM, post-operative tbili ≥18 (HR 24, CI 3.3–174;p = 0.002), post-operative FFP (HR 4.8, CI 1.1–20.2;p = 0.034), and cirrhosis (HR 5.9, CI 1.1–31.3;p = 0.038) were significant predictors. Furthermore, predictors of tbili ≥18 identified on MVA included: older age (HR 1.1, CI 1.0–1.2; p = 0.001) and postoperative FFP (HR 10.1, CI 2.5–40.8; p = 0.001). Conclusion: Total bilirubin ≥18 is significantly associated with an increase in 90-day mortality after major hepatectomy; there are no survivors for patients whose tbili rises ≥30.This information can help clinicians advise patients and families who experience posthepatectomy hyperbilirubinemia; as well, it may be an important marker for intervention as supportive therapies improve.

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

27

OP-I.19 EARLY AND LONG-TERM SURVIVAL OUTCOMES OF PATIENTS WITH COLORECTAL LIVER METASTASES RESECTED AFTER CONVERSION CHEMOTHERAPY M. C. Marques, H. S. Ribeiro, R. S. De Souza, W. L. Da Costa, A. L. Diniz, A. L. De Godoy, I. C. De Farias, C. A. De Mello, F. A. Soares, M. A. Almeida, F. J. Coimbra A.C. Camargo Cancer Center, Sao Paulo, Brasil Liver resection is the standard of care for patients with colorectal liver metastases, but only 15% to 25% are resectable at the initial diagnosis. Improvements in chemotherapy response rates and surgical technique expanded the resectability criteria for patients initially considered as having unresectable disease. These patients have an intermediate prognosis between those who were upfront resectable and those who did not achieve resectability. The aim of this study was to analyze survival outcomes and determine clinicopathological prognostic factors in this scenario. Patients who underwent liver resection for colorectal liver metastases after conversion therapy between 1998 and 2013 were retrospectively analyzed. Unresectability was defined according to the Consensus Guidelines recently published. In the study period, 352 liver resections for colorectal metastases were performed in 268 patients. Fifty-one patients met the inclusion criteria and were analyzed. Thirty and 90-day mortality rate was 5.8% and 9.8%. Major morbidity rate was 29.4% according to Clavien-Dindo Classification. Median number of chemotherapy cycles before hepatectomy was 12. With a median follow-up of 36 months, the 3-year overall survival was 66.1% versus 76.4% in the resectable group (p = 0.060) and 3-year disease-free survival was 10.4% versus 35.8% (p < 0.001). In multivariate analyses, factors that influenced overall survival were the occurrence of Class III and IV postoperative complications and more than 4 liver nodules. There was no independent predictor of disease free survival on multivariate analyses. Liver resection after conversion therapy for colorectal liver metastases is a procedure with high morbimortality but also associated with long-term survival in selected patients.

OP-I.20 RESECTED INTRAHEPATIC CHOLANGIOCARCINOMA: PATTERNS OF ADJUVANT THERAPY AND RECURRENCE I. T. Konstantinidis 1,2, A. X. Zhu1, L. Goyal 1, D. T. Ting 1, V. Deshpande 1, K. K. Tanabe 1, K. D. Lillemoe 1, C. R. Ferrone 1 1 Massachusetts General Hospital Department Of Surgery, Boston, MA; 2University Of Arizona Department Of Surgery, Tucson, AZ Background: The majority of patients who undergo liver resection for intrahepatic cholangiocarcinoma (ICC) suffer from recurrence and succumb to their disease. The role of adjuvant treatment remains unknown. Methods: Clinicopathologic data of resected ICC between 1/2000–12/2013 were evaluated. Patterns of adjuvant therapy and recurrence were analyzed. Results: Of the 76 patients who underwent resection of a ICC the median age was 65 years and 54% were female. The majority of patients (72%) underwent a major hepatectomy HPB 2015, 17 (Suppl. 1), 1–81

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

28

Abstracts

with 87% being an R0 resection. Median ICC size was 5.5 cm. Excluding patients who died within 60 days or were lost to follow-up, 44%(30/68) did not receive adjuvant treatment (median overall survival (OS)not reached), 34%(23/68) received adjuvant therapy within 3 months (OS: 48 months), and an additional 22%(15/68) received therapy after the first recurrence (OS: 44 months). Adjuvant chemotherapy was more likely to be given to patients with an R1 resection (26%vs7%;p = 0.05) and nodal disease (35%vs11%;p = 0.02). Gemcitabine based regimens were the most common first line (85%) (gemcitabine: 32%, gemcitabine/cisplatin: 36%, gemcitabine/oxaliplatin: 32%); 5 FU/chemoradiation was first line in 15%. After median follow-up of 25 months 41% remained free of disease(NED), 12% had a single liver recurrence, 13% multinodular liver recurrence and 34% extrahepatic recurrence. Of the patients who recurred single liver recurrences had the best median survival compared to multinodular recurrences and distant disease (73 vs 23 vs 37 months respectively;p = 0.09). Conclusions: The majority of patients undergoing resection for an ICC either do not receive adjuvant treatment or they receive it after recurrence of their tumor. Single nodule hepatic recurrences have the best outcome.

Conclusion: Liver resection should be considered only in selected patients with melanoma liver metastases without EHD which is associated with poor outcomes. When EHD is controlled with systemic therapy, indication for combined liver ablation needs further investigations.

OP-I.21 ROLE OF LIVER RESECTION OR ABLATION IN METASTATIC MELANOMA MANAGEMENT A. Doussot1, C. Nardin1, H. Takaki2, T. Litchman2, M. I. D’Angelica1, W. R. Jarnagin1, M. A. Postow3, J. P. Erinjeri2, P. Kingham1 1 Department Of Surgery Memorial Sloan Kettering Cancer Center, New York, NY; 2Department Of Interventional Radiology Memorial Sloan Kettering Cancer Center, New York, NY; 3Department Of Medical Oncology Memorial Sloan Kettering Cancer Center, New York, NY Introduction: The median survival for patients with metastatic melanoma is usually limited to approximately one year. Liver ablation and metastasectomy are associated with improved survival in well-selected patients, but their role in era of more effective systemic therapies is uncertain. Methods: Patients undergoing liver ablation or resection for melanoma liver metastases between 1993 and 2013 were included. Outcomes and prognostic factors such as medical therapy impact were evaluated. Results: Forty eight patients underwent ablation (n = 16) or resection (n = 32) for metastases from cutaneous (n = 26) or ocular (n = 22) origin. Median overall survival (OS) was 26 months, with 12 patients (27.3%) alive at 3 years, after resection (n = 9) and after ablation (n = 3) respectively. Patients in the ablation group harbored more aggressive disease at time of presentation, with more extrahepatic disease (EHD) (p = 0.008) and a shorter disease-free interval between primary tumor and liver metastasis diagnosis (p = 0.01). In addition, 68.8% of patients received preoperative systemic therapy before ablation. Median OS in ablation (18 months) and resection (27 months) groups was not different (p = 0.9). EHD was a poor prognostic factor for OS in the resection group (p < 0.05; see Figure). In patients with EHD, 78% of patients receiving preoperative therapy before ablation experienced partial response and tended to have longer OS (12 months) compared to resected patients (6 months, p = 0.2). © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

OP-I.22 RADIOFREQUENCY-ASSISTED LIVER PARTITION AND PORTAL VEIN LIGATION (RALPP): COMPARATIVE SERIES OF A MODIFIED ALPPS TECHNIQUE FOR TWO-STAGE LIVER RESECTION M. H. Sodergren, T. M. Gall, M. Nagendran, L. R. Jiao Imperial College, London, GREATER LONDON Background: The introduction of portal vein embolization and recently the ALPPS technique has rendered a greater proportion of liver tumours surgically resectable by increasing the volume of future liver remnant (FLR) in selected patients. The RALPP technique involves a laparoscopic first stage portal vein ligation and in situ liver splitting using ablation only without complete transection. We hypothesise that this will rapidly increase the size of the FLR limiting any associated morbidity from liver transection. HPB 2015, 17 (Suppl. 1), 1–81

Abstracts Methods: Consecutive patients who underwent RALPP were compared to an age-sex- and liver function-matched cohort of patients undergoing PVE prior to right hepatectomy. The primary endpoint was the percentage increase in FLR volume. Secondary endpoints were morbidity, mortality, and postoperative liver function. Results: There were 12 patients (6M : 6F) in the RALPP group and 8 (4M : 4F) in the PVE group with a median age of 62.5 and 65 yrs respectively. The mean % increase in the FLR volume was 61.5 +/− 16.3 measured after a mean of 20.8 +/− 7.3 days following the first stage for RALPP compared to a % increase of 16.46 +/− 11.7 (p = 0.001) after 52.3 +/− 14.8 days (p < 0.001) following PVE. There was one mortality in the RALPP group at day 19 following right hepatectomy from bowel ischaemia and liver failure. There was no difference in morbidity or post-operative liver function. Conclusion: The RALPP technique is feasible and safe in this limited series, with a greater increase in FLR volume in a shorter time period compared to PVE.

OP-I.23 TEMPORAL TRENDS IN SURGICAL RESECTION AND PERI-OPERATIVE CHEMOTHERAPY FOR COLORECTAL CANCER LIVER METASTASES (CRCLM) IN ROUTINE CLINICAL PRACTICE S. Nanji1,2, W. J. Mackillop2,3,4, X. Wei3, C. M. Booth2,3,4 Department Of Surgery, Queen’s University, Kingston, ON; 2Department Of Oncology, Queen’s University, Kingston, ON; 3Department Of Public Health Sciences, Queen’s University, Kingston, ON; 4Division Of Cancer Care And Epidemiology, Queens University Cancer Research Institute, Kingston, ON

1

Background: The treatment of CRCLM continues to change over time. We report trends in management and outcome of all patients with resected CRCLM in Ontario, Canada. Methods: All cases of CRC in Ontario who underwent surgical resection of liver metastases in 2002–2009 were identified using the population-based Ontario Cancer Registry. Electronic records of treatment and pathology reports were linked to the registry to identify utilization of neoadjuvant (NACT) and adjuvant chemotherapy (ACT) and describe surgical management and pathologic findings. We describe differences over 2 study periods: 2002–2005 and 2006–2009. Results: During 2002–2009, 1711 patients underwent resection of CRCLM. Mean age was 63 years. During the study period there was a 60% increase in patients undergoing resection of CRCLM. For the 2 study periods, mean number of liver lesions resected was 2.0 and 2.2 (p = 0.051), mean size of largest lesion was 4.5 cm and 4.0 cm (p = 0.003), major hepatic resection (≥3 Couinaud segments) rate was 66% and 63% (p = 0.264) and R1 resection margin rate was 6% and 9% (p = 0.021), respectively. 90-day mortality rates for the study periods were 4% and 3% (p = 0.499). Use of NACT and ACT increased from 19% to 41% (p < 0.001) and 42% to 50% (p < 0.001) between study periods, respectively. Five year overall survival during the 2 study periods was 43% (95%CI 40–47) and 45% (95%CI 42–48) (p = 0.402). HPB 2015, 17 (Suppl. 1), 1–81

29

Conclusions: Resection of CRCLM and the use of perioperative chemotherapy increased during the study period. Survival outcomes among patients treated in routine clinical practice are comparable to institution-based studies.

OP-I.24 PATIENTS WITH HEPATITIS B PRESENT WITH MORE ADVANCED LIVER CANCER THAN PATIENTS WITH HEPATITIS C J. L. Pasko1, A. C. Anderton1, C. Costantino1, S. L. Orloff1, L. L. Wong2, W. E. Naugler3 1 Oregon Health And Sciences University, Department Of Surgery, Portland, OREGON; 2University Of Hawaii Cancer Center And School Of Medicine, Department Of Surgery, Honolulu, HI; 3Oregon Health And Sciences University, Department Of Hepatology, Portland, OREGON Background: Well-described factors affecting the incidence of Hepatocellular Carcinoma (HCC) include gender, age, and etiology of liver disease. Factors that may affect the stage of HCC at presentation are poorly understood. Stage at presentation, however, largely dictates available treatments and ultimately prognosis for patients. Aims: Using a large cohort of patients with HCC, identify elements that correlate with HCC stage at initial BCLC stage. Methods: This is a two-institution retrospective review of patients with pathologically or radiographically confirmed HCC from January 1991–August 2014. Results: There were 1134 patients (835 men and 299 women) with HCC in this study. MELD scores ranged from 6–38. Higher MELD scores correlated with higher stages of HCC (p = .01). Screening was found to correlate with a lower BCLC stage (p = .001). Additionally, patients with HBV presented at higher stages (p = 0.03) compared to patients with HCV who presented at lower stages (p = .001). Thirty-three percent of patients with HCV were screened, and 19% of HBV patients were screened. There was no difference in mean MELD score in HCV vs HBV patients. (10.8 vs 10.3, p = 0.10) Mean tumor size for HCV was 4.4 cm compared to 6.7 cm in HBV (p = 0.001). Conclusions: Despite similar MELD scores between patients with HCV and HBV, patients with HBV infection presented with higher stages of HCC and larger tumors. It is unclear if this is biologically driven or is related to a screening disparity between the two groups.

OP-I.25 IMPACT OF POSTOPERATIVE MORBIDITY AND LIVER FAILURE ON SURVIVAL OF PATIENTS WITH RESECTED COLORECTAL LIVER METASTASES H. S. Ribeiro, M. C. Marques, R. S. De Souza, W. L. Da Costa, A. L. Diniz, A. L. De Godoy, I. C. De Farias, H. C. Freitas, F. J. Coimbra A.C. Camargo Cancer Center, Sao Paulo, Brasil Liver resection has become the best chance of cure for patients with colorectal liver metastases. However, studies have shown that postoperative complications could compromise long-term survival results. The aim of this study was to determine the incidence and prognostic factors for postop© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

30

Abstracts

erative morbidity, including liver failure, and their impact on long-term survival. Patients who underwent liver resection for colorectal liver metastases between 1998 and 2012 were analyzed. Any deviation from usual postoperative early outcome was recorded as complications and stratified according to the Clavien-Dindo classification. Three hundred forty-six liver resections were performed in 285 patients. Postoperative morbidity occurred in 57.8% of operations and 90-day mortality was 3.4%. Major complications (grade 3 and 4) occurred in 21.1% of hepatectomies. Twenty-three (6.6%) patients had postoperative liver failure. After a median follow-up period of 37.5 months, patients who had major postoperative complications and/or postoperative liver failure had statistically significant worse 3-year overall survival. In multivariate analyses, both factors were independent predictors of survival (Major postoperative complications HR = 2.1, p = 0.004, 95% CI 1.2–3.7; postoperative liver failure HR = 1.9, p = 0.046, 95% IC 1.0–3.8). However, these factors had no impact on disease free survival. These findings may be explained by a decrease in life expectancy observed in patients who experience life-threatening events. Detailed clinical preoperative evaluation, careful surgical techniques and meticulous postoperative care may improve long-term survival results.

OP-I.26 GADOXETIC ACID (GA) RELATIVE LIVER ENHANCEMENT (RLE) ON PREOPERATIVE MAGNETIC RESONANCE IMAGING (MRI) AS A MARKER OF LIVER FUNCTION TO PREDICT THE RISK OF POST-HEPATECTOMY LIVER FAILURE (PHLF) A. Tremblay St-Germain1, A. Costa2, R. Smoot3, K. Jhaveri2, S. Cleary1 1 Department Of Surgery, Toronto General Hospital, University Of Toronto, Toronto, ON; 2Joint Department Of Medical Imaging, Toronto General Hospital, University Of Toronto, Toronto, ON; 3Department Of Surgery, Mayo Clinic, Rochester, MN Preoperative liver function and future liver remnant (FLR) are typically used to predict the risk of PHLF. The hepatic physiologic reserve and ability to hypertrophy are difficult to predict. We investigated the use of contrast-enhanced MRI with GA as a marker for liver function and potential predictor of PHLF. We retrospectively analyzed all consecutive patients who underwent major liver resection (≥3 segments) and preoperative MRI-GA at our institution between October 2010 and December 2013. Mean RLE was calculated based on regions of interest drawn of the liver on the unenhanced and hepatobiliary phases. The associations between mean RLE and PHLF according to the 50-50 and ISGLS criterias were tested with univariate and multivariate logistic regression analysis. 68 patients (44 men; median age 60.5 years) fulfilled the inclusion criterias: 47 CRLM, 14 HCC, 2 intrahepatic cholangiocarcinoma, 2 NETLM and 3 benign diseases. 1 patient had PHLF according to the 50-50 criteria and 13 patients had PHLF according to the ISGLS criterias (4 grade A, 8 grade B, 1 grade C and death). Mean RLE correlated with the presence and stage of fibrosis on histology (p = 0,032 and p = 0,045). In logistic regression analysis, mean RLE with a cut-off of 100%, in combination with FLR, increased the ability to predict the presence of PHLF according to the ISGLS criteria. However, it did not reach statistical significance as an independent predictor. MRI-GA and mean RLE, in addition to FLR, can improve risk assessment for PHLF after major liver resection.

OP-I.27 THE LEARNING CURVE EFFECT IN LAPAROSCOPIC LIVER RESECTION V. Villani, F. Sabbatino, R. Torabi, D. L. Berger, K. K. Tanabe, K. D. Lillemoe, C. R. Ferrone Massachusetts General Hospital – Department Of Surgery, Boston, MA Background: The expansion of laparoscopic liver resection (LLR) has been considerably slower compared to that of other laparoscopic procedures. In this study we analyzed the learning curve associated with LLR. Method: Retrospective database analysis of consecutive LLR performed between 3/07–6/14. Procedures were divided in three chronological groups: A (03/07–05/11, 43 pts), B (05/11–05/13, 43 pts), C (05/13–06/14, 42 pts). Results: The three groups were comparable for patients’ median age (57 yo; 59 yo; 59 yo), gender (25.6, 39.5, 45.2,% © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts male), percentage of patients who had a resection for a malignancy (86.0%, 83.7%, 65.6%), and median operative time (182 min, 190 min, 197 min). Complex procedures (defined as: trisegmentectomies, left hepatectomies, right hepatectomies or central hepatectomies) increased significantly over time (P = 0.007). One 90-day mortality occurred in group A, while mortality in groups B and C was zero. There was an improving trend in estimated blood loss (568.5, 563.9, 342.0 ml), in number of patients transfused (4, 3, 2 patients), conversion to open procedures (4, 2, 1 conversions), number of complications (8, 3, 4 complications, Clavien grade III or higher). Median length of stay was significantly reduced in groups B and C (6.4 days, 3.8 days, 3.8 days; P = 0.006). Conclusions: Similar to other surgical procedures, LLR is subject to a learning curve. Despite an increase in surgical complexity, after the first 43 procedures there was an improvement in blood loss, rate of conversion, morbidity, and length of stay.

OP-I.28 SAFETY AND EFFICACY OF A NEW ARTICULATING BIPOLAR ENERGY DEVICE FOR PARENCHYMAL TRANSECTION IN LAPAROSCOPIC LIVER RESECTION C. Dural, M. Akyuz, E. Aksoy, P. Yazici, F. Aucejo, C. Quintini, C. Miller, J. Fung, E. Berber Cleveland Clinic, Cleveland, OH Background: The aim of this study is to assess the safety and efficacy of a new articulating vessel sealer (VS) for laparoscopic liver resection (LLR). Methods: A new 5 cm, bipolar VS was used in 28 LLRs (group 1). A comparison was made to 28 patients who underwent LLR (group 2) using other energy devices. T-test and Chi square were used for statistics. Results: Tumor type was malignant in 71% of patients in group 1 and 89% of the patients in group 2 (p = 0.360). Number and size of tumors, and resection type were similar in both groups. In group 1, less number of adjunctive devices (i.e. energy, clip appliers, staplers) were used (median 2) versus group 2 (median 3, p = 0.032). Staplers were used in 28% (n = 8) of cases in group 1 and 54% (n = 15) in group 2, with fewer number of cartridges fired in group 1, when used (median 1.5 vs 4, respectively, p = 0.005). Parenchymal transection time (28.2 ± 3.5 minutes vs 55.2 ± 4.1, respectively, p < 0.001) and total operative time (200.1 ± 13.7 vs 242.7 ± 14.4, respectively, p = 0.036) were shorter for group 1. Morbidity was 11% (n = 3) in group 1 and 18% (n = 5) in group 2 (p = NS). Intraoperative costs were an average of $ 3000 less in group 1 versus group 2 (p = 0.0029). Conclusion: This study demonstrates the safety and efficacy of a new energy device for LLR. The data suggests a potential benefit of this device to reduce operative time and decrease costs by facilitating parenchymal transection.

31

OP-I.30 TREATMENT OF RECURRENCE AFTER RESECTION OF HEPATOCELLULAR CARCINOMA IN CIRRHOTIC LIVER Z. Rong, Y. Collin, S. Turcotte, M. Dagenais, R. Letourneau, M. Plasse, A. Roy, R. Lapointe, F. Vandenbroucke-Menu Centre Hospitalier De L’Universitaire Montreal (CHUM), Montreal, QUEBEC Background: High recurrence rates after liver resection for hepatocellular carcinoma (HCC) remain problematic and no consensus exists on the optimal management of recurrence. Methods: A retrospective analysis was performed on cirrhotic patients with resected HCC in our center between 1992 and 2013. Survival and recurrence outcomes were analyzed. Results: Ninety-four patients with Child A (90 patients; 95.7%) or B (4 patients; 4.3%) cirrhosis were included. There were 79 males with a mean age of 60.8 ± 11.2 years. Major hepatectomy was performed in 43 patients (45.7%). Median HCC size was 3.5 cm (0.8–16.0 cm) and 79 patients (84.0%) had a single tumor. Forty-three patients (45.7%) experienced recurrent disease, mostly intrahepatic (36 patients; 83.7%). Overall and disease-free survivals were 70.0% and 48.4% at 3 years, and 61.6% and 41.6% at 5 years. Inferior overall and disease-free survivals were significantly associated with positive margins (p = 0.005 and p < 0.001), multiple tumors (p = 0.018 and p = 0.03), and multinodularity (p = 0.007 and p < 0.001). Age (<65 years; p = 0.03), tumor size (<3 cm; p = 0.02) and lower T stage (p = 0.02) were linked with longer overall survival. Invasion of adjacent organs (p < 0.001) and major vascular structures (p = 0.01) were associated with recurrence. Thirty-three patients with recurrent HCC received treatment, which included chemotherapy (13 patients), chemoembolization (12), radiofrequency (8), alcoholization (3), repeat resection (4), or transplantation (3). Overall survival was significantly superior in patients with treated recurrence (p = 0.005). Conclusions: Recurrence is frequent after HCC resection, particularly in patients with invasion of adjacent organs or vascular structures. However, satisfactory 5-year survival rates are achievable in patients with treated recurrent disease.

OP-I.31 MULTIMODAL TREATMENT OF UNRESECTABLE HEPATOCELLULAR CARCINOMA TO ACHIEVE COMPLETE RESPONSE RESULTS IN IMPROVED SURVIVAL P. Newell1,2,5, R. Uppal3, Y. Wu1, H. Hoen1,5, J. T. Thiesing4, K. Sasadeusz4, M. Cassera1, R. Wolf1,2, P. Hansen1,2, C. Hammill1,2 1 Providence Cancer Center, Portland, OR; 2The Oregon Clinic, Portland, OR; 3The Portland Clinic, Portland, OR; 4 The Radiology Group, Portland, OR; 5Earle A. Chiles Research Institute, Portland, OR Introduction: With technological advances, questions arise regarding how to best fit newer treatment modalities, such as transarterial therapies, into the treatment algorithm for patients with hepatocellular carcinoma (HCC).

HPB 2015, 17 (Suppl. 1), 1–81

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

32

Abstracts

Methods: Between 2005–2011, 128 consecutive patients initially treated with transarterial radioembolization or chemoembolization using drug-eluting beads were identified. Response was graded retrospectively. Toxicity was measured at 1,3, and 6 months after first and last treatments by recording biochemical adverse events in bilirubin, albumin, and INR. Results: 53% of the patients were considered to have advanced stage disease (BCLC stage C). 16% of patients had an initial complete response, but with additional treatments, this was increased to 36%. Patients with a complete response as their best response to treatment had a median survival (95% confidence interval) of 5.77 (2.58, the upper limit has not yet been reached) years, significantly longer than those whose best response was a partial response, 1.22 (0.84, 2.06) years and those with stable disease as their best response, 0.34 (0.29, 0.67) years. Repeated treatments did not increase the risk of toxicity. Discussion: This retrospective review of patients treated for intermediate and advanced stage HCC revealed a significant survival advantage in patients who achieved a complete response. We did not demonstrate superiority of one modality over the other, but did show that the two could be used sequentially without accumulating significant toxicity. These data support use of a multi-modality approach to intermediate and advanced stage HCC, combining liver-directed treatments as necessary to achieve a complete response.

OP-I.32 HEPATIC RESECTION FOR DISAPPEARING LIVER METASTASIS: A COST-UTILITY ANALYSIS G. Spolverato1, A. Vitale2, A. Ejaz1, D. Cosgrove1, D. Cowzer1, U. Cillo2, T. M. Pawlik1 1 Johns Hopkins Hospital, Baltimore, MARYLAND; 2 Universita Di Padova, Chirurgia Epatobiliare E Trapianto Epatico, Padova, PADOVA Introduction: Data on cost-effectiveness and efficacy of hepatic resection(HR) for colorectal liver metastasis that © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

disappear after systemic chemotherapy(sCT) are lacking. We estimated the cost-effectiveness of HR plus 6 months of sCT in case of recurrence(strategy A) relative to surveillance and 6 months of sCT in case of recurrence(strategy B) for patients with colorectal disappearing liver metastasis(DLM). Methods: Through a Markov model three base cases were evaluated involving a 65 year-old patient with three lesions in the right hemi-liver who underwent 6 months of sCT and 1)had DLM based on MRI; 2)had DLM based on CT scan; 3)had also hepatic artery infusion(HAI) with subsequent DLM based on CT scan. Results: The NHB of strategy A(HR) versus strategy B(surveillance) was negative(-1.7 QALMs) for base case 1. In contrast, the NHB of HR was positive in base case 2(5.4 QALMs); the NHB of HR was positive for base case 3, but the effect was much more modest(0.15 QALMs). The ICER of strategy A versus B was highest for base case 1($105,216/ QALY) and lowest for base case 2($-18,768/QALY); the ICER for HR versus surveillance was intermediate for base case 3($48,924/QALY). Sensitivity analyses demonstrated that HR was cost effective when compared with surveillance when the rate of complete pathological response after 6 months of sCT was estimated to be <60%. Conclusion: Surveillance of DLM is an acceptable strategy when the diagnosis of DLM is made through MRI. NHB and ICER favor HR when the presence of DLM is determined by CT scan alone.

OP-I.33 SIMULTANEOUS RESECTION OF PRIMARY COLORECTAL CANCER AND SYNCHRONOUS LIVER METASTASES: A POPULATION-BASED STUDY S. Nanji1,2, W. J. Mackillop2,3,4, X. Wei3, C. M. Booth2,3,4 Department Of Surgery, Queen’s University, Kingston, ON; 2Department Of Oncology, Queen’s University, Kingston, ON; 3Department Of Public Health Sciences, Queen’s University, Kingston, ON; 4Division Of Cancer Care And Epidemiology, Queens University Cancer Research Institute, Kingston, ON

1

Background: The role of combined resection of primary colorectal cancer (CRC) and synchronous liver metastases (LM) is gaining interest. Here we describe management and HPB 2015, 17 (Suppl. 1), 1–81

Abstracts outcomes of patients in the general population managed with simultaneous or staged resection of the primary tumor and synchronous CRC LM. Methods: All cases of CRC in Ontario who underwent surgical resection of LM in 2002–2009 were identified using the population-based Ontario Cancer Registry. Synchronous disease was defined as having resection of CRC LM within 12 weeks of surgery for the primary tumor. Pathology reports were reviewed to identify extent of disease and surgery. Results: During 2002–2009, 1711 patients underwent resection of CRC LM; pathology reports were identified for 1252 cases. 283 patients had synchronous disease; 116 (41%) patients had simultaneous resections and 167 (59%) had a staged resection. For the simultaneous and the staged groups, mean number of liver lesions resected was 1.7 and 2.3 (p < 0.001), mean size of the largest lesion was 3.1 and 4.7 cm (p < 0.001), major hepatic resection (≥3 Couinaud segments) rate was 26% and 76% (p < 0.001) and the R1 resection margin rate was 10% and 8% (p = 0.46), respectively. 30- and 90-day post-operative mortality rates for simultaneous and staged groups were 0.9% and 2.4% (p = 0.65) and 3.5% and 4.2% (p = 1.00), respectively. Conclusions: Simultaneous resection of synchronous CRC LM is common in routine clinical practice. Compared to a staged approach, patients undergoing simultaneous resections had fewer and smaller liver metastases, less aggressive resections and comparable post-operative mortality.

OP-I.34 COMPARATIVE ANALYSIS OF LAPAROSCOPIC RESECTIONS OF POSTEROSUPERIOR SEGMENTS IN SEMIPRONE POSITION VS. LAPAROSCOPIC LEFT LATERAL SECTIONECTOMY IN SUPINE POSITION. DO WE NEED TO REDEFINE THE DEFINITION OF A MAJOR LAPAROSCOPIC LIVER RESECTION? M. D’Hondt1, E. Yoshihara1, D. Devriendt1, F. Vansteenkiste1, H. Pottel2 1 Dept. Of Digestive And HPB Surgery, Groeninge Hospital, Kortrijk, W-VL; 2Interdisciplinary Research Center, Catholic University Leuven, Campus Kortrijk, Kortrijk, W-VL Introduction: The Louisville-statement defined laparoscopic resections of posterosuperior segments (LPSS) as major hepatectomies. It has been shown that LPSS, are associated with a good field of view,lower conversion rate and less blood loss when performed in semiprone position. All patients whom underwent LPSS at our center were positioned in semiprone since August 2011. The aims of this study were to assess differences in perioperative outcomes between laparoscopic left lateral sectionectomies (LLLS) performed in supine position and LPSS in semiprone. Methods: We reviewed a prospectively collected singlecenter database of all liver resections performed between August 2011 and August 2014. LLLS and LPSS were compared with respect to demographics and perioperative outcomes. Results: Thirty nine patients underwent LLLS(n = 19) or LPSS(n = 20). There were no differences in demographics HPB 2015, 17 (Suppl. 1), 1–81

33

(table) or maximal tumor diameter (p = 0.7569). There were no conversions. Pringle manoeuvre was not used in both groups.There was no difference in peroperative central venous pressure. Operative time in the LLLS group was 100 (60–160) min and 150 (100–270) min in the LPSS group (p = 0.0037) with median intra-operative blood loss in the LLLS group of 50(0–550) ml versus a larger 150(50–700) ml (p = 0.0191) for patients receiving LPSS. No patients required transfusion. Intraoperative and postoperative complication rate was similar in both groups. Mortality rate was nil in both groups. Median hospital stay was 6 days in both groups (p = 0.6382). Conclusion: LPSS in semiprone can be performed with similar clinical outcomes as a minor laparoscopic liver resection except for longer operative time and larger intraoperative blood loss without the need for transfusion.

OP-I.35 ASSOCIATING LIVER PARTITION AND PORTAL VEIN LIGATION IN STAGED HEPATECTOMY (ALPPS) IN SCANDINAVIA. A TRI-INSTITUTIONAL INTRODUCTORY FEASIBILITY STUDY B. I. Rosok1, E. Sparrelid2, B. Bjrnson3, B. A. Bjrnbeth1, B. Isaksson2, L. Lundgren3, E. Pomianowska1, T. Gasslander3, P. Sandstrm3 1 Section For HPB Surgery, Oslo University Hospital, Oslo, NONE; 2Department Of Surgery, Stockholm, NONE; 3 Surgical Department,, Linkg, NONE Introduction: ALPPS has been introduced as an alternative to conventional portal vein embolization or – ligation (PVE/ PVL) in patients with technically resectable liver tumors but insufficient future liver remnant. Initial experiences however, indicated that the complication rate and perioperative mortality following ALPPS exceeded that of PVE/PVL. Materials and Methods: Thirty patients (19 males 11 females) were operated during a 5 month period at our three institutions. Underlying diagnoses were colorectal liver metastases (n = 23), Cholangiocarcinoma (n = 4), HCC (n = 2) and Carolis syndrom The number of lesions varied from 1–20. None of the patients had underlying liver disease. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

34

Abstracts

Results: All patients completed the 2-stage procedure at a median of 8 days (7–15) following procedure 1. Twenty-three patients had extended right hepatectomies, wheras the remaining 7 had conventional right hemihepatectomies. Median hospital stay after the second procedure was 9.5 days (2–50). No perioperative mortality was observed. Complications according to Clavien-Dindo grading was 9 grade 1, 9 grade 2, 4 grade 3A and 3 patient with a grade 3b complication. Discussion/Conclusion: ALPPS may be an alternative to PVE/PVL in some patients. An acceptable complication rate can be obtained if certain selection criteria are met. Many patients have early recurrences following ALPPS, and our short and intermediate oncological results will be presented. Randomized controlled trials are necessary to define what patients may benefit from ALPPS. A multicentre Scandinavian trial, LIGRO (ClinicalTrials.gov NCT02215577) has therefore been initiated and has been enrolling patients since May 2014.

OP-I.36 LIVER ABSCESS: A REVIEW OF MANAGEMENT AND CLINICAL OUTCOMES AT WESTERN HEALTH, AUSTRALIA

OP-I.37 FACTORS PREDICITNG OUTCOMES IN NON-TRAUMATIC EMERGENCY HEPATECTOMYA NSQIP ANALYSIS

N. P. Kohli, J. Choi, S. T. Chan, V. Usatoff Western Health, Footscray, VICTORIA

J. A. Parikh1, S. Anantha Sathyanarayana1, S. Bendix1, M. J. Jacobs1, R. Kather2, I. S. Rubenfeld1 1 St.John Providence And Providence Park Hospitals, Southfield, MI; 2Henry Ford Hospital, Detroit, MI

Background: There is no clear consensus in the literature about definitive liver abscess management. This is the first study aimed to evaluate management and outcomes in Melbourne’s western suburbs. Study Design: A retrospective review of patients with pyogenic and amoebic liver abscess(s) captured using ICD-10 coding over the past five years at Western Health. Primary outcomes were success of management, predictive factors for surgical management, microbial aeitology of abscess and major complications (Clavien grade III to V). Results: Sixty-five patients (43 : 22 male to female) with median age 61 (50.75–72.00 Interquartile range; IQR). Fourteen (22%) were successfully managed with antibiotic treatment. Radiologically guided percutaneous drainage was performed in 41 (63%) cases with median time to drainage of 5.5 (1–11 IQR) days. It was successful in 32 of 41 cases (78%). Seven patients (11%) underwent surgery, of which 3 had failed drainage. Surgical predictive factors were found to be CRP of >100 on admission (Odds ratio 30.750, 95% CI 3.25–291.31, p = 0.003). The median length of stay was 15 (10.5–25.0 IQR) days. Forty-two patients had diagnostic investigations for the cause of their abscess (see Fig. 1). There were three deaths during this period of which two were due to sepsis in context of liver abscess. Conclusion: This study demonstrates that management with intravenous antibiotics and radiologically guided percutaneous drainage is successful in majority of cases. There were seven cases requiring surgery and a CRP >100 was shown to be predictive for need of surgery.

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

Introduction: Although non-traumatic emergent hepatectomies are rarely indicated, their burden to healthcare system in terms of utilization of resources is likely to be high, though has never been quantified. Methods: Using the ACS-NSQIP participant use files for 2005–2012, we identified hepatic resections by Current Procedural Terminology (CPT) code and segregated all non-traumatic hepatectomies into 2 groups: Emergent Hepatectomy (EH) and Non-emergent Hepatectomy(NEH). Preoperative, intraoperative and postoperative factors were analyzed to identify predictors of complications and mortality. Results: Of the 13227 non-traumatic hepatectomies from the NSQIP data, 137 emergency hepatectomies were identified. African Americans required significantly increased EH (2.0% vs 0.92%, OR 2.2, p < 0.001). The most common diagnosis for EH overall, was primary and secondary malignant neoplasm of the liver (38%; n = 33). Preoperative and perioperative transfusion requirements were higher in the EH group compared to NEH (17.5% vs 0.49%, OR 42.5, P < .001; and 52.7% vs 26.4%, OR 3.1, P < .001). Patients in the EH group were significantly more likely to experience a Clavien 4 complication (19.7% vs 7.2%, OR 3.2, p < .001). Mortality rate was higher in the EH group compared to NEH(8.8% vs 2.5%, OR 3.7, P < .001).A multivariate logistic regression analysis revealed ASA score, ascites, and emergent indication as poor outcome indicators. Surprisingly age and length of operation were not significant factors. (Table 1). Conclusion: Emergent hepatectomy has a significantly higher perioperative blood transfusion requirement, with increased morbidity and mortality rate. ASA score, ascites, and emergent indication as poor outcome indicators, while age and length of operation were not significant factors. HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

OP-I.38 DETECTION OF INVISIBLE LIVER TUMORS USING REAL-TIME VIRTUAL SONOGRAPHY A. Miyata, Y. Mise, T. Aoki, J. Kaneko, Y. Sakamoto, K. Hasegawa, Y. Sugawara, N. Kokudo The Univercity Of Tokyo Hospital, Tokyo, TOKYO Introduction: Real-time virtual sonography is an innovative imaging technology that synchronizes an intraoperative ultrasonography (IOUS) with preoperative computed tomography (CT). We validated the effectiveness of the navigation system in 2 cases to locate liver tumors, which had been found in preoperative images and were difficult to detect using conventional intraoperative inspection. Case 1: A 65-year-old man had a metastatic liver tumor from renal cell carcinoma. Preoperative CT revealed that the tumor was 6 mm in size, located in the deep segment 6. Intraoperatively, the tumor was not detected using plain and enhanced ultrasonography. Real-time virtual sonography projected the CT image of the tumor on IOUS image, which helped us to add mobilization of the liver. After mobilizing the liver towards the tumor location, we could find and resect the tumor using additional enhanced IOUS. Case 2: A 58-year-old man had a 6 mm-hepatocellular carcinoma located in segment 4. He had a previous history of left lateral sectionectomy for hepatocellular carcinoma. The recurrent tumor located in the previous cut surface of the liver was not found by conventional IOUS because of the severe adhesion. We used real-time virtual sonography, which navigated us to dissect the adhesion towards the tumor. Conclusions: A novel navigation system using real-time virtual sonography is helpful to locate small tumors that are difficult to find using conventional intraoperative inspection.

HPB 2015, 17 (Suppl. 1), 1–81

35

OP-I.39 THE NEED FOR HEPATOPANCREATOBILIARY SURGEONS: ARE THE COMMUNITY HOSPITALS UNDERSERVED? S. Anantha Sathyanarayana, S. Randhawa, P. Annigeri, G. Marshall, E. Negussie, M. J. Jacobs, J. A. Parikh St.John Providence And Providence Park Hospitals, Southfield, MI Introduction: Surgical educators have recently questioned if too many Hepato-Pancreato-Biliary (HPB) surgeons are being trained. While academic centers may be saturated, many community hospitals may be underserved. Thus, we sought to determine the need for an HPB surgeon at a tertiary care community hospital. Methods: All abdominal computed tomography (CT) scans from February 2014 to May 2014 performed at a community teaching hospital were reviewed and scans with pertinent HPB pathology were isolated. Results: A total of 389 CT scans having pertinent HPB pathology were identified from 3500 scans, for which an HPB surgeon consultation would be appropriate (Table 1). Out of the 291 patients with liver specific pathology, 17 patients had hepatic cysts >4 cm in size, 76 had a solid mass. Eighty three patients were found with pancreatic pathology, out of which 17 patients had cystic lesions >1 cm, 21 of them had a solid mass, 13 had non-specific main duct dilatation and 9 had chronic pancreatitis. Fifteen patients had biliary pathology including 11 with biliary ductal dilatation, one choledochal cyst and one extrahepatic bile duct stricture. For the 3 month study period, a total of 178 patients with significant HPB pathology were identified and the projected volume of patients will be over 700 for a period of 1 year that would require an HPB surgeon consultation. Conclusions: The national need for HPB surgeons should be re-evaluated based on the workload at the community hospital setting and cannot be based on saturation at the academic hospitals.

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

36

Abstracts 4 (8.2%) versus 0 in the pre-POH group. Deaths were due to amioderone related pulmonary fibrosis (1), cardiac arrest (1), and severe multi-organ failure related to a leak(1) and a postoperative bleed (1). Conclusion: The introduction of a hospitalist who specializes in perioperative management of high-risk surgical patients was associated with a decrease in 30-day readmissions and a similar overall complication rate, even in the setting of a higher mean ASA. There was a trend toward a higher mortality rate in the post-POH cohort. More study is required to understand the overall quality and financial impact of POH co-management.

OP-I.40 EVALUATING THE IMPACT OF ADDING PERIOPERATIVE HOSPITALIST CO-MANAGEMENT ON OUTCOMES FOR PATIENTS UNDERGOING PANCREATICODUODENECTOMY

OP-I.41 PERSONALITY TRAITS COMMON AMONG HEPATO-PANCREATO-BILIARY SURGEONS AND THEIR RELATIONSHIP TO JOB SATISFACTION

M. Brown, P. M. Campbell, R. F. Wolf, W. C. Johnston, M. A. Cassera, C. W. Hammill, P. H. Newell, P. D. Hansen Providence Portland Medical Center, Cancer Center, Portland, OREGON

S. N. Osayi1, L. Yu2, J. Drosdeck1, C. E. Ellison1, M. Bloomston1, C. Schmidt1, M. Dillhoff1, S. Weber3, P. Muscarella1 1 Department Of Surgery, The Ohio State University Wexner Medical Center, Columbus, OH; 2Center For Biostatistics, The Ohio State University, Columbus, OH; 3Department Of Surgery, University Of Wisconsin, Madison, WI

Introduction: We hypothesize elderly patients, with multiple comorbidities, undergoing high risk surgical procedures will benefit from pre- and postoperative co-management by a hospitalist who specializes in this field. We report clinical outcomes for two cohorts of patients undergoing pancreaticoduodenectomy (PD), pre and post introduction of a perioperative hospitalist (POH) program. Methods: Data was collected retrospectively on 89 consecutive patients undergoing PD between 2012 and 2014. Analysis was performed on 40 patients prior and 49 patients after the introduction of the POH program. Groups were compared by chi-square and T-test. Results: Results are summarized in the table below. 14 patients in the post-POH cohort were not seen by the POH due to patient selection, distance and transportation issues. Although the ASA was significantly higher in the post-POH group, overall complication rates were similar. 30-day readmissions were also significantly lower in the post-POH group. There were 4 deaths observed in the post-POH group, © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

Background: Personality may influence career choice, and can predict job satisfaction. The purpose of this study was to evaluate the personality traits of hepato-pancreato-biliary (HPB) surgeons and to determine whether they correlate with job satisfaction. Methods: A web-based survey of surgical members of the AHPBA was conducted. Personality traits and job satisfaction were assessed using the Big Five Inventory and the Brief Index of Affective Job Satisfaction instrument. Results: One hundred and thirty-six HPB surgeons completed the survey, 22% response rate (86.8% male, 70.7% >40 years old, 85.3% completed fellowship training, 50% >10 years in practice, and 76.5% academic practice). HPB surgeons scored higher on extraversion and conscientiousness, and lower on neuroticism (p < 0.001) compared to a normal population sample (n = 71,867). High extraversion HPB 2015, 17 (Suppl. 1), 1–81

Abstracts p = 0.02) and low neuroticism (p < 0.001) independently correlated with job satisfaction. Job satisfaction was higher among females (p = 0.004). While 91.9% of respondents indicated that they would choose the specialty again, only 53.7% would recommend it to their child/family. Those who would choose the specialty again were less neurotic (p = 0.039) and more satisfied with their job (p = 0.003). Additionally, those who would recommend the specialty were more agreeable (p = 0.001), more satisfied (p = 0.002), have been in practice longer (p = 0.006), and were more likely to choose the specialty again (p < 0.001). Conclusion: Extraversion and neuroticism correlate with job satisfaction among HPB surgeons. Furthermore, female HPB surgeons appear to have higher levels of job satisfaction. These findings may aid in the recruitment of HPB trainees and may have implications for job performance and patient care.

OP-I.42 IMPACT OF FOCUSED NURSING EDUCATION CURRICULUM IN THE CARE OF HEPATO-PANCREATOBILIARY (HPB) SURGICAL PATIENTS J. Drummond, C. Aviles, A. Cochran, E. Baker, R. Seshadri, J. Martinie, D. Iannitti, R. Swan Carolinas Healthcare System, Charlotte, NC Nurses provide point of contact care for HPB surgery patients. To improve patient education, nurse-physician communication, and nursing comfort with complex HPB patient care, we offered a six hour nursing targeted, educational course, to all nurses from inpatient units at a large, tertiary care hospital. Topics included anatomy, pathology, surgical procedures, nutrition, and pre/post-operative care of the HPB surgical patient. Two weeks prior to the course, a 30 question pre-test was distributed to all participants: 20 questions regarding HPB disease processes, 4 regarding the participants’ background, and 6 describing comfort level, physician communication, and experience caring for HPB patients. The same test was given to participants following the course and scores were compared. Descriptive statistics were performed, survey results tallied. 59 nurses participated in the course: 50 completed the pre-test and 25 completed the post-test. 50% of respondents reported nursing experience of less than 5 years. 68% reported working with HPB surgery patients over half of their shifts. Average pre-test score was 11.1 (55.6%) and post-test was 12.5 (62.5%), an increase of 13.5% (p < 0.01). Nursing confidence, comfort, and communication in the care of HPB surgery patients increased by 37.1% (p < 0.01). The percentage who felt completely confident in answering patient questions regarding HPB diseases more than doubled (38.9% to 84.1%). Our focused HPB nursing core curriculum course was associated with improved understanding of HPB anatomy and disease processes and increased nursing confidence in caring for HPB surgical patients. We hope this will translate to improved patient care and nursing-physician communication.

HPB 2015, 17 (Suppl. 1), 1–81

37

OP-I.43 OPEN DATA FOSTERS QUALITY OUTCOMES FOR HPB SURGERY: RESULT OF 1625 PANCREATIC AND HEPATIC RESECTION C. K. Chang, S. H. TEH, P. Fuchshuber, J. S. Choi, P. D. Peng, B. L. Bolinger, C. E. Binkley, R. M. Ramirez, G. B. Kazantsev, A. L. Spitzer, K. Kojouri, E. E. Rosas, M. M. Mortenson, C. A. Perez, M. A. Schlieman Kaiser Permanente Northern California Hepatico-Biliary-Pancreas Collaborative, Walnut Creek, CA Introduction: Healthcare economics is driving hospital to deliver improved “pay for performance” for all surgical programs. Much of the data regarding efficiency and performance is not actionable on an individual surgeon level as it is kept mostly anonymous. We hypothesize that the development of the Center of Excellence initiative for HPB surgery within KP integrated health care system which comprised of 21 medical centers, can improve surgical outcome by firstly improve individual surgeon performance. Methods: All HPB cases from 2008 to 2014 were retrospectively analyzed. Three-time periods were chosen, and data were given to all HBP surgeons. The first set of data served as a baseline to provide transparency of operative time (OR) and length of stay (LOS). The second set of data was prospective obtained during the initiation of program development The third set of data showed the result of the implementation of such program including a bi-weekly conference to discuss multifaceted best practices. Results: See Graph Conclusion: Transparency of surgical data allows surgeons to self-identify potential surgical outliers among their peers. A continuous open discussion of “best practice” in terms of detail surgical technique, intra-operative management and post-operative allows a gradual transition of “coaching” away from their personal preference. A critical review of data and discussion of best practice allows for systematic change in efficient surgical technique as well as discharge criteria.

OP-I.44 DUODENAL NEUROENDOCRINE TUMORS – LOCATION MATTERS M. R. Sheikh, H. Osman, S. Cheek, S. Hunter, D. R. Jeyarajah Methodist Dallas Medical Center, Dallas, TX Objective: Duodenal neuroendocrine tumors are rare. Historically, when feasible a less aggressive surgical approach is always considered to treat these tumors. The aim of this study was to identify factors associated with necessity for more aggressive surgical procedures. Method: All patients who underwent surgery for duodenal neuroendocrine tumor between September 2005 and June 2014 have been identified retrospectively in our database. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

38

Abstracts

Data collected included clinical presentation, operative findings and histopathological data. Results: 18 patients were identified that underwent surgical management for duodenal endocrine tumors. This included 2 patient with transduodenal excision (11%), 2 patients with duodenal resection (11%), 6 patients had antrectomy (33%) and 8 underwent pancreaticoduodenectomy (44%). On analysis, peri-ampullary location was the most common site of duodenal endocrine tumors (n = 9, 50%). 77% (n = 7) of peri-ampullary lesions led to pancreaticoduodenectomy. The odds of having a pancreaticoduodenectomy is 10 times higher when the lesion is in peri-ampullary location. 6 patients had positive lymph nodes. The odds of having a positive lymph node are almost 9 times higher when the lesion is in ampulla. 83% (n = 5) of tumors with positive lymph nodes were greater than T1 stage. The odds of having positive lymph node is 3 times higher when lesion is greater than T1. Conclusions: Ampullary location of neuroendocrine tumor in duodenum is associated with higher odds of lymph node positivity and need for treatment with more extensive procedures like pancreaticoduodenectomy.

Conclusions: Early post-operative hypophosphatemia is an independent predictor of LRC. A simple LRC risk prediction tool that includes this variable accurately identified low-risk patients and may help identify those most likely to benefit from enhanced postoperative recovery pathways.

OP-I.45 EARLY POST-OPERATIVE HYPOPHOSPHATEMIA AS A NOVEL PREDICTOR OF ANASTOMOTIC FAILURE AFTER PANCREATIC RESECTION: A RISK-PREDICTION TOOL

OP-I.46 SURVIVAL FOLLOWING PANCREATICODUODENECTOMY FOR STAGE 1A PANCREATIC ADENOCARCINOMA IS NOT IMPROVED BY MULTIMODALITY TREATMENT

E. Sadot, L. Roach, C. A. McIntyre, P. J. Allen, A. A. Eaton, M. I. D’Angelica, R. P. DeMatteo, T. P. Kingham, Y. Fong, W. R. Jarnagin Memorial Sloan Kettering Cancer Center, New York, NEW YORK

K. T. Ostapoff, P. Thirunavukarasu, B. W. Kuvshinoff, S. J. Nurkin, S. N. Hochwald Roswell Park Cancer Institute, Buffalo, NY

Introduction: Leak-related complications (LRC) remain serious potential sequela of pancreatic resection. Current LRC risk assessment is inadequate and rarely affects management algorithms. Hypophosphatemia appears to correlate with infective complications after some abdominal operations. This study evaluates early post-pancreatectomy hypophosphatemia as a predictor of LRC. Methods: Consecutive patients who underwent pancreaticoduodenectomy or distal pancreatectomy were analyzed. LRC were defined as pancreatic leak, fistula, or abscess; only grade 2 or higher LRC were recorded. Postoperative serum phosphate levels and other recognized LRC risk factors (duct diameter, soft pancreatic parenchyma, high-risk pathology, excessive blood loss, procedure type, and preoperative chemotherapy) were analyzed. Factors significant on multivariate analysis were used to construct an LRC risk prediction model. Results: From 2011 through 2012, 465 patients were included with a median age of 66 years. LRC (grade >2) were recorded for 85 patients (18%). Univariate analysis identified the following predictors of LRC: hypophosphatemia on postoperative day 3 (p = 0.006), small duct diameter (p = 0.007), soft gland consistency (p = 0.002), and intra-operative blood loss >400 cc (p = 0.01). Hypophosphatemia on postoperative day 3(OR = 2.2, CI: 1.1–4.5), soft gland consistency(OR = 3.1, CI: 1.7–6), and intra-operative blood loss >400 cc(OR = 2.3, CI: 1.3–4) remained significant on multivariate analysis and were used to construct an LRC risk prediction tool, which had a negative predictive value of 93% and a c-index of 0.68(Table 1). © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

Introduction: Pancreaticoduodenectomy is an integral part of pancreatic adenocarcinoma treatment. NCCN guidelines recommend patients undergo multimodality therapy, but few studies validate its significance in early stage patients. Methods: Using NCDB from 1998–2006, patients who underwent a pancreaticoduodenectomy were identified. Patients with invasive histology and stage 1 disease were included. X2 test was used for categorical variables. Median survival was estimated using Kaplan-Meier method with logrank comparison and Cox regression. Results: Over 8 years, 2,801 patients were identified. Median overall survival for Stage 1A (n = 1047) and Stage 1B (n = 1776) was 31.8 and 23.2 months with a median overall follow-up of 23.7 months (0–172.9). Stage 1B patients were more likely to have higher grade tumors (p = 0.002), receive chemotherapy (p = 0.007) or radiation (p = 0.002) and have positive margins (p < 0.0001). There were no differences between groups with respect to lymph node yield (LN) or type of treatment facility. For both stages, chemotherapy, radiation, LN yield ≥17 nodes, age <70, tumor grade and facility type were associated with an improved overall survival on univariate analysis. However on multivariate analysis, for Stage 1A patients only LN yield ≥17 was independently predictive of survival while chemotherapy, radiation, facility type, age and sex had no impact on survival. For patients with Stage 1B, chemotherapy, age <70 and LN harvest ≥17 were associated with an improved overall survival on multivariate analysis. Conclusion: Despite guidelines recommending multimodality treatment for pancreatic adenocarcinoma, there is no improvement in survival for patients with Stage 1A disease. A surgery only approach should be considered for these patients. HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

FRIDAY, MARCH 13, 2015, 6:30PM–7:30PM COCKTAIL VIDEO PRESENTATION VC.01 TOTAL LAPAROSCOPIC CENTRAL PANCREATECTOMY WITH PANCREATIGOGASTROSTOMY FOR HIGH RISK CYSTIC NEOPLASM L. Schwarz, J. B. Fleming, M. H. Katz, J. E. Lee, T. A. Aloia, J. Vauthey, C. H. Conrad UT MD Anderson Cancer Center, Department Of Surgical Oncology, Houston, TEXAS Background: Organ-sparing pancreatic resection is important in prophylactic surgery for cystic neoplasms. There is controversy regarding the optimal surgical approach for pancreatic lesions in the neck or proximal body of the pancreas. Central compared to distal pancreatectomy is technically more challenging but preserves more functional pancreatic tissue. Due to the prophylactic nature of the surgery and long survival of patients with benign and borderline malignant lesions, surgeons need to stratify greater importance to surgical morbidity and sparing pancreatic parenchyma. Patient: The patient is a 59-year-old active woman with a symptomatic cystic neoplasm of the pancreas exhibiting high risk imaging features. The cyst of 2.2 × 1.8 cm in the body of the pancreas was impinging on the portal venous confluence. Technique: The patient was positioned in the French Position, the lesser sac was opened and the pancreatic body exposed. A retropancreatic tunnel was created with staple division of the neck. The body was mobilized off the portal vein and splenic vessels transected. A retrogastric pancreaticogastrostomy was sewn through an anterior gastrotomy. The stent was delivered past the pylorus to decrease pancreatic enzymatic activation. Conclusion: Laparoscopic ultrasound helps in defining cyst borders and minimal blood loss optimizes visualization during the dissection. A minimally invasive pancreaticogastrostomy created through an anterior gastrotomy is technically feasible and safe. This approach can minimize the morbidity of prophylactic pancreatic surgery for patients with cystic neoplasms. Nevertheless, it should not compromise safety, oncologic completeness or an organ-sparing approach.

39

SATURDAY, MARCH 14, 2015, 7:30AM–9:30AM LONG ORAL F – OUTCOMES LO-F.01 COST VARIATION IN LAPAROSCOPIC CHOLECYSTECTOMY AND ASSOCIATION WITH OUTCOMES ACROSS A SINGLE HEALTH SYSTEM: IMPLICATIONS FOR STANDARDIZATION AND IMPROVED RESOURCE UTILIZATION D. Brauer, R. C. Fields, W. G. Hawkins, S. M. Strasberg, L. M. Brunt, N. Mercurio, D. P. Jaques, B. L. Hall Barnes-Jewish Hospital, Washington University School Of Medicine, St. Louis, MO Background: Payers and regulatory bodies are increasingly placing emphasis on cost containment, quality/outcome measurement, and transparent reporting. Significant cost variation occurs in many operative procedures without a clear relationship with outcomes. Clear cost-benefit associations will be necessary to justify increased expenditures in the era of bundled payment structures. Hypothesis: Operating room (OR) supply cost variation in laparoscopic cholecystectomy (LCCK) is not associated with improved outcomes. Methods: All LCCKs performed within a single health system over a one-year period were analyzed for OR supply cost. Results were obtained for individual surgeons and system hospitals. Costs were correlated with NSQIP outcomes. Results: From July, 2013–June, 2014, 2,178 LCCKs were performed by 55 surgeons at 7 hospitals. The median case OR supply cost was $513 ± 156. There was significant variation in cost between individual surgeons, hospitals, and within an individual surgeon’s practice (Figure). There was no significant correlation between cost and individual surgeon volume, hospital, or NSQIP outcomes. The majority of cost variation was explained by selection of trocar and clip applier constructs. Conclusions: Significant case OR cost variation is present in LCCK across a single health system. Differences in cost are not associated with individual surgeon volume or hospital and there is no clear association between increased cost and NSQIP outcomes. Placed within the larger context of overall cost, opportunity exists for OR case standardization and savings with no obvious risk for a reduction in quality of care, which is critical in the era of bundled payment structures. [Correction added on 24 February 2015, after online publication. D. Brauer was added to the list of authors.]

HPB 2015, 17 (Suppl. 1), 1–81

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

40

Abstracts

LO-F.02 ANALYSIS OF LYMPH NODE POSITIVE PATIENTS IN SURGICALLY TREATED INTRAHEPATIC CHOLANGIOCARCINOMA: A REVIEW OF THE NATIONAL CANCER DATABASE Z. Jutric, C. W. Johnston, H. M. Hoen, P. H. Newell, M. A. Cassera, C. W. Hammill, R. F. Wolf, P. D. Hansen Providence Portland Cancer Center, Portland, OR Introduction: The role of routine lymphadenectomy in the surgical treatment of intrahepatic cholangiocarcinoma (ICC) has been poorly defined. Recent studies have recommended consideration of lymphadenectomy given its prognostic implications. We aim to define predictive indicators of survival in patients with positive lymph nodes. Methods: The National Cancer Data Base (NCDB) was queried for patients who underwent surgical resection for ICC between 1998 and 2011. Single predictor univariate analyses were performed on 23 variables including demographics, tumor characteristics, surgery outcomes and adjuvant therapy details. Both single predictor univariate and multivariate Cox proportional hazards survival analysis were then performed on 160 patients identified to have positive lymph nodes. Results: Of 823 patients with complete data, 57% had at least one lymph node examined. Median survival for lymph node negative patients was 37 months versus 15 months for lymph node positive patients, results shown. Other univariate factors associated with decreased survival include male sex, urban location, tumor size, grade and positive margin status. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

In lymph node positive patients, multivariate analysis showed poorer survival in the patients not receiving chemotherapy or radiation (HR 1.43, p = .006), tumor size >5 cm compared to <5 cm (p = .018), and older age (p < .0001). Lymph node positive patients
HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

41

LO-F.03 NATURAL HISTORY AND TREATMENT TRENDS IN HEPATOCELLULAR CARCINOMA SUBTYPES: INSIGHTS FROM A NATIONAL CANCER REGISTRY P. L. Jernigan, K. Wima, D. J. Hanseman, R. S. Hoehn, A. Ertel, E. Midura, I. M. Paquette, S. A. Ahmad, S. A. Shah, D. E. Abbott Department Of Sugery, University Of Cincinnati School Of Medicine, Cicinnati, OH Introduction: Histopathological advancements have enabled more sophisticated characterization of hepatocellular carcinoma (HCC), but the clinical significance of these distinctions is incompletely understood. Our aim was to investigate pathologic and treatment differences between HCC variants. Methods: The American College of Surgeons National Cancer Data Base (1998–2011) was queried to identify 784 patients with surgical management of six HCC subtypes (1.8% of all HCC patients): fibrolamellar (FL, n = 206), scirrhous (Sc, n = 29), spindle cell (Sp, n = 20), clear cell (CC, n = 169), mixed type (M, n = 291), and trabecular (T, n = 69). Chi-square, Kaplan-Meier and Cox regression analysis were used to identify associations between demographic, tumor and treatment-specific variables and clinical outcomes, namely overall survival (OS). Results: Patients with FL-HCC were younger than other variants (median age 27 vs. 54–61, p < 0.001), more commonly female (56.3%, p < 0.001), and less likely to receive a transplant (3.66%, p < 0.001). Patients with FL- and Sp-HCC presented more frequently with larger tumors (>5 cm, p < 0.001) and node-positive disease (p < 0.001). Median OS (years) of the six subtypes was: FL 2.99 (CI 2.08–4.55), Sc 4.12 (1.47–7.01), Sp 0.56 (0.13–1.04), CC 2.99 (2.08–4.55), M 2.22 (1.61–2.7), and T 2.91 (1.88–8.86). On multivariate analysis, better OS was associated with FL-HCC, lower pathologic stage, node-negative disease, and liver transplant (Table 1). Eighty-six patients (11%) received adjuvant therapy, which was not associated with better OS. Conclusion: These data represent the largest series of recognizable HCC variants, demonstrating distinct differences in presentation and natural history. These findings can help clinicians and patients discuss treatment decisions and prognosis for rare clinical entities.

HPB 2015, 17 (Suppl. 1), 1–81

LO-F.04 UNDERSTANDING DRIVERS OF COST VARIATION FOR EPISODES OF CARE AMONG PATIENTS UNDERGOING HEPATOPANCREATOBILIARY SURGERY G. Spolverato, A. Ejaz, N. Ahuja, K. Hirose, M. A. Makary, C. L. Wolfgang, M. J. Weiss, J. Cameron, T. Pawlik Johns Hopkins Hospital, Baltimore, MD Background: Understanding factors associated with variation in procedure-related costs may help identify means to increase savings. We sought to define potential variation in hospital charges associated with hepatopancreatobiliary (HPB) surgery. Methods: Patients who underwent a HPB procedure between 2009–2013 were identified. Perioperative morbidity was ascertained through ICD-9 codes. Total hospital charges were tabulated for room&board, surgical/anesthesia services, medications, laboratory/radiology services, and other charges. Results: 2,545 patients underwent either a pancreas(66.8%) or liver/biliary(33.2%) resection. Mean total charges for all patients were $42,357 ± $33,745(pancreas: $46,352 ± $34,932 vs. liver: $34,303 ± $29,639;P < 0.001). Major morbidity(pancreas, range: 7%-18%; liver, range: 9%18%) and observed : expected(O : E) length-of-stay(LOS) (pancreas, range: 0.67–1.64; liver, range: 1.06–3.35) varied among providers(both P < 0.001). While a perioperative complication resulted in increased total hospital charges(complication: $66,401 ± $55,124 vs. no complication: $39,668 ± $29,250;P < 0.001), total charges remained variable even among patients who did not experience a complication(P < 0.001). For example, mean total charges for a pancreaticoduodenectomy ranged from $41,413$49,543 vs. $29,376-$43,420 for a partial hepatectomy (Figure). Surgeons within the lowest quartile of O : E LOS had lower total charges($33,879 ± $27,398) versus surgeons in the highest quartile($49,498 ± $40,971)(P < 0.001). Surgeons with the highest O : E LOS had higher across-theboard charges (operating room, highest quartile: $10,514 ± $4,496 vs. lowest quartile: $7,842 ± $3,706; medication, highest quartile: $1,796 ± $3,799 vs. lowest © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

42

Abstracts

quartile: $925 ± $2,211; radiology, highest quartile: $2,494 ± $4,683 vs. lowest quartile: $1,424 ± $3,247; laboratory, highest quartile: $4,236 ± $5,991 vs. lowest quartile: $3,028 ± $3,804; all P < 0.001). Conclusions: After accounting for in-hospital complications, total mean hospital charges for HPB surgery remained variable by case type and provider. While the variation in cost was associated with LOS, provider-level differences in across-the-board charges were also noted.

LO-F.05 AN ECONOMIC ANALYSIS OF PANCREATICODUODENECTOMY: SHOULD COSTS DRIVE CONSUMER DECISIONS? T. B. Tran, M. M. Dua, D. J. Worhunsky, G. A. Poultsides, J. A. Norton, B. C. Visser Stanford University School Of Medicine, Stanford, CA Background: Consumer groups campaign for cost transparency believing that patients will select hospitals accordingly. The aim of this study was to determine whether the “cost” of a Whipple might be considered in choosing a hospital. Methods: Using Nationwide Inpatient Sample Database, we analyzed charges for patients who underwent pancreaticoduodenectomy (PD) from 2000–2010. Outcomes were compared between high volume hospitals (HVH, >20/yr) and © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

low volume hospitals (LVH). The relationship of inflationadjusted charges and outcomes were stratified by uncomplicated PD (length of stay <14 days) vs complicated PD (>14 days). Results: A total of 15,599 PD were performed in 1,186 hospitals at a median cost of $87,444 (IQR $60,015$144,869). While only 94 (8%) hospitals performed >20 PD/year, 57% of all PD were performed in these HVH. HVH had shorter hospital stay (11 vs 15 days, p < 0.001) and mortality (3% vs 7.6%, p < 0.001). PD performed at LVH had higher median charges compared to HVH ($97,923 vs. $81,581, p < 0.001). The cost of uncomplicated PD was significantly lower than a complicated PD ($67,238 (IQR 51,112–91,401) vs $138,325 (IQR 95,206–224,919), p < 0.001). When comparing uncomplicated PD between HVH and LVH, the median cost was very similar ($67,389 vs $66,922). Among uncomplicated PD, a multivariate analysis controlling for demographics and co-morbidities revealed that cost did not affect the risk of mortality (OR 1). Conclusions: The cost of a Whipple (even uncomplicated) remains surprisingly variable. PD at HVH are associated with better outcomes, which is reflected in lower charges. But ultimately, patients should choose volume over price.

LO-F.06 THE IMPACT OF PERIOPERATIVE BLOOD TRANSFUSION ON SHORT-TERM OUTCOMES FOLLOWING PANCREATECTOMY: AN ANALYSIS FROM THE AMERICAN COLLEGE OF SURGEONS NATIONAL SURGICAL QUALITY IMPROVEMENT PROGRAM (ACS-NSQIP) J. Hallet1,2, A. L. Mahar3, M. Tsang2, C. H. Law1,2, N. G. Coburn1,2, P. J. Karanicolas1,2 1 Division Of General Surgery, Sunnybrook Health Sciences Centre – Odette Cancer Centre, Toronto, ONTARIO; 2 Department Of Surgery, University Of Toronto, Toronto, ONTARIO; 3Department Of Public Health Sciences, Queen’s University, Kingston, ONTARIO Background: While perioperative red blood cell transfusions (RBCT) are associated with worse outcomes in colorectal surgery, their impact following pancreatectomy HPB 2015, 17 (Suppl. 1), 1–81

Abstracts remains unclear. We sought to examine the association between RBCT and post-operative morbidity following pancreatectomy. Methods: Using the ACS-NSQIP database, we identified patients undergoing elective pancreatectomy from 2006 to 2012. Patients missing data on key variables were excluded. We compared post-operative morbidity and length of stay based on RBCT status using univariate and multivariate analyses. A sensitivity analysis was conducted excluding patients with higher baseline risk for RBCT. Results: From 21,132 pancreatectomies, we included 14,322 patients of whom 1624 (11.3%) received RBCT. Major morbidity (34.9% Vs. 21.6%; p < 0.0001) and mortality (15.7% Vs. 11.5%; p < 0.0001) were higher, and median length of stay was prolonged (15.7 Vs. 11.5 days; p < 0.0001) with RBCT. After adjustment for baseline characteristics including comorbidities, malignant diagnosis, procedure, and operative time, RBCT was independently associated with increased major morbidity (Relative Risk – RR 1.45; p < 0.0001), post-operative infections (RR 1.30; p < 0.001), thrombo-embolic events (RR 1.41; p = 0.01), cardiac events (RR 2.41; p < 0.0001), respiratory failure (RR 2.60; p < 0.0001), and mortality (RR 2.51; p < 0.0001). Length of stay was prolonged with RBCT (adjusted mean estimate 1.22; p < 0.0001). Excluding patients with higher baseline risk of RBCT did not substantially alter the results. Conclusion: Perioperative RBCT is independently associated with worse short-term outcomes and prolonged length of stay following pancreatectomy. This observation holds true in patients with lower baseline risk of RBCT. Comprehensive multidisciplinary strategies to minimize and rationalize the use of RBCT are warranted.

43

study was to evaluate the oncologic outcomes of patients with specified CRLM-EHD vs. CRLM-only and to determine the effect of PET-CT on disease-free (DFS) and overall survival (OS). Methods: This is a sub-study of a randomized trial studying the effect of PET-CT before liver resection on surgical management of CRLM with or without EHD (limited to specified resectable metastases: portal lymph nodes/lung/local recurrence) in patients with resected colorectal cancer from 2005– 2013. Survival data adjusted for PET-CT was analyzed using standard statistics. Results: From 404 patients in the trial, 25 had EHD at randomization [14 PET-CT group and 11 control (no PET-CT)]. The most common EHD-site was lung (n = 18). All EHD sites known at randomization were resected (PET-CT and control group). After median follow-up of 36-months, median DFS for CRLM-EHD was 5.9-months (95%CI: 3.6–11.7) and 16.2-months (95%CI: 13.7–18.9) for CRLM-only [unadjusted-HR: 3.03 (95%CI: 2.00–4.59)]; the estimated OS was similar between groups [36.7-months (95%CI: 26.8–43.4) vs. 40.7-months (95%CI: 40.5–57) respectively; unadjusted-HR: 1.68 (95%CI: 0.98, 2.89)]. After adjusting for the use of PET-CT, DFS for CRLM-EHD remained significantly worse compared to CRLM-only without significant differences in the adjusted OS. Conclusions: Metastasectomy for specified and limited CRLM-EHD is associated with similar OS to CRLM-only despite a lower DFS. The use of PET-CT prior to complete metastasectomy did not affect the time to recurrence or the OS of patients with CRLM-EHD.

LO-F.08 TIMING, INCIDENCE, AND RISK FACTORS ASSOCIATED WITH UNPLANNED POSTOPERATIVE HOSPITAL READMISSIONS IN THE HEPATO-PANCREATICO-BILIARY PATIENTS

LO-F.07 RESECTION OF COLORECTAL CANCER LIVER METASTASES IN THE SETTING OF EXTRAHEPATIC DISEASE; RESULTS FROM A RANDOMIZED TRIAL EVALUATING THE EFFECT OF PET-CT P. E. Serrano1, C. Moulton2, C. Gu1, C. H. Law3, L. Ruo1, K. Y. Gulenchyn1, D. Quan4, R. Fairfull Smith5, D. W. Jalink6, M. Husien7, J. A. Julian1, M. N. Levine1, S. Gallinger2 1 McMaster University, Hamilton, ON; 2University Health Network, University Of Toronto, Toronto, ON; 3Sunnybrook Health Sciences Centre And Odette Cancer Centre, Toronto, ON; 4London Health Sciences Centre, London, ON; 5The Ottawa Hospital, Ottawa, ON; 6Cancer Centre Of Southeastern Ontario, Kingston, ON; 7Grand River Regional Cancer Centre, Kitchener, ON Introduction: Selected patients with colorectal cancer liver metastases (CRLM) and extra-hepatic disease (EHD) are considered for curative surgery. The main objective of this HPB 2015, 17 (Suppl. 1), 1–81

G. C. Edwards1, L. Du1, J. R. Miller2, J. Ehrenfeld1, H. R. Mir1, Y. Shyr1, A. A. Parikh1, N. B. Merchant1, K. Idrees1 1 Vanderbilt University Medical Center, Nashville, TN; 2 Meharry Medical College, Nashville, TN Introduction: As part of the Affordable Care Act, the Hospital Readmissions Reduction Program (HRRP), has started to impose financial sanctions on hospitals with increased readmission rates. The purpose of this study is to define the incidence and identify peri-operative factors associated with 30 and 90-day readmission after hepatic (HR) and pancreatic resections (PR). Methods: HR and PR patients were retrospectively reviewed over an 8 year period. Pre-operative factors [patient demographics, ASA class, Charlson Comorbidity Index, Elixhauser Comorbidity Index], intra-operative factors [Surgical Apgar Score (SAS), operative duration] and postoperative factors [postoperative major complications (PMC), hospital length of stay (LOS), ICU LOS, discharge disposition] were evaluated. Multivariable Cox regression (MVR) analysis was used to examine associations for hospital readmission. Results: 30 and 90-day readmission rates in 878 patients are shown in the table. By MVR, PMC (HR = 8.9, 95%, © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

44

Abstracts

p = 0.0003), PR (HR = 2.3, p = 0.006), ICU LOS (HR = 1.1, p = 0.02) and SAS (HR = 0.75, p = 0.02) were associated with 30-day readmission. PMC (HR = 7.5, p = 0.0001) and PR (HR = 2.1, p = 0.007) were also associated with 90-day readmission. In patients without complications, ICU LOS (HR = 7.0, p = 0.03), PR (HR = 2.4, p = 0.006) and SAS (HR = 0.7, p = 0.04) were independently associated with 30-day readmission while only PR (HR = 2.1, p = 0.007) was associated with 90-day readmission. Conclusions: The 90-day readmission rate in HPB patients remains high (18.8%) and the majority of readmissions (88.8%) occur within 30 days of discharge. Identifying factors associated with an increased risk of readmission is critical in the development of interventions and resource utilization to help reduce unplanned readmissions and decrease costs within this population.

Conclusions: Psychiatric disease and substance abuse are highly prevalent among veterans with HCC. Most patients are surviving on very meager income. These profound socioeconomic and psychosocial problems must be recognized when providing care for HCC to this population to provide adequate treatment and surveillance.

SATURDAY, MARCH 14, 2015, 7:30AM–9:30AM LONG ORAL G – BASIC/ TRANSLATIONAL/EDUCATION LO-G.01 TAUROURSODEOXYCHOLIC ACID ALLEVIATES ISCHEMIA/ REPERFUSION INJURY IN STEATOTIC MOUSE LIVER C. D. Anderson, J. Zhang, N. Singh, W. Dorsett-Martin, T. M. Earl University Of Mississippi Medical Center, Jackson, MS

LO-F.09 MISSING THE OBVIOUS: PSYCHOSOCIAL OBSTACLES IN VETERANS WITH HCC K. J. Hwa1,2, M. M. Dua1,2, S. M. Wren1,2, B. C. Visser1,2 Stanford University School Of Medicine, Stanford, CA; 2 Palo Alto Veterans Affairs Hospital, Palo Alto, CA

1

Introduction: Socioeconomic disparities in patients with hepatocellular carcinoma (HCC) influence medical access and treatment. However, in addition to socioeconomic barriers, the veteran population suffers from significant psychosocial obstacles. This study identifies the often overlooked social challenges that veterans face while undergoing treatment for HCC. Methods: We retrospectively reviewed prospectively gathered data regarding social/behavioral hardships for 100 veterans at the Palo Alto VA who had been treated for HCC between 2009–2014 (50 consecutive patients who underwent resection and 50 treated with intra-arterial therapy). Results: Substance abuse history was identified in 96%, with 34% actively abusing alcohol or drugs. Half were unemployed. Most patients survived on very limited income (median $1340/mo, IQR $900–2125); 36% on <$1000/mo, 37% between $1001–2000/mo, and 27% with >$2000/mo. Fully, 37% were homeless at time of index treatment, which was more common in those with the lowest income (62% of <$1K/mo group, 32% of $1–2K/mo group, and 6% of >$2K/mo group, p < 0.05). Psychiatric illness was present in 64/100 (64%) patients; among these the majority (59/64, 92%) received ongoing psychiatric treatment. The top two diagnoses were depression and PTSD. Fifty-one percent of all patients had been incarcerated. Transportation was provided to 23% of patients who would otherwise have been unable to attend medical appointments. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

Background: Tauroursodeoxycholic acid (TUDCA) decreases endoplasmic reticulum (ER) stress, autophagy, and cell death in cultured rat hepatocytes. We hypothesized that TUDCA could reduce the injury caused by total warm ischemia reperfusion (WIR) in steatotic mouse liver. Methods: Male ob/ob mice underwent 100% hepatic warm ischemia by clamping the portal triad for 30 minutes. For the experiment group, 200 mg/kg TUDCA was injected IP 1 hour before the surgery. Animals were sacrificed at 12 hours and 48 hours after reperfusion. Quantitative real time PCR measured ER stress markers such as C/EBP homologous protein (CHOP), glucose regulated protein 78 (GRP78), protein kinase dsRNA-dependent-like ER kinase (PERK), and activating transcription factor-6 (ATF6). Western blot examined autophagy marker microtubule-associated protein 1 light chain 3 (LC3 II). ELISA determined interleukine-6 (IL6) levels (liver and serum). Results: Compared to controls, WIR increased ER stress in the liver [CHOP (∼3 fold, p = 0.004), GRP78 (∼4 fold, p = 0.001), PERK (∼2 fold, p = 0.005), and ATF6 (∼1.5 fold, p = 0.004)] at 12 but not 48 hours. LC3 II protein levels were increased at both 12 (∼3 fold, p = 0.019) and 48 hours (∼4 fold, p = 0.025). Serum IL6 levels were increased at 12 (∼40 fold, p = 0.034) and 48 hours (∼33 fold, p = 0.034). TUDCA treatment decreased LC3 II at 12 (p = 0.018) and 48 hours (p = 0.034), decreased serum IL6 at 12 (p = 0.025) and 48 hours (p = 0.025), and improved animal survival (median 26 hours vs 41 hours, p = 0.02). ER stress levels were not changed. Conclusion: TUDCA improves survival and reduces the inflammation following WIR in steatotic liver through a non-ER stress pathway.

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

LO-G.02 CHARACTERIZATION OF A PORCINE MODEL FOR ASSOCIATING LIVER PARTITION AND PORTAL VEIN LIGATION FOR STAGED HEPATECTOMY (ALPPS) K. P. Croome, S. A. Mao, J. M. Glorioso, S. L. Nyberg, D. M. Nagorney Mayo Clinic, Rochester, MN Background: Publications using the ALPPS procedure have demonstrated a future liver remnant(FLR) growth of 40–160% in only 6–9 days. The present study aimed to develop and describe the first large animal model of ALPPS that can be used for future studies. Methods: A total of 13 female domestic swine were studied. ALPPS stage 1 (portal vein division and parenchymal transection) was followed by ALPPS stage 2 (completion left extended hepatectomy) 7 days later. An abdominal CT scan was performed immediately prior to ALPPS stage 1 surgery and again 7 days later to assess hypertrophy immediately prior to ALPPS stage 2 surgery. Blood samples as well as tissue analysis were performed. Results: On CT volumetric analysis mean size of the FLR prior to ALPPS stage 1 was 21.4 ± 1.8% and 39.8 ± 4.6% prior to ALPPS stage 2. Median degree of hypertrophy was 74.5% with a median kinetic growth rate of 10.6% per day. Liver weights at autopsy correlated well with CT volumetric analysis(p = 0.65). There was no significant difference in mean lab values (AST,ALT,ammonia,INR or bilirubin) from baseline until immediately prior to ALPPS stage 2. Post ALPPS stage 2 there was a significant increase in INR from baseline 1.1 ± 0.1 and 1.6 ± 0.1 (p = 0.005), respectively. No post-operative deaths secondary to liver failure were observed. Conclusion: The present study describes the first reproducible large animal model of the ALPPS procedure. Degree of hypertrophy and kinetic growth rate were similar to that which has been demonstrated in human publications. This model will be valuable as future laboratory studies are performed.

HPB 2015, 17 (Suppl. 1), 1–81

45

LO-G.03 PREOPERATIVE ANAEMIA AND POSTOPERATIVE OUTCOMES AFTER HEPATECTOMY: A RETROSPECTIVE COHORT STUDY S. Tohme, P. Varley, M. Khreiss, A. Tsung University Of Pittsburgh, Pittsburgh, PA Background: Preoperative anemia is associated with adverse outcomes after surgery in general but outcomes after hepatectomy specifically are not well established. We aimed to assess the effect of preoperative anemia on 30-day postoperative morbidity and mortality in patients undergoing major hepatectomies. Methods: All elective hepatectomies for the period 2005– 2012 recorded in the NSQIP database were evaluated. We selected to study partial lobectomies, total left, total right, and trisegmentectomies and exclude minor procedures. We obtained anonymized data for 30-day mortality and morbidity, demographics, and preoperative and perioperative risk factors. We used multivariate logistic regression to assess the adjusted and modified effect of anemia, which was defined as (hematocrit <39% in men and <36% in women), on postoperative outcomes. Results: We obtained data for 13,198 patients, of whom 4,383(33.2%) had preoperative anemia. Postoperative mortality at 30-days was higher in patients with anemia than those without anemia (odds ratio[OR]2.15,95%CI 1.70– 2.71). Morbidity at 30-days was also higher in patients with anemia (for any complication1.93,1.79–2.09; for serious complications 2.05,1.90–2.22). After adjustment for predefined clinical and laboratory risk factors, postoperative morbidity was higher in patients with anemia than in those without anemia (adjusted OR any complication 1·48,1.36–1.61, serious complications 1.54,1.41–1.68). Postoperative mortality was similar in both groups after adjustment (1.09, 0.836–1.433). Conclusion: Preoperative anemia is independently associated with an increased risk of morbidity in patients undergoing hepatectomy. Therefore, it is crucial to readdress preoperative blood management in anemic patients prior to hepatectomy. Anemia was not an independent predictor of mortality which may be due to the rare event of 30-day mortality after elective hepatectomy.

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

46

Abstracts

LO-G.05 INTERACTION OF GLYCOGEN SYNTHASE KINASE-3 AND NOTCH1 IN PANCREATIC CANCER M. Kunnimalaiyaan, S. Kunnimalaiyaan, T. Gamblin Medical College Of Wisconsin, Milwaukee, WI Abstract: Glycogen synthase kinase-3 (GSK-3) can act as either tumor promoter or suppressor by its inactivation depending on the cell type. There are conflicting reports on the roles of GSK-3 isoforms and their interaction with Notch1 in pancreatic cancer. We hypothesize that GSK-3α stabilizes Notch1 in pancreatic cancer cells thereby promoting cellular proliferation. Methods: Pancreatic cancer cell lines MiaPaCa2, PANC-1, and BxPC-3 were treated with 0–20 μM of AR-A014418 (AR). Cell growth was determined by MTT assay and LiveCell Imaging. The levels of Notch pathway members (Notch1, HES-1, survivin, cyclinD1), phosphorylated GSK-3 isoforms, and apoptotic markers were determined by Western blot. Immunoprecipitation was performed to identify the binding of GSK-3 specific isoform to Notch1. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

Results: AR-A014418 treatment had a significant dosedependent growth reduction (p < 0.001) in pancreatic cancer cells compared to control. The growth suppression effect is due to apoptosis. Importantly, reduction in GSK-3 phosphorylation leads to a reduction in Notch pathway members. Over expression of active Notch1 in AR-A014418-treated cells resulted in negation of growth suppression. Immunoprecipitation analysis revealed that GSK-3α binds to Notch1. Conclusions: This study demonstrates for the first time that the growth suppressive effect of AR-A014418 in pancreatic cancer cells is mainly mediated by reduction in phosphorylation of GSK-3α with concomitant Notch1 reduction. GSK-3αappears to stabilize Notch1 by binding and may represent a target for therapeutic development. Furthermore, down regulation of GSK-3 and Notch1 may be a viable strategy for possible chemosensitization of pancreatic cancer cells to standard therapeutics.

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

LO-G.06 NOVEL CONCEPT OF ELECTROCOAGULATION & TUMOR CELL IMPLANTATION: CREATION OF MINIMALLY INVASIVE ORTHOTOPIC MURINE MODEL OF PANCREATIC CANCER J. S. Bhullar1, Y. Cozakov1, N. Varshney2, S. Bindroo1, S. Chaudhary1, J. Tilak3, M. Decker3, M. Jacobs1, V. K. Mittal1 1 Department Of Surgery, Southfield, MI; 2Department Of Pathology, Toledo, OHIO; 3Department Of Patient Care Research, Southfield, MI Background: Orthotopic murine models of pancreatic cancer represent an important tool for evaluating treatment strategies. Several genetically modified mouse tumors and xenograft models have been reported. Genetic models have unpredictable growth & variable waiting period, while orthotopic models are operative ones, difficult to create and result in irregular metastasis. There is a constant endeavor to create an orthotopic model which replicates the human disease process. Study Design: Orthotopic pancreatic tumors were induced in 20 SCID mice using a novel technique. Low dose electrocoagulation of pancreas under laparoscopic guidance (using Coloview-mouse colonoscope) with thin electrode, followed by injection of 0.1 cc BxPC3 pancreatic cancer cells was done (n = 12, study group). Control mice underwent electrocoagulation alone (n = 4, group 1) and tumor cell injection alone (n = 4, group 2). Mice were evaluated for tumor growth and metastasis by necropsy (4 and 8 week for experimental group; 8 weeks for control group). Results: Tumors were detected in 11/12 mice in experimental group, 1/4 in control group 2, and none in control group 1. Over time there was an increase in tumor growth, tumor volume, lymphovascular invasion of pancreas, with metastasis to lymph nodes and surrounding organs. Conclusions: We report a novel concept of tumor cell implantation at site of electrocoagulation of pancreas. Combined with the minimally invasive technique, yields a replicative orthotopic murine model of pancreatic cancer. Our model is minimally invasive, easy to create, and overcomes the limitations of the existing models while questions the possibility free floating tumor cell implantation at resection site.

HPB 2015, 17 (Suppl. 1), 1–81

47

LO-G.08 ADEQUACY OF HPB TRAINING: POTENTIAL DISCONNECT BETWEEN FELLOW AND PROGRAM DIRECTOR PERCEPTIONS? A. K. Bressan1, J. P. Edwards1, E. Dixon1, R. M. Minter2, D. R. Jeyarajah3, S. C. Grondin1, C. G. Ball1 1 Department Of Surgery, Foothills Medical Center And The University Of Calgary, Calgary, AB; 2Department Of Surgery, University Of Michigan, Ann Arbor, MI; 3 Methodist Dallas Medical Center, Dallas, TX Background: Hepatopancreatobiliary (HPB) fellowship programs have undergone recent significant changes with regard to training standards, case volume thresholds and multimodality educational platforms. The goals of this study were to (1) compare perspectives of residents and program directors on perceptions of readiness to enter practice and (2) identify core HPB procedures that require increased emphasis during training. Methods: This survey targeted program directors (PDs) and trainees participating in the Fellowship Council / AHPBA pathway. Demographics, education, and career plans were collected. A comparative analysis of PD and trainee opinions on their confidence to perform thirteen core HPB procedures was completed (p < 0.05). Results: The response rate was 88% for both fellows (21/ 24) and PDs (23/26). Amongst fellows, 72% believe there is an excessive number of trainees, 81% aim to work in university-based or academic institutions, 90% expect to have an HPB practice combined with non-HPB cases, and 95% do not plan to pursue additional training. For all thirteen HPB procedures, the volume of cases during training was more often considered good or excellent by PDs than by fellows. This difference reached statistical significance for: major hepatectomies (PDs: 87% vs. fellows: 57%, p = .042); pancreaticoduodenectomies (100% vs. 81%, p = .044); and laparoscopic distal pancreatectomies (78% vs. 43%, p = .029). Trainees also systematically rated their confidence to perform HPB procedures lower compared to PDs’ perception (Figure 1). Conclusions: This study provides insight into content domains which may require additional attention during fellowship to achieve an appropriate level of proficiency and confidence upon completion of training.

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

48

Abstracts

LO-G.09 WHAT TO EXPECT WHEN YOURE EXPECTING A HEPATOPANCREATOBILIARY SURGEON: SELF-REPORTED EXPERIENCE OF HPB SURGEONS FROM DIFFERENT TRAINING TRACKS S. G. Warner1, A. Alseidi2, J. C. Hong3, T. M. Pawlik4, R. M. Minter1 1 Departments Of Surgery And Medical Education, University Of Michigan Health System, Ann Arbor, MI; 2 Department Of Surgery, Virginia Mason Medical Center, Seattle, WA; 3Department Of Surgery, Medical College Of Wisconsin, Milwaukee, WI; 4Department Of Surgery, Johns Hopkins Hospital, Baltimore, MD Background: With a recent increase in fellowships offering HPB training through multiple routes, prospective trainees and employers must understand the differences between available HPB training pathways. This study highlights selfreported fellowship experience and current scope of practice across 3 different training pathways. Methods: A survey was disseminated to 654 surgeons – active AHPBA members and recent graduates of HPB, transplant-HPB, and surgical oncology fellowships using SurveyGizmo®. Descriptive statistics were calculated. Results: 416 (66%) surgeons responded. Most respondents were male (89%), and most (83%) practice in an academic setting. Table 1 demonstrates fellowship operative experience and current case mix in practice. MIS training was the most commonly identified training deficiency, with 47% HPB, 49% transplant, and 52% SSO-trained respondents in agreement. Ultrasound was also a commonly identified training gap with 34% HPB, 40% transplant, and 25% SSO-trained respondents in agreement. Non-HPB cases routinely performed in practice were most commonly GI surgery and general surgery (56% & 49%) for HPB-trained respondents, transplant and general surgery (87% & 21%) for transplant-trained respondents, and GI surgery and non-HPB surgical oncology (70% & 28%) for surgical oncology-trained respondents. Conclusions: HPB surgery fellowship training experiences vary by training pathway, though perceived deficiencies in MIS and US training are common across all pathways. Despite this variability, the ultimate scope of non-transplant HPB practice is similar across training pathways. Thus, selection of a training pathway may best be guided by the desired training experience and planned focus of other components of one’s future practice.

LO-G.10 TEACHING PREOPERATIVE PLANNING: A NEW EDUCATIONAL INITIATIVE N. Zilbert1, T. Lam2, S. Gallinger1, L. St. Martin2, C. Moulton1,2 1 University Of Toronto Department Of Surgery, Toronto, ON; 2The Wilson Centre, Toronto, ON Background: Previous research has identified the importance that expert surgeons place on preoperative planning. Currently the teaching and assessment of preoperative planning is limited. This video demonstrates a novel education initiative to teach surgical trainees strategies for preoperative planning for complex HPB procedures. Methods/Results: This video reviews one preoperative planning module for a case of a patient with a colorectal liver © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

metastasis requiring a left hemihepatectomy. There are five screens that make up the module. The introductory screen provides a brief clinical history. The second screen allows the trainee to review the patient’s preoperative imaging. The next screen displays two videos of attending surgeons reviewing the same preoperative imaging. The fourth screen displays intraoperative video clips narrated by the operating surgeon. Each video clip focuses on an issue that one of the attendings on the preceding screen. The final screen shows a video from the operating surgeon reviewing the key learning points for the case to reinforce these for the trainee. Conclusions: In conclusion this video demonstrates a novel strategy for teaching preoperative planning for HPB surgery. Following the completion of this module the trainee should have a deeper understanding of the issues and considerations that are relevant for left hemihepatectomies. However the ultimate strength of this program is that several different left hemihepatectomy cases will be presented in the same format, so collectively the trainee will gain a richer and more complete understanding of the procedure. The process will be repeated for the other index procedures for HPB surgery.

LO-G.11 UNDERSTANDING SURGICAL ANATOMY OF THE LIVER: THERES AN APP FOR THAT L. M. Postlewait, M. Konomos, J. A. Matlock, T. White, K. A. Delman, S. K. Maithel Carlos And Davis Center For Surgical Anatomy And Technique, Emory University, Atlanta, GA Introduction: Surgical anatomy of the liver is complex and difficult to visualize. The two-dimensional renderings available for trainees make translation to practical application challenging. Our aim was to create an interactive App to teach liver anatomy to improve trainee preparation for hepatic surgery. Methods: Liver model and animation storyboards were created from radiographic images by a certified medicalillustrator under the guidance of a hepatobiliary fellowshiptrained surgeon in the education center for anatomy and simulation at the author’s institution. Animations were completed in Adobe Photoshop and Illustrator. 3D polygonal models were completed with detailed attention to liver shape, vessel placement, and internal divisions. Interactivity of the App was designed in Unity3D. The animated video models were UVMapped and brought into Cinema 4D. The App was designed for use on the Apple iPad. Results: The App is an interactive model with a narrated video to teach liver anatomy. The liver rotates in space permitting a better understanding of its 3-D structure. Parenchyma can be removed to reveal vascular and biliary anatomy. A narrated video provides a detailed overview of hepatic anatomy with a logical progression from whole liver topography sequentially down to segmental detail. Initial feedback via internal assessment is exceptional. Conclusions: This portable, mobile-device based instrument is a novel educational tool to teach liver anatomy via an interactive approach. Studies to assess its educational utility are underway. Integration with cross-sectional imaging is planned to enhance clinical applicability. The current iteration is applicable for student, resident and fellow-level trainees. HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

49

LO-H.02 RESECTION IS NOT INFERIOR TO LIVER TRANSPLANTATION IN NODE NEGATIVE INTRAHEPATIC CHOLANGIOCARCINOMA O. C. Kutlu, S. Garcia, M. V. Williams TTU HSC Dept Of Surgery, Lubbock, TEXAS Introduction: Intrahepatic cholangiocarcinoma(IHCCC) is the second most common malignancy of the liver. Despite the increasing incidence, few studies have been published on therapeutic options and outcomes. Although survival benefits of transplantation are well established for HCC, there is little information on the outcomes between liver transplantation and resection for IHCCC. In this study we investigated the survival of IHCCC in a large population database and identified if there was a survival advantage of transplantation over resection for stage I and II tumors. Material Methods: SEER database was used to identify IHCCC patients. Patients diagnosed between 1990 and 2008, histologically proven IHCCC, T1 and T2 tumors, N zero, no metastasis, no radiotherapy, and not lost to follow up were included in the study. Analyses were performed using SPSS 20 with Kaplan-Meier statistics and Cox proportional hazards regression. Results: A total of 297 patients, 221 underwent resection and 76 underwent transplantation met the criteria. Mean survival for resection was 36.8 months and 41.1 months for transplantation. Survival for transplantation vs surgery is as follows, 80% and 78% at one year, 62% and 63% two years, 54% and 51% three years, 45% and 36% four years, 28% and 36% at 5 years respectively. Survival between both groups were similar (P = 0.29). Conclusion: We evaluated if transplantation offered a survival benefit in patients with early IHCCC. Results showed no difference in survival between resection and transplantation. This study questions the utility of transplantation for stage I and II IHCC in the era of organ shortage.

LO-H.03 A NATIONWIDE ASSESSMENT OF OUTCOMES AFTER BILE DUCT RECONSTRUCTION M. F. Eskander, L. A. Bliss, O. K. Yousafzai, S. W. De Geus, S. Ng, M. P. Callery, K. Khwaja, J. F. Tseng Beth Israel Deaconess Medical Center, Boston, MA Background: Bile duct reconstruction (BDR) is used to manage benign and malignant neoplasms, choledochal cysts and congenital anomalies, trauma and iatrogenic bile duct injuries, and other non-malignant diseases. We compared BDR outcomes overall and by indication. Methods: Retrospective analysis of Nationwide Inpatient Sample discharges (2004–2011) including ICD-9 codes for BDR. All statistical testing performed using survey weighting. Univariate analysis of patient, hospital, admission characteristics, and outcomes by indication using chi square testing. Multivariate modeling for inpatient complications and inpatient death by logistic regression. Results: Identified 67,160 weighted patient discharges in which BDR was coded: 2.5% for congenital anomaly, 37.4% for malignant neoplasm, 2.3% for benign neoplasm, 9.9% for biliary injury, and 48% for other non-malignant disease. 68.4% of BDR discharges for neoplasm were elective vs. HPB 2015, 17 (Suppl. 1), 1–81

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

50

Abstracts

60.8% for congenital anomaly, 46.1% for other nonmalignant and 37.8% for biliary injury (p < 0.0001.) 79.8% of neoplasm discharges were from teaching hospitals vs. 62.3% for other non-malignant disease, 65.2% for biliary injury and 66.1% for congenital anomaly (p < 0.0001.) 33.3% of total BDR discharges involved at least one complication and 84.8% were discharges to home. Median length of stay was 9 days (IQR 6, 15) and median cost was $22,230 (IQR 14,399, 38,358.) Significant multivariate predictors of inpatient death include indication of biliary injury or malignancy (figure), and predictors of any complication include public insurance and non-elective admission. Conclusion: This is the first national description of BDR using a large database. In this diverse sampling, both procedure indication and patient characteristics influence morbidity and mortality.

included surgery (n = 256, 57.5%), antibiotics alone (n = 117, 26.3%), and cholecystostomy tube (n = 72, 16.2%). For all patients, length of stay (p < 0.001), disposition to home (p < 0.001), and morbidity (p = 0.003) were related to increasing TG13 grade. For surgical patients, worsened outcomes with increasing TG13 grade were seen for conversion to open (p = 0.001), OR duration (p < 0.001), length of stay (p = 0.009), disposition to home (p < 0.001), and readmission (p = 0.037). On multivariate analysis, TG13 grade was an independent predictor of increasing length of stay (p = 0.009) and conversion to open surgery (grade 2 = OR 7.63 (2.25–25.90), grade 3 = OR 24.2 (5.0–116.37)). Conclusion: The TG13 criteria for grading acute cholecystitis accurately stratify patient outcomes in a US population. Wide adoption of TG13 can better inform patients, hospital systems, and payers of the expected outcomes of acute cholecystitis.

LO-H.05 TARGETING DEFINITIVE MANAGEMENT IN PATIENTS WITH ACUTE GALLSTONE PANCREATITIS AND CHOLEDOCHOLITHIASIS L. C. Ewan, H. Jenkins, D. A. Subar Department Of HPB Surgery, Blackburn, LANCASHIRE

LO-H.04 PREDICTING LENGTH OF STAY AND CONVERSION TO OPEN SURGERY FOR ACUTE CHOLECYSTITIS: VALIDATING THE 2013 TOKYO GUIDELINES IN A US POPULATION G. Wright1,2, M. T. Hefty1,2, K. Stilwell2, J. Johnson2, M. H. Chung1,2,3 1 GRMEP/Michigan State University General Surgery Residency Program, Grand Rapids, MI; 2Michigan State University College Of Human Medicine, Grand Rapids, MI; 3Spectrum Health Medical Group, Grand Rapids, MI Introduction: Predicting expected patient outcomes based on disease severity is becoming increasingly important in the US healthcare system. The 2013 Tokyo Guidelines (TG13) for the diagnosis and severity of acute cholecystitis were put forward by a consensus panel. Methods: A retrospective review of patients presenting with acute cholecystitis to a single center from 2009–2013 was performed. The diagnosis and severity of cholecystitis were assigned according to the TG13. The primary outcome measures were length of stay and conversion to open surgery. Regression models were constructed for risk-adjusted analysis. Results: A total of 445 patients were eligible for study. Patients were divided as follows: 137 (30.8%) grade 1, 191 (42.9%) grade 2, and 117 (26.3%) grade 3. Primary treatment © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

Background: Appropriate management of common bile duct stones in patients with gallstone pancreatitis often varies from recommended guidelines. Aim: To determine if patients with gallstone pancreatitis with common bile duct stones (CBDS) were appropriately investigated and managed according to guidelines. Methods: This retrospective study identified 165 patients from 2009–2013 with a first episode of gallstone pancreatitis. Cumulative scoring (0–5) based on age >55 yrs, CBD >7 mm, bilirubin >22 mmol/L, ALP >200, ALT >53 was used to predict probability of CBDS. Investigation with either MRCP only, ERCP only, or MRCP then ERCP was determined for each group. The presence of CBDS was used determine whether the primary investigation was appropriate. Results: In the high CBDS probability group (Score 5,n = 18) 83% demonstrated a stone. 72% of this group had ERCP directly and 28% had MRCP then ERCP. The mean delay between MRCP and ERCP was 6.8d. For those with 4 risk factors for CBDS (n = 44) 45% were found to have a CBDS. The majority had an MRCP before ERCP (45%), 20% went straight to ERCP and 16% had MRCP only. Average delay to ERCP from MRCP was 5.3d (inpatients) and 99d (outpatients). For those at low risk of CBDS (Score 0–2,n = 55), 9% had CBDS. Most underwent MRCP only (34.5%) however 24% had ERCP of which 85% negative for CBDS. Conclusion: Acute gallstone pancreatitis cases predicted as high risk for CBDS should undergo ERCP directly. For those at moderate risk outpatient delays for MRCP should be avoided. ERCP should be avoided in low risk groups unless MRCP demonstrates a CBDS.

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

LO-H.06 BILE DUCT RESECTION IN THE TREATMENT OF HEPATOBILIARY AND GALLBLADDER MALIGNANCY: EFFECT OF ASSOCIATED PROCEDURES ON OUTCOMES P. Shen, N. Fino, E. Levine, C. Clark Wake Forest School Of Medicine, Winston Salem, NC Introduction: Resection of the bile duct is required for the treatment of cholangiocarcinoma and sometimes indicated when resecting gallbladder and hepatic tumors. Methods: The American College of Surgeons National Surgical Quality Improvement Program Participant Use File was used to analyze surgical outcomes in a database of patients with hepatobiliary and gallbladder malignancies undergoing bile duct resection with or without hepatic or vascular resection (n = 787). Patients were divided into three groups based on type of procedure performed: 1)Bile duct resection only (n = 289); 2)Bile duct resection with hepatic resection (n = 454); and 3)Bile duct resection with hepatic resection and vascular resection (n = 44). Postoperative complications were compared between groups and regression-adjusted risk factors were analyzed to produce observed and expected (O/E) morbidity and mortality rates and indices. Results: Performing additional procedures significantly increased rates of organ space surgical site infection (p < 0.0001), being on ventilator >48 hours (p = 0.0388), acute renal failure (p = 0.0055), sepsis (p = 0.0292), septic shock (p = 0.0208) and overall risk of having at least one complication (p < 0.0001). Thirty-day mortality rates for Groups 1, 2 and 3 were 6.23%, 8.15% and 18.18%, respectively (p = 0.0242). Risk-adjusted morbidity and mortality rates also increased when Group 1 (O/E = 1.27 and 1.64) was compared to Group 2 (O/E = 1.61 and 2.31) and Group 3 (O/E = 1.88 and 8.00). Conclusion: Hepatic and vascular resection significantly increase morbidity and mortality when performed with bile duct resection for malignancy. Patients undergoing all three procedures experience an 8-fold increase in mortality risk with a morbidity risk that is 2-fold higher.

LO-H.07 CLINICAL AND PATHOLOGICAL FEATURES OF INTRADUCTAL PAPILLARY NEOPLASM OF THE BILIARY TRACT AND GALLBLADDER S. Bennett1, C. Marginean2, M. Paquin-Gobeil1, J. Wasserman2, J. Weaver1, R. Mimeault1, F. K. Balaa1, G. Martel1 1 University Of Ottawa, Dept Of Surgery, Ottawa, ON; 2 University Of Ottawa, Dept Of Pathology And Laboratory Medicine, Ottawa, ON Background: Intraductal papillary neoplasms of the bile duct (IPNB) and intracholecystic papillary neoplasms (ICPN) are rare tumors of biliary epithelium, characterized by papillary growth within lumen that can be associated with invasive carcinoma. Their natural history remains poorly understood. This study examines clinicopathological features and outcomes. HPB 2015, 17 (Suppl. 1), 1–81

51

Methods: Patients who underwent surgery for IPNB/ICPN between 2009 and 2014 were identified. Descriptive statistics were generated. Results: Of 23 patients found to harbor IPNB/ICPN, 43% were male and average age was 68. Most common presentations were jaundice (43%), abdominal pain (29%), and incidental ultrasound finding (14%). Preoperative ERCP with brushing/biopsy showed at least cytologic atypia in 8/10 cases. Tumor locations were: 5 intrahepatic, 3 hilar, 8 extrahepatic bile duct, and 7 gallbladder. Mean tumor size was 3.8 cm, 25% had positive lymph nodes, 47% had lymphovascular invasion, and 37% had perineural invasion. The R0 resection rate was 83%. The average number of lymph nodes sampled was 4.8. Epithelial subtypes included pancreatobiliary (52%) and intestinal (48%), and 87% demonstrated invasive carcinoma, either tubular type or mucinous. Median follow-up was 25 months. The 3-year overall and disease-free survivals were 70% and 61%, respectively. Of the 6 recurrences, 4 occurred in patients who had extrahepatic bile duct cancers and 2 in gallbladder cancers. Conclusion: IPNB/ICPN are rare tumors that spread along the entire biliary epithelium, including the gallbladder. At pathology, the majority of patients demonstrate invasive carcinoma, thus warranting radical resection. Oncologic prognosis may be superior to that of other biliary tract cancers.

LO-H.09 INTRAOPERATIVE NEAR-INFRARED CHOLANGIOGRAPHY: OPTIMIZATION OF TIMING AND DOSE A. Zarrinpar, E. P. Dutson, C. Mobley, R. W. Busuttil, C. E. Lewis, A. Tillou, A. Cheaito, O. J. Hines, V. G. Agopian, D. T. Hiyama Department Of Surgery, David Geffen School Of Medicine, UCLA, Los Angeles, CA Introduction: Intraoperative cholangiography is the gold standard for clear delineation of biliary anatomy. However, logistical difficulties lead to its low utilization. Near-infrared fluorescence cholangiography (NIRFC) with indocyanine green (ICG) has been developed for real-time, intraoperative biliary imaging. While several studies have shown its feasibility, dosing and timing for its practical use have not been systematically optimized. Objective: We undertook a prospective observational study with varying doses and elapsed times from injection of ICG to visualization. Image quality of NIRFC and its utility to the operating surgeon were assessed. Methods: Adult patients undergoing laparoscopic biliary and hepatic operations were enrolled. A single intravenous dose of ICG (0.02–0.25 mg/kg) was administered at various times (15–180 mins) prior to planned visualization. The porta hepatis was examined using a dedicated laparoscopic system. Each operating surgeon evaluated the intraoperative recognition of biliary structures using a qualitative scoring system (1-poor to 5-excellent). Quantitation studies were also performed on the images obtained during the operation. Results: Thirty-four patients were enrolled. Visualization scores of the extrahepatic biliary tract improved with increasing doses of ICG up to 0.08 mg/kg. The score also improved with increased time up to 45 min after ICG administration. Similarly the CBD-to-liver intensity ratio increased with © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

52

Abstracts

both dose and time. These results suggest that a dose of 0.08 mg/kg administered 45 minutes prior to visualization is optimal for visualization, with diminishing improvements with increased dose and time. Conclusion: NIRFC is safe, practical, and effective in delineating extrahepatic biliary anatomy during laparoscopic biliary and hepatic operations.

LO-I.02 LONG-TERM OUTCOME OF PATIENTS UNDERGOING LIVER TRANSPLANTATION FOR MIXED HEPATOCELLULAR CARCINOMA AND CHOLANGIOCARCINOMA: AN ANALYSIS OF THE UNOS DATABASE V. Vilchez1, L. Pena2, M. Shah1, M. F. Daily1, C. Tzeng1, D. Davenport1, R. Gedaly1, E. Maynard1 1 Department Of Surgery – University Of Kentucky, Lexington, KENTUCKY; 2Department Of Internal Medicine-Gastroenterology – University Of Kentucky, Lexington, KENTUCKY Objective: To compare long-term outcomes in patients undergoing liver transplantation (LT) for mixed hepatocellular carcinoma/cholangiocarcinoma (HCC-CC) versus those with hepatocellular carcinoma (HCC) or cholangiocarcinoma (CC). Methods: A retrospective analysis of patients undergoing LT for HCC-CC was performed using the United Network for Organ Sharing (UNOS) database from 1994–2013. Overall and disease-free survival (OS, DFS) in patients with HCC-CC, HCC, and CC were compared. Results: Of the 123,167 patients who underwent LT, 4,049 patients had a primary malignancy (94 HCC-CC; 3,515 HCC; 440 CC). Within the HCC-CC cohort 47(50%) had diagnosis of HCV, compared to 1260 (35%) with HCC and 11(2%) with CC. The mean age of the patients with HCC-CC was 57 ± 10 years and 77% were male. MELD at time of listing did not differ among the three groups. Forty-six percent of the patients with HCC-CC recurred. OS at 1, 3 and 5-years for HCC-CC (82%, 47%, 40%) was similar to CC (79%, 58%, 47%) but significantly worse compared to HCC (86%, 72%, and 62% p = 0.002). Similarly, DFS at 1, 3, and 5 years, for HCC-CC (78%, 45%, 38%) was similar to CC (75%, 55%, 44%) but significantly worse than HCC (82%, 68%, 54%, p = 0.005). Conclusion: LT for mixed HCC-CC have inferior OS and DFS compared to those with HCC, suggesting that HCC-CC outcomes more closely follow the CC phenotype. Attempts © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

should be made to identify HCC-CC patients prior to transplant and if transplanted undergo close surveillance and consideration for immunosuppression modification and/or adjuvant therapy.

LO-I.03 LIVER TRANSPLANTATION FOR HEPATOCELLULAR CANCER IN HIV POSITIVE PATIENTS N. Nissen1, R. Rogers2, B. Barin4, P. Stock2 Cedars-Sinai Medical Center, Los Angeles, CA; 2 University Of California, San Francisco, San Francisco, CA; 4EMMES Corporation, Rockville, MD

1

Introduction: Liver transplantation (LT) is an excellent treatment option for hepatocellular carcinoma (HCC) in many patients, but whether this applies to the HIV+ patient is unknown. Aim: To evaluate the outcomes of LT for HCC in patients with HIV co-infection by analyzing results of a US multicenter trial of solid organ transplantation in HIV+ patients. In this trial, 125 patients underwent LT between 2001 and 2007 at 11 US centers. Methods: Database review. Patients: Forty-five out of 125 patients (30%) undergoing LT in this trial had HCC. These patients averaged 50 years of age and the median native MELD at transplant was 15 (IQR 11–15). All patients were within Milan criteria at transplant and only 1 patient had been previously downstaged. Hepatitis C (HCV) was present in 30 of 45 patients. Results: One and 3 year graft survivals for the HCC+/HIV+ patients were 82% and 67%, compared to 75% and 58% in the HCC-/HIV+ patient group from the same trial (p ns). Most grafts were lost due to HCV. At a median followup of 4.5 years, 4 patients (11%) developed HCC recurrence, which occurred at a mean of 14 months after LT. Of the recurrences one patient survived >24 months. Conclusion: LT is an effective treatment for HCC in HIV+ patients, with survival rates similar to that of HIV+ patients without HCC. Patient and graft survival are primarily determined by HCV status. HIV+ patients do not have excessive risk of tumor progression after LT. HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

LO-I.04 PREDICTIVE FACTORS FOR EXTRAHEPATIC RECURRENCE OF HEPATOCELLULAR CARCINOMA FOLLOWING ORTHOTOPIC LIVER TRANSPLANTATION A. Andreou, J. Pratschke, D. Seehofer Department Of General, Visceral And Transplant Surgery, Charite Campus Virchow Klinikum, Berlin, DEUTSCHLAND Background: Recurrence of hepatocellular carcinoma (HCC) in patients treated with orthotopic liver transplantation (oLTX) is associated with diminished survival. Particularly extrahepatic localization of HCC recurrence contributes to poor prognosis. Patients and Methods: Clinicopathological data of patients who underwent oLTX for HCC between 1989 and 2010 in a high-volume transplant center were retrospectively evaluated and predictors of extrahepatic recurrence were identified. Results: Three hundred and sixty-seven patients underwent oLTX for HCC. After a median follow-up time of 77 months, 93 patients (25%) were diagnosed with a recurrence. Median time to recurrence was 18.9 months. Recurrence was located exclusively in the liver in 19 cases (20%) and 74 patients (80%) had extrahepatic recurrence. Factors associated with extrahepatic recurrence in multivariate analysis included HCC beyond the Milan criteria (P < .0001) and the presence of major vascular tumor invasion (MVI) (P = .035). In patients with HCC beyond the Milan criteria who developed a recurrence (n = 73), MVI was the only positive predictor of extrahepatic recurrence in multivariate analysis (P = .0001). In patients with HCC within the Milan criteria who recurred after oLTX, DNA-index >1.5 (P = .04) was the only predictive factor for extrahepatic recurrence. Conclusions: Advanced HCC beyond the Milan criteria and the presence of MVI are associated with an increased risk for extrahepatic recurrence and are currently considered as contraindications to oLTX. In patients with HCC within the Milan criteria, the DNA-index represents a valuable prognostic marker for the development of extrahepatic recurrence and may support the selection of patients for intensive postoperative tumor surveillance.

SATURDAY, MARCH 14, 2015, 4:00PM–5:30PM ORAL POSTER II (PANCREAS, TRANSPLANT) OP-II.01 PANCREATIC DEBRIDEMENT FOR NECROTIZING PANCREATITIS: NATIONAL OUTCOMES AND PREDICTORS OF MORTALITY S. W. Ross, E. M. Hanna, R. C. Kirks, R. Seshadri, J. B. Martinie, R. Z. Swan, D. A. Iannitti Carolinas Medical Center, Division Of HPB Surgery, Charlotte, NC Background: Necrotizing pancreatitis often requires pancreatic debridement, which can result in a high rate of HPB 2015, 17 (Suppl. 1), 1–81

53

morbidity and mortality. Risk factors that predispose patients to post-operative complications or death are not well-defined. Using a national surgical database, we sought to identify pre-operative predictors of mortality after debridement. Methods: The American College of Surgeon’s NSQIP database was queried from 2005–2011 by CPT code for pancreatic debridement. Risk factors for complications and death at 30 days were examined using univariate tests and predictors of mortality were identified using step-wise logistic regression. Results: 1,162 patients underwent pancreatic debridement. On average patients were middle aged(54.9 ± 14.2 years), male(70.0%), and obese(30.6 ± 8.0 kg/m2). The most common comorbidities included diabetes(33.2%), smoking(22.3%), COPD(6.5%), steroid use(4.0%) and cardiac history (4.0%). Prior to surgery, 21.2% had ascites, 7.6% had acute renal failure, 14.7% had recent >10% body mass loss, and 63.2% had pre-operative sepsis. Emergent surgery occurred in 29.3%. Wound, general, and major complications occurred in 7.2%, 56.6% and 31.3% respectively. Mean length of stay was 34.1 ± 31.0 days. 30 day mortality was 7.7%, and was higher in patients with COPD, cardiac history, ascites, acute renal failure, steroid use, sepsis, older age, higher BMI, and emergent operations(p < 0.05). The table summarizes independent predictors of mortality identified using multivariate analysis. Conclusion: Nationally, rates of adverse outcomes following pancreatic debridement are high. Patients who are older, with higher BMI, are on dialysis, have COPD, with poor functional status, hypoalbuminemia, azotemia or hyperbilirubinemia have increased odds of mortality following debridement for necrotizing pancreatitis.

OP-II.02 EXPERIENCE WITH PANCREAS-SPARING DUODENECTOMY FOR FAMILIAL ADENOMATOUS POLYPOSIS M. Dong, N. Ali, S. Reddy, C. O’Rourke, G. Morris-Stiff, R. M. Walsh Department Of HPB Surgery, Cleveland, OHIO Introduction: Duodenal adenomas are a common finding in patients with familial adenomatous polyposis (FAP) and individuals with Spigelman stage IV adenomas are at high risk of developing duodenal carcinoma. These patients are traditionally treated by pancreatoduodenectomy (PD) though an alternate approach is pancreas-sparing duodenectomy © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

54

Abstracts

(PSD). We report a 22-year experience with PSD for the treatment of duodenal polyps in FAP. Methods: A retrospective review was performed of a prospectively maintained database containing all patients undergoing PSD from 1992 to 2013. Phone interviews were conducted to confirm current status of patient at follow-up. Results: Fifty-four patients underwent PSD during the study period, all for Spigelman stage IV polyps. An unsuspected invasive cancer was found in one patient on final pathology. The mean operative time was 305 ± 70 minutes with a mean blood loss of 300 ± 170 mL. There was one peri-operative mortality, unrelated to the operative procedure. Thirteen patients (24%) had an immediate postoperative complication including eight (15%) biliary/ pancreatic leaks, and 1 (2%) enteric anastomotic leak. Pancreatitis was observed in 4 (10%). 42(78%) of patients were available for follow-up. Recurrent polyps were found in 16(34%). Of these, only 3(19%) patients required operative intervention, two proximal jejunal resections and one PD for development of a polyp at the ampullary anastomosis. Conclusion: Our experience with PSD reinforces its value as a definitive prophylactic procedure for duodenal polyposis in FAP and allows for full preservation of pancreatic function.

OP-II.03 THE EFFECT OF EARLY POST-OPERATIVE NON-STEROIDAL ANTI-INFLAMMATORY DRUGS ON OUTCOMES FOLLOWING PANCREATICODUODENECTOMY R. Behman1,2, P. Karanicolas1,2, M. Lemke1,2, S. Hanna1,2, C. Law1,2, N. Coburn1,2, J. Hallet1,2 1 University Of Toronto, Toronto, ONTARIO; 2Odette Cancer Centre, Toronto, ONTARIO Introduction: Non-steroidal anti-inflammatory drugs (NSAIDs) are used commonly for post-operative analgesia, but can potentially impair healing. Their effect on pancreaticoduodenectomy (PD) outcomes is unknown. We sought to examine the impact of early post-operative NSAIDs on pancreatic fistula (PF) after PD. Methods: We reviewed our prospective pancreatectomy database supplemented by medication administration records, including all PDs from 2002 to 2012. Primary outcome was occurrence of clinically significant (Grade B-C) PF. Secondary outcomes included major morbidity (Clavien grade III-V) and 90-day mortality. Patients were compared based on early post-operative NSAIDs use (first 3 days following surgery) using univariate and multivariate analyses. Sub-group analyses were conducted based on NSAIDs type (ketorolac and COX2-inhibitor). Results: We identified 251 PDs, of whom 127 (50.6%) patients received NSAIDs postoperatively (15.1% ketorolac, 32.7%, COX2-inhibitor, 2.8% both). Use of any NSAIDs was associated with a non-significant increase in PF (16.5% Vs 11.3%%; p = 0.23), and no difference in major morbidity and mortality. Use of ketorolac was not associated with an increase in PF (8.7% Vs. 15.1%; p = 0.256). COX2inhibitors were associated with increased PF (20.2% Vs. 10.5%; p = 0.033), but no difference in major morbidity or mortality. After adjusting for age, Charlson comorbidity © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

score, diagnosis, and estimated blood loss, use of COX2inhibitors was independently associated with PF (odds ratio 2.32; p = 0.026). Conclusions: COX2-inhibitors are associated with PF in the early postoperative period. While ketorolac appears safe in this setting, caution is warranted with the use of COX2inhibitors.

OP-II.04 NEOADJUVANT THERAPY WITH ANATOMICAL BORDERLINE PANCREATIC DUCTAL ADENOCARCINOMA. DOES IT MAKE DIFFERENCE? A. M. Zaki1, N. Rezaee1, J. He1, J. L. Cameron1, K. Hirose1, T. M. Pawlik1, N. Ahuja1, M. A. Makary1, H. E. Horeya2, M. J. Weiss1, C. L. Wolfgang1 1 Johns Hopkins University School Of Medicine, Baltimore, MARYLAND; 2Mansoura University School Of Medicine, Mansoura, DAKAHLIA Background: The benefit of neoadjuvant therapy over a surgery-first approach in patients with borderline pancreatic ductal adenocarcinoma (PDAC) has not been well defined. Aim: To compare postoperative outcomes of patients with borderline PDAC who underwent pancreatectomy after neoadjuvant treatment with those of patients who underwent upfront surgery. Methods: Between 2008 and 2014, 231 patients were identified as anatomical borderline PDAC. 117 of 231 (50.6%) patients received neoadjuvant therapy and 114 (49.4%) patients had a surgery-first approach. Univariate, multivariate and survival analyses were performed. Results: Compared to surgery first group, neoadjuvant group was associated with smaller tumor size in the pathological specimen (P < .001), lower incidence of metastatic lymph nodes (39% vs. 80%; P < .001), less perineural invasion (61% vs. 97%: P < .001), less micro-vascular invasion (32% vs. 68%: P < .001), less vascular resection rate (31% vs. 57%; P < .001) and a lower rate of positive resection margin (32% vs. 44%; P < .055). Univariate analysis identified nodal status, lymph node ratio and tumor size as predictors for survival. Multivariate analysis identified only lymph node ratio (P = .005) as independent predictor of patient survival. Postoperative mortality and morbidity rates were similar in the 2 groups. However, survival analysis starting from date of first dose of neoadjuvant therapy versus date of upfront surgery showed better median overall survival in favor of the neoadjuvant group (44 months vs. 20 months; P = .011). Conclusion: Neoadjuvant treatment for borderline PDAC is associated with better pathological outcomes and overall survival. Lymph node ratio can provide significant prognostic information after pancreatectomy for patients with borderline PDAC.

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

OP-II.05 PANCREATODUODENECTOMY PROVIDES EFFECTIVE LONG-TERM PAIN RELIEF FOR CHRONIC PANCREATITIS IN SELECT PATIENTS AT GREATER THAN 15 YEARS FOLLOW-UP K. P. Croome, D. M. Nagorney, M. Tee, M. J. Truty, K. Reid-Lombardo, F. G. Que, M. L. Kendrick, M. B. Farnell Mayo Clinic, Rochester, MN Background: We have employed pancreaticoduodenectomy (PD) for selected patients with small duct, headdominant chronic pancreatitis (CP) with intractable pain. Information examining very long term outcomes in patients undergoing PD for CP is lacking. Patients and Methods: All patients who underwent PD for CP from 1976 to 2013 were reviewed. Surviving patients were contacted for a follow-up questionnaire and SF-12 Quality of Life Survey. Results: A total of 166 patients were identified (Cohort 1: 1976–1999(N = 105) and Cohort 2: 2000–2013(N = 61)). Median time from presentation until surgery was significantly longer in Cohort 2(2.09 years) compared to Cohort 1(1.13 years)(p = 0.017). A higher proportion of patients in Cohort 2(98%) had intractable pain prior to surgery than in Cohort 1 (82%)(p = 0.002). Prior to PD a higher proportion of patients in Cohort 2 had undergone endoscopic stenting, 67% vs 10%(p < 0.001) and/or celiac plexus block 15% and 5%(p = 0.026). Median follow-up for all survey respondents was 15 years. On the SF-12, mean physical component score(PCS) was 43.8 ± 11.8 and mental component score(MCS) was 54.3 ± 7.9. Patients were significantly lower on the PCS(p < 0.001) and significantly better on the MCS(p = 0.001) than the general US population. Mean pain score out of 10 was significantly lower after surgery 1.6 ± 2.6 than before surgery 7.9 ± 3.5(p < 0.001). Diabetes developed in 28% of patients who were not diabetic prior to surgery. Conclusion: Although practice has changed so that patients have a longer time from presentation until surgery as less invasive techniques are attempted, PD appears to provide effective long-term pain relief and acceptable quality of life in appropriately selected patients with chronic pancreatitis and intractable pain.

OP-II.06 GETTING THE FULL PICTURE: SIGNIFICANCE OF INDETERMINATE LUNG NODULES IN PANCREATIC ADENOCARCINOMA K. T. Hemingway1, E. Halpern2, P. V. Pandharipande3, A. L. Warshaw1, C. Fernandez Del Castillo1, K. D. Lillemoe1, C. R. Ferrone1 1 Department Of Surgery-Massachusetts General Hospital, Boston, MA; 2Department Of Biostatistics-Massachusetts General Hospital, Boston, MA; 3Department Of Radiology-Massachusetts General Hospital, Boston, MA Objectives: Patients presenting with resectable pancreatic adenocarcinoma (PDAC) often have indeterminate lung nodules on pre-operative chest CT scans. The aim of this study was to determine whether nodules on chest CT are predictive of lung first metastases (LFM). HPB 2015, 17 (Suppl. 1), 1–81

55

Methods and Materials: Retrospective clinicopathologic data was recorded for patients undergoing resection of their PDAC between 1993–2012 who also underwent a preoperative chest CT scan. Time to metastasis and overall survival (OS) were calculated. Results: Of 792 patients who underwent surgical resection of their PDAC, 171 underwent a pre-operative chest CT scan. Median age was 67.7 years and 54.7% were female. Median PDAC size was 30 mm. Indeterminate nodules were identified in 111 patients. Indeterminate lung nodules of 0–3 mm in 35 patients, 3–6 mm in 44 patients, 6–10 mm in 12 patients and >10 mm in 14 patients. Lung-first metastasis (LFM) developed in 20.7% of patients with nodules vs. 10.8% of patients with no nodules (p = 0.142). Patients with LFM had the longest median OS compared to patients developing lung + liver (6.3%) or lung + local (5.4%) metastases (32.7 mo vs 22.3 mo vs. 18.6 mo, respectively). The risk of lung-first metastases did not increase with increasing size, number, or location of indeterminate nodules, smoking history or pre-operative Ca19-9. Of the 28 patients with LFM, 6 underwent surgical removal of one or more nodules. Conclusions: Neither the location, number, or size of indeterminate lung nodules on pre operative chest CT scans are predictive of developing lung metastases as the first site of metastatic disease.

OP-II.07 READABILITY AND ACCURACY OF ONLINE PATIENT MATERIALS FOR PANCREATIC CANCER BY TREATMENT MODALITY AND WEBSITE AFFILIATION A. Storino1, M. Castillo-Angeles1, A. A. Watkins1, C. Vargas1, J. Mancias1, A. Bullock1, A. Demirjian2, A. J. Moser1, T. S. Kent1 1 Beth Israel Deaconess Medical Center – Harvard Medical School, Boston, MA; 2University Of California, Irvine, Orange, CA Introduction: Patients search online health information frequently but there is little quality control. Patient/family understanding of presented information may depend, partly, upon readability and accuracy. This study was undertaken to evaluate reading level and accuracy among commonly searched websites about pancreatic cancer treatment options. Methods: An online search on 5 pancreatic cancer treatment modalities was conducted. Readability was measured by 9 standardized tests and accuracy was assessed by an expert panel. Readability and accuracy were compared by treatment modality and website affiliation by Kruskal-Wallis test. Results: Significant differences existed by treatment modality for both readability and accuracy (Table 1), with surgeryrelated websites having the lowest readability level. Alternative therapy-related websites had the lowest accuracy. Readability varied by affiliation, with lower readability for non-profits than media-owned (p = 0.00001) and academic center sites (p = 0.0001). Privately-owned websites had lower readability than media sites (p = 0.009). Accuracy was highest for government websites; government, academic, and non-profit sites were more accurate than privately-owned or media-owned websites (p = 0.0001). There was no association between accuracy and readability level. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

56

Abstracts

Conclusions: Although variation existed in both readability and accuracy, improvement is needed throughout. The readability level for all treatment modalities is higher than recommended, which may negatively impact patient/family understanding of treatment options. Accuracy was reasonable, except for alternative therapy websites. Website affiliation impacted both readability and accuracy. Privately-owned and media sites had lower accuracy. In accordance with patient-centered care, improvement is needed in the quality of online resources in order to empower patients in the shared-decision making setting.

OP-II.08 ENGLISH AND SPANISH LANGUAGE READABILITY OF ONLINE PATIENT RESOURCES FOR PANCREATIC CANCER M. Castillo-Angeles, A. Storino, A. A. Watkins, C. R. Vargas, J. F. Tseng, M. P. Callery, A. J. Moser, T. S. Kent Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA Introduction: 50% of cancer patients search the Internet for health information. The NIH and AMA recommend that patient-oriented health information should be written at a sixth-grade reading level. We evaluated commonly-searched English and Spanish-language pancreatic cancer websites in light of these recommendations. Methods: The top ten websites for “pancreas cancer” in English and Spanish were identified in a patient-simulated manner using the most-visited internet search engine. 122 English and 76 Spanish articles were assessed using ten readability analyses. Spanish-language websites native to 3 representative Spanish-language countries with AHPBA chapters were identified. Mean readability scores were determined for each article and website. T-tests were used to compare reading levels by website, language, and country. Results: Overall mean reading level across US websites was 12.4 ± 1.42 (English) and 10.5 ± 0.95 (Spanish). Mean readability by website ranged from 9.5 to 15.2 in English, and from 8.6 to 12.3 in US Spanish. Argentine, Chilean, and Mexican websites had mean reading levels 9.7, 10.9, and 11 respectively. Spanish websites in each country were significantly easier to read than those in English (Figure 1). Conclusion: Commonly searched online pancreatic cancer resources in English and Spanish exceed the recommended sixth grade reading level. Spanish-language websites, US-based or not, were significantly easier to read than English-language sites, though still above the recommended 6th grade level. With the growing focus on patient-centered care, attention should be paid to ensuring availability of © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

information at a level understandable by the general public. More easily understandable materials may facilitate patient participation in shared decision-making.

OP-II.09 PLASMA CANCER ANTIGEN 19-9 (CA19-9) LEVELS DIFFERENTIATE PATIENTS WITH MALIGNANT INTRADUCTAL PAPILLARY MUCINOUS NEOPLASM (IPMN) FROM THOSE WITH IPMN ALONE T. K. Nguyen, D. Joyce, G. A. Falk, S. Chalikonda, G. Morris-Stiff, R. M. Walsh Cleveland Clinic Foundation, Cleveland, OH Introduction: Malignant transformation is a recognized complication of IPMNs, in particular when the disease affects the main pancreatic duct. However they are often only recognized during histopathological examination of resection specimens. Plasma CA19-9 has potential use in the diagnostic work-up of pancreatic adenocarcinoma but has not been well investigated in IPMN. The aim of this study was to evaluate the role of CA19-9 in differentiating between malignant IPMN (invasive [IPMN Ca] and high grade dysplasia [HGD]) IPMN from those with IPMN and low/ moderate (LGD & MGD) dysplasia. Methods: The institutional pancreatic cyst database was interrogated to identify all patients with a histopathological diagnosis of IPMN. Patients were sub-divided into three categories based on the degree of neoplastic change: IPMN carcinoma; IPMN HGD; and IPMN LGD&MGD. Ca19-9 levels were assessed in relation to the 3 categories. The IPMN Ca and IPMN HGD were then co-assessed. Results: During the period January 2000 to December 2013, 158 patients underwent resection, with final histopathology indicating: IPMN [n = 45] carcinoma, IPMN HGD [n = 24]; and IPMN LGD/MGD [n = 89]. There was no difference in comparing incidental versus pre-operatively diagnosed cancers or when comparing main duct and side-branch IPMNs. Conclusions: Plasma CA19-9 is independently valuable in the evaluation of HGD and carcinoma arising in the setting of IPMN. [Correction added on 24 February 2015, after online publication. Author’s initial changed from R. Walsh to R.M. Walsh] HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

57

OP-II.10 PANCREATIC NECROSIS: A SINGLE INSTITUTIONS REVIEW OF PRACTICAL ADHERENCE TO A STEP-UP APPROACH S. Downs-Canner, B. Boone, J. Steve, A. Zureikat, K. K. Lee, H. J. Zeh, M. Hogg University Of Pittsburgh Medical Center, Pittsburgh, PA Introduction: Level 1 evidence demonstrated the “step-up” approach to infected pancreatic necrosis improves outcomes. We examined our institution’s compliance with this approach and its impact on outcomes. Methods: We performed a retrospective review of the index admission of all patients with necrotizing pancreatitis treated in 2012 at a tertiary referral academic medical center. Results: 41 patients (median age 52, 75% male) were treated for necrotizing pancreatitis. 80% were transferred a median of 3 days from presentation. By the revised Atlanta Classification, 23 (56%) had severe acute pancreatitis and 19 (46%) had suspected/confirmed infected necrosis. Of those infected, 18 (95%) required pancreatic drainage procedures; 10 (56%) underwent step-up (Table). Median time from presentation to first pancreatic drainage was 1 day (step-up = 0.5 days versus surgery first = 4.5 days). Five (50%) of step-up went on to surgery a median of 1 day later. Nine (47%) had multi-system organ failure (step-up = 30% versus surgery first = 75%). Median length of stay was 23 days for step-up versus 21 days for surgery first. ICU admission rate was 90% for step-up versus 80% for surgery first. In-hospital mortality was 20% in step-up and 12.5% in surgery first. Overall survival for all necrotizing pancreatitis patients was 78% and 71% of deaths were related to pancreatitis. Pancreatitis related readmission rate was 59% (step-up = 30% versus surgery = 88%). Conclusion: Step-up adherence was 56%; however, half of those went on to surgery. Further analysis is necessary to determine if the subset of patients undergoing surgery first represent deviations from level 1 recommendations or medically appropriate deviations.

HPB 2015, 17 (Suppl. 1), 1–81

OP-II.11 NATIONAL DISPARITIES IN MINIMALLY INVASIVE SURGERY FOR PANCREATIC CANCER E. M. Gabriel1, P. Thirunavukarasu1, K. Attwood2, S. Hochwald1, B. Kuvshinoff1, S. Nurkin1 1 Roswell Park Cancer Institute, Buffalo, NY; 2Roswell Park Cancer Institute, Buffalo, NY Introduction: Social and racial disparities have been related to differences in access to care. This study investigated patterns in minimally invasive surgery (MIS) across different social, racial and geographic populations of patients with pancreatic cancer. Methods: We utilized the National Cancer Database, 2004 to 2011, to identify patients with pancreatic cancer who underwent surgery through either an open, laparoscopic or robotic approach. Multivariate analysis was performed to characterize differences in patient demographics in relation to surgical approach. Results: A total of 11,464 patients were identified. The initial surgical approach included 82.5% open (9,461), 15.8% laparoscopic (1,815) and 1.6% robotic (188). Table 1 shows the results of our analysis. Race was not statistically significant across the different surgical approaches. There was a trend toward increased MIS in patients with private insurance. Academic centers performed more MIS compared to community cancer programs. On multivariate analysis, only national location was shown to be a statistically significant factor associated with increased rates of MIS. Patients in the Middle Atlantic region of the US were most likely to have robotic surgery. Regarding laparoscopic surgery, the Mountain and West South Central states had the lowest rates of laparoscopic procedures, but among the other national regions there were no statistically significant differences. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

58

Abstracts

Conclusions: Minimally invasive approaches for pancreatic cancer comprise 17.5% of surgical procedures. Race and insurance status were not statistically significant factors associated with MIS. Although academic centers performed most of the MIS, specific geographic regions comprised the only statistically significant factor on multivariate analysis.

whereas only 47 (38.5%) were reported in their pathology reports. Thirty patients were found to have pathologic evidence of PV/SMV resection, but not based on billing. Conclusion: Histologic characteristics and margin distances are frequently missing from PAC pathology reports. High rates of R1 resections exist, as well as discrepancies regarding reporting of major vein resections between billing and pathology reports, which holds substantial implications for utilization of pathology reports for determination of PV/SMV resection.

OP-II.13 COMPARATIVE ANALYSIS OF INSTITUTIONAL CLINICAL PATHWAYS FOR MANAGEMENT OF BORDERLINE RESECTABLE PANCREATIC CANCER O. M. Rashid1, J. M. Pimiento1, P. Nguyen1, G. Springett1, P. Hodul1, S. Hoffe1, R. Shridhar1, B. L. Johnson2, K. Illig2, P. A. Armstrong2, W. J. Fulp1, M. P. Malafa1 1 H. Lee Moffitt Cancer Center And Research Institute, Tampa, FL; 2University Of South Florida, Tampa, FL

OP-II.12 WHAT IS THE COMPLETENESS OF PATHOLOGY REPORTS FOR RESECTED PANCREATIC ADENOCARCINOMA SPECIMENS? A POPULATION-BASED ANALYSIS A. El-Sedfy1,2, D. J. Kagedan1,3, E. Shin1, R. Raju1, M. E. Dixon1,4, M. Elmi 1,3, C. Rowsell3, Q. Li1,3, N. Mittmann1,3, N. G. Coburn1,3 1 Sunnybrook Research Institute, Toronto, ONTARIO; 2Saint Barnabas Medical Center, Livingston, NJ; 3Sunnybrook Health Sciences Centre, Toronto, ONTARIO; 4Maimonides Medical Centre, New York, NY Introduction: Following oncologic resection, pathology reports provide critical information for determining prognosis and directing treatment. However, reports are often incomplete. We assessed pathology report completeness from pancreatic adenocarcinoma (PAC) resections on a population level. Methods: All patients who underwent resection for PAC in Ontario between 2005 and 2010 were identified using provincial databases. Pathology reports were evaluated for completeness based on the College of American Pathologists (CAP) 2012 Protocol. Portal (PV) and superior mesenteric vein (SMV) resections identified from physician billing codes were compared to the corresponding pathology report. Results: 475 patients had available pathology reports for review. 421 were pancreaticoduodenectomy and 54 were distal pancreatectomy. CAP pathologic variables and the percentage of reports missing each is as follows: T stage 0.8%; N stage 0.4%; number of lymph nodes examined 3.4%; margin status 2.7%; tumor grade 1.7%; lymphovascular invasion 20.2%; perineural invasion 13.5%; microscopic tumor extension 5.5%. 32.4% of pathology reports (n = 156) revealed R1 resection. 21.4% of reports for specimens with R0 resection (n = 308) were missing margin distance. Based on billing, 122 patients underwent PV/SMV resections, © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

Background: Currently there is no consensus combination therapy clinical pathway (CP) for the management of borderline resectable pancreatic adenocarcinoma (BPA) based on randomized prospective studies, therefore, BPA CP’s vary by institution. We recently reviewed outcomes of our BPA CP and now seek to compare results at other centers to further evaluate optimal therapy. Methods: We performed a literature review of all BPA CP reports, excluding case reports and feasibility studies. BPA diagnostic criteria, pancreatectomy rate, margin status, pathologic response (grade IIa-IV) rate (PR), disease free (DFS), disease specific (DSS), and overall survival (OS) were reviewed and compared to our institutional data. Standard statistical methods were used for statistical comparison. Results: There were three studies which met inclusion criteria (Table). Including the 101 patients treated at our institution, 279 entered 4 CP’s. 175(59%) met NCCN/AHPBA BPA diagnostic criteria, 144(41%) MDACC Type A. The CP regimens are listed (Table). With 95% confidence, resection rate at our institution was 44.2–64.4% versus 27.7–49.3% with CP 1, 16.6–46.5% with CP 2, and 35.8–61.3% with CP 3. With 95% confidence, the PR at our institution was 69.9– 93.7% versus 55.2–63.5% with CP 1, 77.7–89.0% with CP2, and 8.6–10.7% with CP3. DFS, DSS, and OS are reported (Table). Conclusions: A comparative analysis of outcomes demonstrates that many BPA patients benefit from the reported CP’s. However, the differences in reported resection and pathologic response to treatment rates between institutional CP’s warrant further investigation to determine optimal therapy.

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

59

OP-II.14 DOMAIN-BASED ASSESSMENT OF THE LEARNING CURVE FOR NEW SURGICAL TECHNOLOGY: ROBOT-ASSISTED VS. OPEN DISTAL PANCREATECTOMY

OP-II.15 TOTAL PANCREATECTOMY FOR PANCREATIC ADENOCARCINOMA IS ASSOCIATED WITH EQUIVALENT PERI-OPERATIVE MORTALITY AND LONG TERM SURVIVAL AS COMPARED TO PARTIAL PANCREATECTOMY

S. Klompmaker, A. A. Watkins, W. J. Van Der Vliet, S. J. Thoolen, A. Storino, M. Castillo-Angeles, J. F. Tseng, M. P. Callery, T. S. Kent, A. J. Moser Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA

E. H. Baker1, R. M. Seshadri1, M. A. Templin2, R. Z. Swan1, J. B. Martinie1, D. A. Iannitti1 1 Department Of General Surgery, Division Of HPB Surgery, Carolinas Medical Center, Charlotte, NC; 2 Dickson Advanced Analytics Group, Charlotte, NC

Introduction: The Learning Curve for minimally-invasive surgery monitors operating time, conversion rate, and incremental cost as surrogate markers of proficiency and outcome. We expanded this concept to include four aggregate domains for new technology assessment (Table 1) based on Institute of Medicine principles to evaluate overall risk/benefit. The initial Learning Curve for robot-assisted distal pancreatectomy (RADP) was compared to unmatched consecutive open DP (ODP) at an expert center. Methods: Unmatched comparison between 29 RADP and 169 consecutive ODP performed between 2006–2012 prior to implementation of RADP. Cumulative treatment burden at 90 days was assessed. Propensity scoring controlled for selection bias. Results: No differences in age, gender, race, Charlson Comorbidity Index, suspected pathology, tumor location, or size were observed between the RADP and ODP cohorts. Within the efficiency domain, RADP patients had a reduced 90-day total hospital stay (6 vs. 7 days, p = .002), but longer mean operative time as compared to ODP patients (p < .001). Cumulative morbidity and oncological efficacy for malignancy was similar to ODP within the limits of sample size. Safety, as measured by blood loss and laparotomy rate (3.3%), was improved following RADP compared to ODP. Propensity-scored sensitivity analysis did not alter these results. Conclusion: Domain-based evaluation of the initial RADP learning curve was comparable to the established phase of ODP in consecutive patients at an expert center. Operating time and associated costs should be re-evaluated in the context of reduced total hospital stay and increased patient eligibility for the minimally-invasive approach. Prospective validation of these metrics is required.

Introduction: Historically, total pancreatectomy (TP) for pancreatic adenocarcinoma (PAC) has been performed for large or advanced tumors but carries inherent significant postoperative morbidity secondary to endocrine and exocrine insufficiency. Questions regarding postoperative mortality and long term survival remain important when comparing utility of TP to partial pancreatecomy (PP) procedures such as pancreaticoduodenectomy (PD) or distal pancreatectomy (DP) for pancreatic cancer. Methods: Retrospective analysis of the National Cancer Data Base between the years 1998–2011 was performed. 224,335 patients were identified with diagnosis code of PAC; 128,368 patients had relevant data from 1998–2006 and were eligible for 5 year survival analysis. Chi-square, t-test, Kaplan-Meier curves and log rank tests were used for statistical analysis. Results: 16,744 patients underwent surgical resection for PAC during the study period. 1,954 patients underwent TP while 16,111 underwent PP (959 DP and 13,357 PD). Rates of TP increased during the study period (164 in 1998 to 295 in 2006). Mean tumor size was 38 mm for TP and 34.8 mm for PP (p = 0.001). More lymph nodes were examined following TP (11.8 vs 10.8, p < 0.001). 30 day mortality rates following TP were similar to PP (6.05% vs 5.35%, p = 0.20). When compared stage for stage, there were no differences in 30 day mortality between groups. Median overall survival following TP was 14.98 months vs 15.64 months for PP (p = 0.16). No difference was seen in long term overall survival when patients were compared stage for stage. Conclusions: Post-operative mortality and long term overall survival rates are similar between TP and PP procedures.

HPB 2015, 17 (Suppl. 1), 1–81

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

60

Abstracts and lack of at least a 50% decrease in CA19-9 levels (OR 13.2 [2.5–69.1]) were. On sub-analysis, CA19-9 decrease <50% remained predictive for any progression while tumor size predicted distant progression only. A receiver operating characteristic curve showed that tumor size >3 cm was 87% sensitive for progression (AUC 0.785). Conclusion: LAPD patients undergoing neoadjuvant chemotherapy with tumors >3 cm or that exhibit less than 50% reduction in CA 19-9 maybe at higher risk for progression on chemotherapy. Patients with these risk factors may benefit from additional treatment prior to an attempt at resection.

OP-II.17 SURGERY VERSUS NON-OPERATIVE MANAGEMENT OF PANCREATIC ADENOCARCINOMA WITH SMA INVASION P. Thirunavukarasu, E. Gabriel, B. Kuvshinoff, S. Hochwald, S. Nurkin Roswell Park Cancer Institute, Buffalo, NY

OP-II.16 PREDICTORS OF PROGRESSION OF LOCALLY ADVANCED PANCREATIC CANCER ON NEOADJUVANT CHEMOTHERAPY J. B. Rose1, F. G. Rocha1, A. Alseidi1, T. Biehl1, B. Lin2, V. Picozzi2, S. Helton1 1 Virginia Mason Medical Center, Section Of Hepatobiliary Surgery, Seattle, WA; 2Virginia Mason Medical Center, Department Of Hematology/Oncology, Seattle, WA Introduction: Neoadjuvant treatment for patients with locally advanced pancreatic adenocarcinoma (LAPD) is becoming more widely utilized and may be associated with improved survival. However, even with careful patient selection, many patients do not undergo curative resection due to disease progression. The objective of the present study is to identify predictors of disease progression prohibiting resection. Methods: A retrospective review was performed on all patients with LAPD at a high volume tertiary hospital between January 2008 and August 2014 who received extended neoadjuvant gemcitabine/docetaxel chemotherapy. Clinicopathologic predictors of disease progression prohibiting resection were determined by univariate and multivariate logistic regression analysis. Results: Eighty-four patients with LAPD were initiated on neoadjuvant chemotherapy. 16 patients (19%) progressed on treatment by RECIST criteria (9 distant, 7 local). Multivariate logistic regression analysis found that sex, age by quartile, McGill-Brisbane score, clinical stage III, or multivessel involvement were not predictive of progression prohibiting resection. However, tumor size (OR 2.6 [1.3–5.5]) © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

Background: Pancreatic adenocarcinoma (PDAC) invading the Superior Mesenteric Artery (SMA) is traditionally considered unresectable, with surgery offered to select patients. We evaluated outcomes of surgery versus non-operative management. Methods: Data for patients with SMA-invading PDAC without evidence of distant metastases was extracted from the National Cancer Database, 1998–2006. Results: Of 3,445 patients, 227 (6.9%) had surgery, specifically pancreaticoduodenectomies (73.2%), distal pancreatectomies (14.5%) and total pancreatectomies (12.3%). Overall R0 resection rate was 43.7%. Median hospital stay was 9 days, 30-day mortality – 7.0%, and 30-day readmission rate – 8.7%. In patients who underwent surgery plus systemic therapy, median overall survival (OS) was 21 months when margin-negative and 13.6 months when margin-positive (p < 0.001). In contrast, among patients who underwent surgery alone, OS was uniformly poor regardless of margin status (8.3 vs 6.7 months, p = 0.09). Patients with R0 resection plus systemic therapy had significantly better OS compared to patients in whom surgery was not recommended (e.g.: due to comorbidities) or given systemic therapy alone despite recommendation for surgery (Figure 1). Among patients without comorbidities, R0 resection plus systemic therapy had significantly better survival than systemic therapy alone (22.2 months vs. 11.4 months, p < 0.01). The operative mortality of patients with ≤1 comorbidity was lower than with ≥2 comorbidities (6.0% vs. 12.6%, p = 0.05). Conclusion: For appropriate patients with SMA-invading PDAC, a combination of R0 resection with systemic therapy offers the best outcome. Given the higher mortality, these operations should be offered selectively in specialized centers.

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

61

OP-II.19 751 CONSECUTIVE PANCREATICO-DUODENECTOMIES AT A TERTIARY CARE CENTRE: MAKING A CASE FOR STANDARDIZATION AND CENTRALIZATION E. Pai, A. Mitra, S. Patkar, M. Goel, S. V. Shrikhande Tata Memorial Hospital, India, Mumbai, MAHARASHTRA

OP-II.18 EARLY DRAIN REMOVAL IS A BEST PRACTICE IN SELECTED PANCREATIC SURGERY PATIENTS H. A. Pitt1, B. L. ZarZaur2, S. W. Behrman3, E. M. Kilbane2, B. L. Hall4, A. Parmar5, R. S. Riall5 1 Temple University School Of Medicine, Philadelphia, PA; 2 Indiana University School Of Medicine, Indianapolis, IN; 3 University Of Tennessee College Of Medicine, Memphis, TN; 4Washington University School Of Medicine, St. Louis, MO; 5University Of Texas Medical Branch, Galveston, TX Background: The morbidity of pancreatic surgery remains unacceptably high. Recent reports suggest that drain management may influence postoperative complications. While early drain removal may improve outcomes, pancreatic surgeons have been slow to adopt this practice. Therefore, the aim of this analysis was to compare outcomes when drains were removed early or late after pancreatic surgery. Methods: Data were gathered through the American College of Surgeons – National Surgical Quality Improvement Program (ACS-NSQIP) Pancreatectomy Demonstration Project (PDP). Over a 14-month period, 2,805 patients underwent a pancreatoduodenectomy, distal pancreatectomy, total pancreatectomy or pancreatic enucleation at 43 institutions. After exclusion of patients without drains or data on drain removal, 1,841 patients were available for analysis. Early drain removal was defined as on or before POD#3. Early drain patients (n = 148, 8.1%) were younger, had more women and were less likely to undergo a pancreatoduodenectomy. Therefore, propensity score matching and sensitivity analyses were performed. Outcomes were determined by ACS-NSQIP and PDP definitions. Standard statistical tests were applied. Results: After propensity score matching early (n = 127) and late (n = 127) drain removal patients were well-balanced for age, gender, BMI, serum albumin, ASA class, operation type, time and approach, vascular resection, gland texture, duct size, pathology and POD#1 drain amylase. Outcomes are presented in the table. Conclusions: This analysis suggests that early drain removal is associated with reduced morbidity following pancreatic surgery. Early drain removal should be considered in selected pancreatectomy patients. HPB 2015, 17 (Suppl. 1), 1–81

Background: The outcome of Pancreaticoduodenectomy(PD) has been closely linked to hospital volume and experience. The low incidence of pancreatic cancer, coupled with few specialized Hepato-PancreaticoBiliary teams and lack of referral patterns and service centralization, contribute to sparse data from the Indian subcontinent. Methods: Prospective database of PDs from 1992 to 2014 was evaluated retrospectively over 4 time periods based on changing practice trends: A (1992–2001), B (2003-July 2009), C (August 2009-December 2011) and D (January 2012-August 2014). Peri-operative parameters were compared using SPSS v.21.0. Results: 751 patients underwent PD. The average resections increased from 14 to 94, over periods A to D, respectively. While post-operative pancreatic fistula(POPF) rates increased from 16% to 21.5% over periods A to D, the incidence of bile leaks and post-pancreatectomy hemorrhage declined to 0.8% and 3.9%, respectively. Morbidity and hospital stay was more in period D compared to B and C, but mortality declined from 6.3% to 2.8% over periods A to D. Overall series morbidity and mortality was 35.2% and 4.5%, respectively. Conclusion: This series represents the largest single-centre experience with PD from India and the surrounding region. The higher morbidity in period D is likely attributable to a combination of increasing surgical complexity and reduced selection bias. The gradually increasing experience should serve as a benchmark for developing dedicated pancreatic surgery teams even in regions of low incidence of pancreatic cancer where the need for training and centralization is greater compared to high incidence regions, such as USA and Europe.

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

62

Abstracts

OP-II.20 LEAKAGE OF AN INVAGINATION PANCREATICOJEJUNOSTOMY MAY HAVE LETHAL CONSEQUENCES H. Lavu1, S. W. Keith 1, E. M. Kilbane3, A. Parmar4, B. L. Hall 5, H. A. Pitt2 1 Thomas Jefferson University, Philadelphia, PA; 2Temple University School Of Medicine, Philadelphia, PA; 3Indiana University Health, Indianapolis, IN; 4University Of Texas Medical Branch, Galveston, TX; 5Washington University School Of Medicine, St. Louis, MO Background: No consensus exists regarding the most effective form of pancreaticojejunostomy following pancreatoduodenectomy. The aim of this analysis was to determine whether the type of pancreaticojejunostomy influences morbidity or mortality. Methods: Data were gathered through the American College of Surgeons-National Surgical Quality Improvement Program, Pancreatectomy Demonstration Project. Over 14 months, 1,781 patients underwent a pancreatoduodenectomy (PD) at 43 institutions. After exclusion of patients undergoing minimally invasive PD and those without information on gland texture or duct size, 890 patients were analyzed. Patients were divided into duct-to-mucosa (n = 734, 82%) and invagination (n = 156, 18%). Type of pancreaticojejunostomy (PJ) was then included in eight separate morbidity and mortality multivariable analyses. Results: Invagination patients had higher serum albumin (p < 0.01) lower BMIs (p < 0.01) and preoperative serum bilirubin (p < 0.02), were less likely to have a preoperative biliary stent (p < 0.01) or chemotherapy (p < 0.04), were more likely to have a soft gland (p < 0.01) and were less likely to undergo pylorus preservation (p < 0.01). Multivariable analyses demonstrated that age, gender, BMI, preoperative albumin and biliary stents, gland texture and pancreatic duct size were related (p < 0.05) to multiple postoperative morbidity outcomes. PJ anastomosis type was not associated with morbidity but did affect mortality (duct-tomucosa vs. invagination Odds Ratio 0.22, p < 0.01). Among patients who developed a pancreatic fistula, none of the 119 duct-to-mucosa compared to five of 20 invagination patients died (p < 0.01). Conclusions: Patients who undergo a pancreaticojejunostomy (PJ) by duct-to-mucosa or invagination differ with respect to pre- and intra-operative variables. When an invagination PJ leaks, the consequences may be lethal.

OP-II.21 SHOULD ACUTE PANCREATITIS BE AN INDICATION TO RESECT IPMN? J. L. Cioffi, S. J. Lee, J. A. Waters, C. M. Schmidt, A. Nakeeb, M. G. House, E. P. Ceppa, N. J. Zyromski Indiana University, Indianapolis, IN Introduction: Intraductal papillary mucinous neoplasms (IPMN) cause acute pancreatitis (AP) more commonly than generally appreciated. The natural history of IPMN with AP is unclear, and whether an episode of AP should be an indication for surgery in the setting of IPMN is controversial. We sought to determine the natural history of IPMN causing AP. Methods: 348 patients with pathologically proved, resected IPMN were analyzed. Patients with single versus multiple © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

episodes of AP were compared to determine clinical and pathological differences between these groups. Results: 114 (33%) IPMN patients had at least 1 episode of AP. Among IPMN/AP patients, 22 (19%) had more than one episode of AP. IPMN type (main duct, branch duct, mixed type), location (head versus body/tail), and size were similar between patients with single versus multiple AP episodes. Duration of symptoms related to IPMN was significantly longer for multiple AP patients (single episode 17 months, multiple episodes 37 months). Invasive carcinoma was present in 13 (11%) of all IPMN/AP patients, and was more common in patients with multiple AP episodes (23%) than those with a single AP episode (9%). Conclusions: Acute pancreatitis occurs in 33% of patients with resected IPMN, 19% of whom had multiple AP episodes prior to resection. Patients with multiple AP episodes were more likely to harbor invasive carcinoma compared to those with a single episode of AP. These data support early resection of IPMN patients who develop acute pancreatitis.

OP-II.22 ACCURACY OF PREOPERATIVE IMAGING FOR VASCULAR INVOLVEMENT IN LOCALLY ADVANCED, BORDERLINE RESECTABLE PANCREATIC ADENOCARCINOMA FOLLOWING NEOADJUVANT CHEMOTHERAPY J. Clanton, J. B. Rose, A. Alseidi, T. Biehl, S. Helton, F. Rocha Virginia Mason Medical Center, Seattle, WA Background: Radiographic imaging with arterial/venous phase, thin-slice computed tomography (MDCT) and endoscopic ultrasound (EUS) are utilized for assessment and staging of locally advanced, borderline resectable pancreatic adenocarcinoma (BR-LAPD). Neoadjuvant therapy followed by pancreaticoduodenctomy with vascular resection is typically required for BR-LAPD. However, little is known about the correlation of MDCT and EUS findings with need for vascular resection and pathologic invasion. Methods: A retrospective review of a prospectivelymaintained database of consecutive patients with BR-LAPD treated with neoadjuvant chemotherapy at our institution between 2011 and 2014 was performed. MDCT and EUS results were compared to operative findings and the final pathologic reports regarding vascular involvement. Results: A cohort of 33 patients underwent staging by MDCT and EUS followed by neoadjuvant chemotherapy and successful resection. All patients were reported to have venous involvement on CT, EUS, or both. Based on preoperative imaging and operative findings, a venous resection was performed in 15/33 (45.5%) patients, but histological evidence of vascular invasion was only noted in 6/33 (18.2%) pathologic specimens. These six patients were noted to have vascular invasion on both CT and EUS. CT demonstrated vascular involvement without EUS confirmation in 4/33 (12%) patients, while EUS demonstrated vascular involvement without CT confirmation in 3/33 (9%) patients. None of these patients had true pathologic vascular invasion. Conclusion: CT and EUS can be complementary modalities to detect BR-LAPD, but may overestimate the actual HPB 2015, 17 (Suppl. 1), 1–81

Abstracts incidence of venous involvement. This may be due to inherent limitations of these techniques versus a true downstaging effect of neoadjuvant chemotherapy.

OP-II.23 DISTAL PANCREATECTOMY PERFORMED BY HIGH VS LOW VOLUME SURGEONS IN A LARGE INTEGRATED HEALTH CARE SYSTEM: A RISK ADJUSTED ANALYSIS S. Woo, A. Difronzo Kaiser Permanente Los Angeles Medical Center, Los Angeles, CALIFORNIA Introduction: Improved outcomes for pancreatectomy are observed when performed at high-volume centers. The study objective was to compare predicted and actual risk-adjusted outcomes of distal pancreatectomy (DP) performed by high vs low volume surgeons in an integrated health care system. Methods: A retrospective review of patients having DP from 2005–2013. A standardized risk calculator was used to predict risk of complications and mortality. The Clavien grading system was used for actual complications. High volume was 7 or more DP per year. Results: 50 surgeons performed 247 DP. 3 high volume surgeons performed 54 cases. In the low volume cohort, 193 DP were performed (median 2 per surgeon). The calculated risk of any complication and risk of mortality was no different between the two groups, but was significantly lower than actual observed complications. There were more laparoscopic (56 vs 33%, p = 0.001) and spleen-preserving cases in the high volume group (59 vs 18%, p < 0.001). There was a significant difference in overall actual 60-day complications: 33% in the high volume cohort vs 52% (p = 0.008). There were fewer Clavien grade 1/2 complications (15 vs 37%, p < 0.001), and less postoperative pancreatic fistula (PPF) (11% vs 33%, p = 0.0009) in the high volume group. There was no difference in mortality between the groups. Conclusion: Distal pancreatectomy performed by high volume surgeons in a large health care system results in decreased risk-adjusted overall morbidity including PPF. Some observed differences may be related to greater use of laparoscopy and splenic preservation.

OP-II.24 PORTAL VEIN THROMBOSIS AFTER TOTAL PANCREATECTOMY AND AUTOLOGOUS ISLET CELL TRANSPLANTATION W. P. Lancaster, D. B. Adams, K. A. Morgan Medical University Of South Carolina Department Of Surgery, Charleston, SC Introduction: Portal vein thrombosis (PVT) is a rare complication of total pancreatectomy with autologous islet transplantation (TPIAT). Little is reported about the risk factors, consequences, or treatment for this complication. Methods: A retrospective review and analysis of a prospectively-collected database of patients undergoing TPIAT from March 2009 to August 2014 was conducted. Two-tailed t-tests were used comparing continuous data and Fisher’s exact test comparing categorical data. Results: 135 patients (102 women,76%) underwent TPIAT; Nine(7%) had PVT. All patients with PVT were women. HPB 2015, 17 (Suppl. 1), 1–81

63

There were no differences in age or islet equivalents transplanted in patients with and without PVT. Mean BMI of patients with PVT was lower than those without (21.8 vs 26.5 kg/m2,p = 0.03).Mean portal pressure post-islet infusion was higher in patients with PVT (25.2 vs 16.0,p = 0.0007), with 4/9 having pressures over 30 mmHg. The median time to diagnosis of PVT was 10.5 days postoperative(range 7 to 210),with 7/9 having negative duplex POD1. Eight of 9 patients with PVT were treated with systemic anticoagulation and 7/8 had resolution on repeat imaging. One patient died from complications of anticoagulation. Two patients developed cavernous transformation(CTPV), one untreated and one diagnosed after CTPV. All patients with PVT were insulin-requiring at latest follow-up versus 72/94 patients(77%) without PVT with at least 1-year follow-up(p = 0.035). Conclusions: PVT following TPIAT is an uncommon but serious complication. It occurs late in the postoperative period in women with a low BMI. A standardized follow-up imaging protocol is suggested. The treatment for PVT is anticoagulation. Patients with PVT can expect to be insulindependent.

OP-II.25 READMISSION AFTER PANCREATECTOMY FOR CANCER; A DATA-DRIVEN APPROACH TO QUALITY IMPROVEMENT S. M. Misustin, K. K. Christians, F. M. Johnston, E. A. Krzywda, S. L. Lahiff, S. Tsai, D. B. Evans Medical College Of Wisconsin, Milwaukee, WI Background: Hospital readmissions have been proposed as a hospital quality measure. Analysis of readmissions has largely focused on post-discharge management. Assessment of pre-discharge data may reveal additional information. Methods: Pancreatic cancer (PC) pts who underwent resection were identified. The causes of readmissions within 30 days of discharge were reviewed. Results: Surgical resection was performed on 210 consecutive pts. The median length of stay of the index operation was 8 days (IQR: 4). Postoperative complications were identified during the index hospitalization in 45 (22%) pts, with one perioperative death. Pts were discharged to: home with family (48; 23%), home with visiting nurse (144; 69%), other facility (15; 8%). Readmission occurred in 29 (14%) pts. Post-discharge analysis identified 21 (72%) of 29 readmissions related to the operation; infection (6), GI bleed (3), delayed gastric emptying (3), pancreatic fistula (6), ileus (1), SBO (2). An additional 8 (28%) of the 29 readmissions were due to potentially preventable causes including; hypoglycemia (1), dehydration/anorexia (5), and pain (2). However, pre-discharge analysis using data available on the day of discharge suggested that 9 (31%) of the 29 readmissions should not have been discharged and could have been identified prior to discharge. Conclusions: Pre-discharge analysis identified data elements in 9 (31%) of the 29 readmissions that could have prevented discharge, but were undetected in post-discharge analysis. To prospectively identify these pts prior to discharge, the causes of these 29 readmissions were used to create a checklist and a formal “time out” discharge procedure has been implemented. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

64

Abstracts

OP-II.26 TITLE D. P. Nussbaum, L. M. Youngwirth, R. R. White, B. M. Clary, J. A. Sosa, D. G. Blazer Duke University Medical Center, Durham, NORTH CAROLINA Introduction: Pancreatic acinar cell carcinoma (pACC) has cure rates up to 40% following resection, yet many patients with localized disease do not undergo surgery. Methods: The 1998–2011 National Cancer Data Base was queried for patients with pACC. Among patients with localized disease, multivariable analysis was used to predict the likelihood of undergoing resection. Cox proportional hazards modeling was then used to assess variables associated with survival following resection. Results: 933 patients were identified. Median age at diagnosis was 64 years. Tumors were most common in men (66%) and white patients (88%), and occurred most frequently in the pancreatic head (57%). Mean size was 6.6 cm. While 42% of patients presented with localized disease, nearly one-quarter of these patients did not undergo resection. Median survival was 55 months following resection, compared to 23 months without surgery (p < 0.01). Failure to undergo surgery was associated with older age (OR 1.32, p = 0.02), male sex (OR 2.30, p < 0.01), black race (OR 2.86, p = 0.03), higher grade (OR 2.45, p = 0.03), location within the head (OR 3.33, p < 0.01), and treatment at a nonacademic facility (OR 2.09, p < 0.01). Following adjustment, only older age (HR 1.17, p = 0.01) and lymph node metastases (HR 2.58, p = 0.04) were associated with increased mortality following resection (Table 1). Conclusions: Survival following resection of pACC is nearly five years, yet specific subsets of patients appear less likely to undergo surgery. Of these groups, only older age is independently associated with mortality. Efforts to increase access to care could result in improvements in survival for patients with pACC.

OP-II.27 RADIOGRAPHIC RESPONSE AND RESECTABILITY OF LOCALLY ADVANCED, BORDERLINE RESECTABLE PANCREATIC ADENOCARCINOMA AFTER EXTENDED NEOADJUVANT CHEMOTHERAPY J. Clanton, J. B. Rose, A. Alseidi, T. Biehl, S. Helton, F. Rocha Virginia Mason Medical Center, Seattle, WA Background: Neoadjuvant therapy is often utilized for locally advanced, borderline resectable pancreatic cancer (BR-LAPD) prior to resection, despite previous reports suggesting minimal downstaging. However little is known about the effect of extended preoperative chemotherapy on disease response. Methods: Retrospective review of a prospective database of consecutive patients with BR-LAPD treated with a 24-week course of gemcitabine and docetaxel between 2011 and 2014 was performed. Patients with high-quality imaging before and after treatment, and who completed full course of therapy with intention to resect were included. RECIST 1.1 criteria were used to assess radiographic response. Results: Forty-five patients who completed extended neoadjuvant chemotherapy were included for analysis. Thirty-two (71.1%) patients underwent pancreaticoduodenectomy and 14 (31.1%) received a vein resection. Thirteen (28.9%) were not resected due to local or distant progression or reduced performance status. The majority of patients demonstrated stable tumor size (24/45, 53.3%) or partial response (20/45, 44.4%). Suspicious lymph nodes remained stable (27/45, 60%) or improved (11/45, 24.4%) while vascular involvement was stable (21/45, 46.7%) or improved (16/45, 35.6%) after treatment. None of the ten patients not resected after chemotherapy had an increased tumor size, three had enlarging lymph nodes, and one had more extensive vascular involvement by CT. None of these factors were associated with ability to undergo resection (p > 0.05) Conclusions: Extended neoadjuvant chemotherapy for BR-LAPD commonly results in either stable disease or a partial response. One-third of patients have demonstrable downstaging of vascular involvement, however radiographic response or downstaging is not associated with successful resection.

OP-II.28 PATIENT SELECTION FOR ROBOT-ASSISTED PANCREATODUODENECTOMY FAVOURS OBESITY N. Ali, A. T. Stafford, G. Morris-Stiff, J. Wey, K. ElHayek, S. Chalikonda, R. M. Walsh Department Of HPB Surgery, Cleveland, OHIO Introduction: There are potential advantages to the application of minimally invasive techniques to pancreatoduodenectomy (PD), with current data indicating its technical feasibility. We present a single center experience of robot-assisted PD (RAPD) focusing on patient selection and outcomes, in particular infections, comparing results to open PD. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts Methods: A retrospective review of a prospectively maintained database of all open PDs and RAPDs from March 2009 to June 2014 was performed. RAPD has been introduced selectively during this period. As the main outcome assessment was infection rate (total/deep organ space and surgical site [SSI]), patients converted from RAPD to open were included in the open group. Results: 69 patients underwent completed RAPD, and 372 open PD during the study period. There was a change in disease etiology from predominantly premalignant lesions and ampullary carcinomas to include all indications for PD. There was a significant difference in BMI between groups with RAPD preferentially used in obese patients (p = 0.004) that evolved during the study. Comparing RAPD and open PD, the SSI rate was less in the RAPD group (9% vs.14%) with no difference in overall infective complication rate (26% vs. 26%) or intra-abdominal infection (10% vs. 10%). Estimated blood loss was significantly less (200 versus 400 cc (p < 0.001) in favor of RAPD. Conclusion: Increasing experience with RAPD has led to a change in utilization of the procedure and a transition to its use in all pathologies. To maximize its benefits we have a positive selection bias towards obese patients.

OP-II.29 PREDICTORS OF ACTUAL SURVIVAL IN RESECTED PANCREATIC ADENOCARCINOMA: A POPULATION-LEVEL ANALYSIS D. J. Kagedan1,2, R. Raju2, M. Dixon3, E. Shin1, Q. Li5, N. Liu5, M. Elmi1,2, A. El-Sedfy4, C. Earle2, N. Mittmann2, N. G. Coburn2 1 University Of Toronto, Toronto, ON; 2Sunnybrook Health Sciences Centre, Toronto, ON; 3Maimonides Medical Center, Brooklyn, NY; 4Saint Barnabas Medical Center, Livingston, NJ; 5Institute For Clinical Evaluative Sciences, Toronto, ON Introduction: Among patients diagnosed with pancreatic adenocarcinoma, numerous clinicopathologic factors have prognostic value following curative-intent resection. We sought to assess actual survival following resection and to determine factors predictive of survival on a population level. Methods: Patients undergoing resection for pancreatic adenocarcinoma between 2005–10 were identified within the provincial cancer registry and administrative databases that include actual survival for all patients in Ontario, Canada (population 13 million). We fully abstracted pathology reports for 473. Kaplan-Meier survival analysis and Cox proportional hazards multivariate regression were performed to determine the clinicopathologic variables associated with decreased survival. Results: The actual 1-, 3-, and 5-year survival rates were 65%, 23%, and 15% respectively, with median survival 1.48 years. Follow-up time ranged from 2.07–7.22 years, and 377 (79.7%) were censored for death before the end of follow-up. Multivariate regression revealed the following variables to be negatively associated with survival: age >70 (p = 0.001), T stage (p < 0.01), nodal metastasis (p < 0.001), tumor grade (p < 0.001), positive margin status (p < 0.01), lymphovascular invasion (p < 0.001), lymph node positivity ratio >0.2 (p < 0.001). Patients with multivisceral or major vascular resections, and patients with low socioeconomic status HPB 2015, 17 (Suppl. 1), 1–81

65

did not have worse survival. Receiving treatment at a high-volume hepatopancreatobiliary center was associated with improved survival (HR = 0.49, 95%CI = 0.36–0.67, p < 0.0001). Conclusion: Advanced age, positive margins, and histopathologic tumor characteristics predict poor prognosis, and undergoing more extensive resection does not worsen survival. Receiving treatment at a hepatopancreatobiliary centre improves survival. In a publicly-funded healthcare system, poor socioeconomic status does not worsen survival.

OP-II.30 DOES THE MD ANDERSON (MDA) CRITERIA OF BORDERLINE RESECTABLE (BLR) PANCREATIC CANCER IMPACT SUCCESSFUL COMPLETION OF NEOADJUVANT THERAPY AND SURVIVAL? A. N. Krepline, K. K. Christians, B. George, P. S. Ritch, B. A. Erickson, K. Oshima, P. Tolat, D. B. Evans, S. Tsai Medical College Of Wisconsin, Milwaukee, WI Background: BLR PC has been classified based on tumorvessel abutment/encasement or the radiographic presence of indeterminate metastatic lesions. The importance of this subclassification on outcome is unclear. Methods: BLR PC patients were classified as: type A, anatomic criteria; type B, indeterminate CT findings including suspicious peripancreatic lymphadenopathy (LAD), or a CA19-9 >2000 (hiCa19-9). Patients meeting both type A and B criteria were classified as type B. Results: Of the 110 BLR patients, 56 (51%) were type A and 54 (49%) were type B. Restaging after neoadjuvant therapy (Neoadj), demonstrated metastases in 8 (14%) of 56 type A vs. 11 (20%) of 54 type B patients (p = 0.82). Neoadj and surgery (Surg) was completed in 66 (60%) of the 110 patients; 37 (66%) type A and 29 (54%) type B (p = 0.19). Of the 54 type B patients, Neoadj and Surg was successfully completed in: 8 (42%) of 19 with indeterminate CT findings; 15 (56%) of 27 with suspicious LAD; and 6 (86%) of 7 with hiCa19-9 (p = 0.14). Overall survival (OS) for type A vs. B patients was 17 vs.19 months (p = 0.93). Of the 66 patients who completed Neoadj + Surg, type A and B patients had OS of 31 vs. 42 months respectively (p = 0.38); there were no OS differences between type B subgroups. Conclusions: In a contemporary series of 110 pts with BLR PC, 60% completed neoadjuvant therapy including surgery. Subgroup categorization of BLR patients into type A and B had no prognostic significance.

OP-II.31 THE IMPACT OF MINIMALLY INVASIVE DISTAL PANCREATECTOMY ON 90-DAY READMISSIONS AND COST: IS IT ANY BETTER THAN OPEN? J. Parikh, S. Anantha Sathyanarayana, S. Bendix, M. J. Jacobs Providence Hospital Medical Center, Southfield, MI Introduction: Laparoscopic distal pancreatectomy (LDP) is commonly performed for lesions of pancreatic body and tail. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

66

Abstracts

Surprisingly, recent literature suggests an increase in readmission rates after LDP, hence potentially negating any gained length of stay (LOS) benefit compared to open distal pancreatectomy (ODP). Therefore, we sought to examine readmission rates and total cost of LDP versus ODP at a high-volume community hospital. Methods: Between January 2003 to December 2013, 81 distal pancreatectomies were performed at a community teaching hospital. A retrospective analysis on demographics, 90-day outcomes, readmission rates, length of stay (LOS), and total cost were collected. Results: Eighty-one patients underwent distal pancreatectomy (41 open and 40 laparoscopic). Median age was 62 years. Two-thirds of patients were female. LDP had significantly shorter mean operative time (150 vs. 183 minutes; p < 0.01) and decreased blood loss compared to ODP (135 vs. 568 mL; p < 0.001). Table 1 compares tumor characteristics, LOS, readmission rates, and costs. Pancreatic fistula rates were comparable with no Grade C fistulae in either group. Overall 90-day morbidity was lower in the LDP group with no mortalities. The 30-day and 90-day readmission rate was lower in LDP; hence LDP has lower total hospital days. The overall costs for both the index admission and the total hospital stay (including readmission) were lower for LDP group. Conclusion: LDP has significantly lower index LOS, fewer total hospital days and lower overall costs compared to ODP. LDP should be the standard of care for amenable lesions in the body or tail of the pancreas.

OP-II.32 SHOULD SMALL PANCREATIC NEUROENDOCRINE TUMORS BE OBSERVED OR RESECTED? R. E. Eldert, N. Valsangkar, M. Kilbane, J. L. Cioffi, M. G. House, N. J. Zyromski, C. M. Schmidt, A. Nakeeb, E. P. Ceppa Indiana University School Of Medicine, Indianapolis, IN Background: Pancreatic Neuroendocrine Tumors (PNET) are heterogeneous in behavior and metastatic potential. Recent series recommend surveillance of PNET less than 2 cm in size; consensus is lacking whether a size cutoff for PNET can be used to define malignant potential and guide therapy. The aim of this study was to determine if there were differences in cancer-specific outcomes following resection for tumors 2 cm or less versus greater than 2 cm. Methods: Consecutive cases of resected nonfunctional PNET (n = 48) were reviewed retrospectively at a high volume academic center between 2004 and 2009. Data were gathered through the American College of SurgeonsNational Surgical Quality Improvement Program and electronic medical records. Survival was determined from the institutional cancer registry and Social Security Death Index. Results: Patients with PNET 2 cm or less (n = 16) versus those greater than 2 cm (n = 32) had no significant difference in demographic, preoperative, or intraoperative variables. Surgical pathology was reviewed (Table). Patients underwent distal pancreatectomy (56%), pancreaticoduodenectomy (29%), enucleation (8%), central pancreatectomy (4%), and Frey-procedure (2%). Mortality at 30 and 90 days was 0% for both groups. The 2 cm or less group had a lower 30-day and 90-day morbidity (p < 0.01). There was no significant difference in five-year survival between groups. Conclusions: Resected nonfunctional PNET 2 cm or less appear to have a similar rate of lymph node and distant metastasis and no significant benefit of 5 year survival compared to larger PNETs. The lower rate of postoperative morbidity and equivalent overall malignant nature of PNET supports resection over observation of PNET 2 cm or less.

OP-II.33 CHOLEDOCHODUODENOSTOMY IS A SAFE AND EFFICIENT ALTERNATIVE FOR BILE DUCT RECONSTRUCTION DURING LIVER TRANSPLANT T. L. Nydam University Of Colorado School Of Medicine, Aurora, CO In our institution, choledochoduodenostomy (CDD) has become the bile duct reconstruction of choice in liver transplants when a duct-to-duct choledochocholedocostomy (D2D) is not possible. We provide evidence that CDD is a safe option for bile duct reconstruction with significant advantages during postoperative care. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts Methods: All orthotopic liver transplants performed at The University of Colorado Hospital from July 2006 to July 2013 were retrospectively reviewed. Patient demographics, donor type, post-transplant complications, ERCP times, and biliary percutaneous transhepatic interventions (PTC) were collected. Statistical analysis was performed using a paired student’s t-test assuming equal variances. Results: 632 liver transplants were performed. Eighty-two patients underwent CDD, 28 patients underwent Roux en Y choledochojejunostomy (CDJ), and 522 patients underwent D2D. There was no statistical difference in cholangitis, bile leak, anastomotic stricture, or other complications. However, there was a statistically significant difference in mean length of ERCP and number of PTCs between the CDD and CDJ cohorts. (Table 1). Conclusions: Contrary to traditional teaching, a CDD reconstruction appears to have no difference in complications compared to a CDJ reconstruction. In addition, the length of time spent during ERCP and the number of PTCs required were significantly lower in the CDD cohort. In our institution, CDD is a safe option for bile duct reconstruction during liver transplant that provides improved postoperative access to the graft biliary system.

67

Methods: Twelve high risk LTx from 2011 to 2014 were reviewed. Sorafenib started within 2 months if: microvasular invasion (MVI), beyond Milan, or AFP >500. Dose was escalated from initial 400 mg/d. Results: 75% were male, average BMI was 31, and 66% had bilobar tumors. Pathologic staging revealed ave. 2 tumors, total tumor burden 5.1 cm, 75% either stage B or C (BCLC staging), and MVI in 42%. AFP was >100 in 17%. Ave. dose was 265 mg in all patients but 300 mg in those that tolerated treatment. Four (33%) of patients are still on treatment, but 5 ( 42% ) received it for less than 50 days. Side effects included hand foot syndrome, flushing, anxiety,nausea, diarrhea, hypertension, neutropenia, arthralgias and abdominal pain. No recurrences were observed. Eleven (92%) of patients are alive, all without recurrence. Conclusion: This pilot study of adjuvant sorafenib post LTx revealed that 68% of targeted patients received adequate treatment at tolerable doses of 300 mg daily. Recurrences were not observed in this high risk group, although follow up is short. The study gives insight to the current multicenter sorafenib adjuvant post LTx trial.

OP-II.35 DIPS VS. TIPS – IS THERE ANY DIFFERENCE WHEN IT COMES TO TRANSPLANTATION? J. L. Pasko, R. C. Schenning, B. D. Petersen, S. L. Orloff Oregon Health And Science University, Portland, OR

OP-II.34 A PILOT STUDY OF ADJUVANT SORAFENIB IN HIGH RISK LIVER TRANSPLANTATION FOR HCC A. E. Alsina1,2, E. S. Franco1,2, A. Makris4, J. Shim3, J. Allison3, R. Claudio3, M. Johnstone3, K. Barber5, C. Albers1, N. Kemmer1 1 Tampa General Medical Group, Tampa, FL; 2University Of South Florida, Department Of Surgery, Morsani College Of Medicine, Tampa, FL; 3Tampa General Hospital, Office Of Clinical Research, Tampa, FL; 4 University Of South Florida, School Of Business, Tampa, FL; 5Tampa General Hospital, Tampa, FL Introduction: Sorafenib can prevent post liver transplant (LTx) recurrences for hepatocellular carcinoma (HCC) (Yan, J. Liver Transpl 2013), but data in humans is scarce. We hypothesized that recurrences in high risk HCC LTx can be reduced. Aims: Determine applicability, tolerability, and efficacy. HPB 2015, 17 (Suppl. 1), 1–81

Background: Direct intrahepatic portocaval shunt (DIPS) is a modification of transjugular intrahepatic portosystemic shunt (TIPS) commonly used for ESLD patients with refractory ascites, recurrent variceal bleeding, and end-stage liver disease. DIPS uses ultrasound guidance to stent a communication between the IVC and the portal vein. While TIPS is more readily utilized, DIPS have shown to be helpful for patients with difficulty anatomy and failed TIPS. Objective: To assess whether liver transplantation is safe and feasible in recipients with DIPS. Secondary aims were to do a matched comparison of DIPS vs.TIPS transplant recipients with respect to autologous and PRBC transfusions, warm ischemia time, cold ischemia time, LOS, graft survival, and overall patient survival. Methods: Retrospective review January 1994–December 2011 in a shared two academic institution transplant program of patients who underwent liver transplantation with an existing DIPS or TIPS. The transplant operation was either bicaval (before 2007), or onlay cavacavostomy (after 2008). Results: Nine DIPS and 27 TIPS patients underwent transplantation during the study period. Nine DIPS were matched to 18 TIPS patients by age, gender, etiologic disease, and MELD at transplantation. There was was no statistical difference in amount of autologous and PRBC blood transfusion, warm ischemia, cold ischemia, LOS, graft survival, or overall survival. Conclusion: DIPS when compared to TIPS is safe, and feasible, and carries no added risk to patients undergoing liver transplantation. Potential liver transplant candidates who are referred for DIPS based on unfeasible or failed TIPS should not be denied this treatment option. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

68

Abstracts

OP-II.37 INCIDENCE AND LONG-TERM SURVIVAL OF PATIENTS WITH DE-NOVO HEAD AND NECK CARCINOMA AFTER LIVER TRANSPLANTATION A. Andreou1, M. Lenarz2, D. Seehofer1, J. Pratschke1, A. Coordes2 1 Department Of General, Visceral And Transplant Surgery, Charite Campus Virchow-Klinikum, Berlin, BERLIN; 2 Department Of Otorhinolaryngology, Head And Neck Surgery, Charite Campus Benjamin Franklin, Berlin, BERLIN Background: Liver transplant recipients have an increased risk for the development of de-novo malignancies. Methods: Clinicopathological data of patients who developed head and neck cancer after liver transplantation (LT) were evaluated and predictors for overall survival (OS) were identified. Results: Thirty-three of 2040 patients (1.6%) who underwent LT between 1988 and 2010 developed de-novo squamous cell carcinoma of head and neck (HNSCC). The incidence of HNSCC in LT recipients with end stage alcoholic liver disease was 5%. After a median follow-up of 9 years, 1-, 3-, and 5-year OS rates were 74%, 47% and 34%, respectively. Tumor size, cervical lymph-node metastasis, tumor site and tumor therapy (surgery only versus surgery and adjuvant radio/chemoradiotherapy versus radio/ chemoradiotherapy only, P < 0.0001) were significantly associated with OS in univariate analysis. However, surgery only independently predicted OS in multivariate analysis. Conclusions: Early diagnosis and surgical treatment for de-novo HNSCC are crucial for outcome. Therefore, HNSCC risk should be taken into close consideration during post-transplant follow-up examinations especially among patients with positive history of smoking and alcohol consumption.

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

OP-II.38 IMPACT OF BODY MASS INDEX ON LIVER TRANSPLANT PATIENT SURVIVAL: 1994–2013 S. Ayloo1, P. Talbot2, M. Molinari1 Dalhousie University, Halifax, NS; 2Diabetes Care Program Of Nova Scotia, Halifax, NS 1

Back Ground: Obesity is a world-wide epidemic impacting the field of transplantation. In 2012, more than one-third of the United States liver transplantations (LT) were performed on patients with body mass index (BMI) ≥30. Objectives: To explore the OPTN/UNOS database for donor and recipient characteristics and transplant outcomes among first time, whole organ, cadaveric LT patients stratified by different weight-status. Methods: Retrospective study of the SRTR database from January 1994 to September 2013 was conducted to analyze donor/recipient characteristics and transplant outcomes stratified by underweight (BMI ≤18.5), normal (BMI ≥18.5– 24.99), overweight (BMI 25–29.99), Class I (BMI 30–34.99), class II (BMI 35–39.99) and class III obese (BMI ≥40). Patients with missing BMI were excluded. Transplant outcomes-Patient survival by Kaplan-Meier, cause of death, length of hospital stay was analyzed. Results: Of the 222,000 cases, 66,461 met the inclusion criteria. The donor characteristics of mean age, median BMI, cold ischemia time, percent male, Caucasian ethnicity and leading cause of death of head trauma were 40 years, 25.4(7.4–73.22), 7 hours(0–49.5), 60%, 71% and 41% respectively. Recipient characteristics and outcomes are presented in Table 1. Kaplan-Meier shows lower survival for underweight and obese. BMI is significantly associated with gender, ethnicity, education, functional status, primary diagnosis, comorbidities, graft failure (acute/de novo hepatitis), primary cause of death. Conclusion: LT in obese patients is increasing, with increasing proportion of NASH as an indication for transplantation. Cardiovascular events were significantly higher in higher BMI groups. Patient survival was lower in underweight and obese Class III in comparison to other BMI grouping.

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

OP-II.39 THE EFFECT OF A PORTAL PRESSURE < 15 ON THE PREVENTION OF SMALL-FOR-SIZE SYNDROME IN ADULT-TO-ADULT LIVING-DONOR LIVER TRANSPLANTATION: AN OBSERVATIONAL STUDY A. S. Helmy1, A. Abdelhady1, A. Hosny1, A. Dahaba3, M. A. Al-Shazly1, M. El-mansy2, S. Uemoto4 1 Cairo University, Cairo, GIZA; 2Department Of Clinical Pathology, Theodor Bilharz Research Institute, Cairo, GIZA; 3Department Of Anaesthesiology And Intensive Care Medecine, Graz, GRAZ; 4Department Of Surgery, Division Of HPB Surgery And Transplantation, Graduate School Of Medicine, Kyoto, KYOTO Background: Small-for-size liver graft injury in Adult-toadult living-donor liver transplantation (A-LDLT) can contribute to severe postoperative graft dysfunction, known as Small-for-size syndrome (SFSS). Recent evidence implicates portal hypertension in SFSS pathogenesis. The aim of our study was to investigate the effect of a portal venous pressure (PVP) cut-off value of <15 mmHg on the prevention of SFSS in A-LDLT. Methods: Thirty-three patients underwent A-LDLT between October 2009 and June 2013. We aimed to keep the PVP under 20 mmHg at the end of the operation using graft inflow modulation (GIM). Patients fell into 2 groups; group A with final PVP <15 mmHg (n = 16) and group B with final PVP ≥15 mmHg (n = 17). We diagnosed postoperative SFSS according to the Clavien definition. HPB 2015, 17 (Suppl. 1), 1–81

69

Results: Final PVP was controlled under 20 mmHg in all patients. Three patients suffered SFSS in group B (17.6%) compared with no patients in group A (P = 0.078). There was a higher proportion of dead patients in group B [4, two of whom died of SFSS (23.5%)] versus group A [1 (6.2%); P = 0.166]. Conclusion: A final PVP <15 mmHg correlates with better outcomes and seems to be a safe target level for the prevention of postoperative SFSS in A-LDLT.

OP-II.40 COMPLETE PATHOLOGIC TUMOR KILL IS NOT REQUIRED AFTER BRIDGING THERAPY FOR HCC E. Beal, S. Black, M. Bloomston, C. Schmidt The Ohio State University, Columbus, OH Introduction: In patients with hepatocellular carcinoma (HCC), bridging therapy is common prior to liver transplant (LT) using ablation or trans-arterial chemoembolization (TACE). It is unknown whether complete pathologic tumor kill is needed for superior outcome. Methods: The medical records of all patients who underwent LT between 2008 and 2013 at a single hospital were reviewed. The incidence of viable HCC within the hepatic explant was determined for patients who underwent LT after bridging therapy. Outcomes were compared between those with complete and incomplete tumor kill. Results: There were 125 patients who underwent LT during the study period, and 42 (34%) had HCC. Twenty-eight © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

70

Abstracts

(65%) were treated with bridging therapy, specifically 13 (46%) underwent ablation and 15 (54%) underwent TACE. Viable tumor within the hepatic explant was found in 11 (69%) patients after ablation and 10 (67%) after TACE. Median follow-up after LT in the HCC group was 17.3 months. Estimated 5-year survival was similar in the ablation (75%) and TACE (76%) groups. One death occurred in the perioperative period (3%). There was one patient with recurrent cancer in the cohort initially treated with ablation who died 13.3 months after LT with multifocal disease. Conclusions: After bridging therapy with either ablation or TACE, rates of viable HCC are high in the explant after LT. Despite this, recurrence is uncommon and overall survival is comparable to patients transplanted without HCC. Depending on regional waitlist mortality, it may not always be necessary to achieve complete pathologic tumor kill with bridging therapy.

OP-II.41 EXPANSION OF HPB AND LIVER TRANSPLANTATION IN AN UNDERSERVED STATE: CREATING GROWTH AND MAINTAINING QUALITY C. D. Anderson, A. Seawright, B. Borg, T. M. Earl University Of Mississippi Medical Center, Jackson, MS Significant disparities in HPB/liver transplant care prompted expansion of Mississippi’s only transplant program. The strategic plan hypothesized that a systems based approach would allow growth while maintaining quality. Methods: Systems were implemented to start the program in 8/2011 including a streamlined transplant clinic, increased outreach, a robust QAPI system, and the creation of a liver tumor program. This study examined institutional wide data from 1/1/2010 until 3/31/2014. Period 1 (1/1/2010–7/31/ 2011) represented pre-expansion, and period 2 (1/1/2013–3/ 31/2014) was after full implementation. The performance of the liver tumor, and liver transplant programs was analyzed. Data is reported by median. Periods are compared using the student t-test. Results: From 8/2011 until 3/2014, 222 patients with liver tumors were evaluated. Loco-regional liver tumor treatments increased from 5/quarter to 25/quarter (p < 0.001). Major HPB operations increased from 3/quarter to 14/quarter (p < 0.001). This increase was within the transplant program (14 vs 2/quarter). The first liver transplant occurred in 3/2013. 29 liver transplants were performed in 13 months. 1-month allograft and patient survivals were 93% and 100%. Actuarial 1-year allograft and patient survivals were 89.66% and 96.55%. MELD at transplant was 22 (range 15–40). 35% of transplant recipients were African American (13% regionally, 10% nationally). LOS was 9 days. 30 day readmission was 20.69%. Conclusion: A disparity in HPB and liver transplant access existed in Mississippi which was addressed via a comprehensive expansion of the abdominal transplant programs. The approach allowed rapid creation of a high volume transplant/ HPB center and improved patient access while maintaining expected outcome metrics.

OP-II.43 RACIAL DISPARITY IN LIVER TRANSPLANTATION IN AN INNER CITY POPULATION A. Smith, A. T. Hauch, E. Kandil, A. Paramesh, M. Killackey, M. Moehlen, L. A. Balart, J. F. Buell Tulane Transplant Institute, Tulane University And Louisiana State University Medical System, New Orleans, LA Several studies have identified disparity in access to liver transplantation among African Americans. This study examines the experience of an inner city University transplant program with a significant proportion of socioeconomically challenged African American patients. Methods: A retrospective analysis of all liver transplant (LT) recipients at our institution from 2007–2014. Pretransplant demographics and post-transplant outcomes were analyzed. A multivariate analysis was performed to examine patient and graft survival as well as risk for complications. Conclusions: African American patients that were successfully transplanted at our institution were significantly younger with trends of lower portal vein thrombosis and acutely decompensated (higher INR, inpatient transplant event). Despite limited evidence these data suggest a selection bias. To further delineate the origins of this bias we plan on further analyzing waitlist mortalities and dropout events along racial lines to identify if African Americans are at higher risk for dropout under this current allocation system.

OP-II.44 ADDITION OF STEREOTACTIC BODY RADIATION TO PREVENT LIVER TRANSPLANT WAITLIST DROPOUT: IS IT SAFE? T. J. Smith3, J. E. Hooper6, A. E. Castillo1,2, W. Naugler5, P. Worth1,2, K. Kolbeck7, K. Farsad7, K. Enestvedt1,2, M. Fuss4, T. Mitin4, C. Thomas4, S. Orloff1,2 1 Oregon Health And Science University. Department Of Surgery, Portland, OR; 2Portland VA Medical Center. Department Of Surgery, Portland, OR; 3Gundersen Health System. Department Of Surgery, La Crosse, WI; 4Oregon Health And Science University. Department Of Radiation Oncology, Portland, OR; 5Oregon Health And Science University. Department Of Medicine, Portland, OR; 6Johns Hopkins. Department Of Pathology, Baltimore, MD; 7 Oregon Health And Science University. Department Of Radiology, Portland, OR Background: Waitlist dropout is a major consideration in liver transplantation (LT) for HCC. Loco-regional treatment is used to prevent dropout. Little is known about Stereotactic

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts Body Radiation (SBRT) safety and its influence on waitlist dropout, perioperative LT complications, or recurrence post-LT. Aim: Whether adding SBRT to HCC LT waitlist patients having received Trans-Arterial-Chemo-Embolization (TACE), impacts the safety/efficacy profile as measured by waitlist dropout, perioperative complications or posttransplant HCC recurrence. Methods: Retrospective analysis from a two-institution transplant program, of 10 consecutive waitlisted HCC patients receiving SBRT + TACE matched with 10 such patients receiving TACE only. Results: Median treatment follow-up: SBRT + TACE group 45 vs 43 mo TACE-only group; median LT follow-up: 34 and 38 mo, respectively. Mean wait-time for the SBRT-group was 330 d vs 150 d for the TACE-only group. At last follow-up 8/10 SBRT-group patients were alive vs 6/10 in the TACEonly group. One HCC recurrence occurred in the SBRTgroup vs two in the TACE-only group. All HCC recurrences died. Pre-transplant median AFP was higher in the SBRTgroup. Tumor response by explant percent necrosis was similar between groups. LOS, ICU days and median EBL trended higher for the SBRT group (p = 0.60). There were no waitlist drop-outs in either group. Conclusions: Despite longer waitlist time, HCC patients receiving SBRT + TACE had no waitlist drop-out, and lower HCC recurrence post-LT. SBRT + TACE patients trended toward more difficult operations, but no significant difference in post-LT survival. Thus, addition of pre-LT SBRT to TACE appears safe and effective as a bridge to LT. Future prospective randomized clinical trials are warranted.

SUNDAY, MARCH 15, 2015, 7:30AM–8:30AM VIDEO A – LIVER V-A.01 TIPS AND TRICKS FOR GLISSONIAN APPROACH DURING LAPAROSCOPIC RIGHT HEPATECTOMY F. F. Makdissi, R. C. Surjan, M. A. Machado Sirio Libanes Hospital, Sao Paulo, SAO PAULO Background: Laparoscopic Glissonian approach is a technique that avoids dissection of the hilar plate. It is useful for anatomical liver resection and can reduce the operative time. However, it has not been frequently used because needs expertise and knowledge of liver anatomy. Aim: To present a video of a laparoscopic right hepatectomy using the Glissonian approach. The detailed technique for control of the right pedicle is highlighted in this video. Patient and Methods: A 29-year-old man with anabolic steroids abuse presented with an acute abdominal pain. CT scan showed 11-cm mass in the right liver. Patient is referred for surgical treatment. Operation begins with liver mobilization. Glissonian approach of the right pedicle is achieved with two small incisions made in specific anatomic landmarks around hilar plate. Clamp is introduced through these incisions to occlude right Glissonian pedicle. The vascular clamp is then replaced by an endoscopic vascular stapling HPB 2015, 17 (Suppl. 1), 1–81

71

device, the right liver ischemic delineation is confirmed, and the stapler is fired. All of these steps are performed without the Pringle maneuver and without hand assistance. Liver transection and vascular control of the right hepatic vein are accomplished with as usual. The specimen is extracted through suprapubic incision. Results: Operative time was 160 minutes with minimum blood loss. Recovery was uneventful and patient was discharged on the third postoperative day. Final pathology showed liver cell adenoma with no signs of malignancy and free margins. Conclusion: Laparoscopic Glissonian approach for right hepatectomy is safe, fast and can be accomplished with proper technique.

V-A.02 CAVO-ATRIAL THROMBECTOMY COMBINED WITH RIGHT HEPATECTOMY EXTENDED TO SEGMENT IV FOR VASCULAR INVASION FROM HEPATOCELLULAR CARCINOMA THROUGH AN ABDOMINAL, TRANSDIAPHRAGMATIC AND INTRAPERICARDIAC APPROACH E. Vicente1,2,3, Y. Quijano1,2,3, B. Ielpo1,2,3, H. Duran1,2,3, I. Fabra1,2,3, E. Diaz1,2,3, S. Olivares1,2,3, R. Caruso1,2,3, A. Prestera1,2,3, M. De Luca1,2,3, J. Maupoey1,2,3, E. Vicente1,2,3 1 Sanchinarro University Hospital., Madrid, MADRID; 2 Clara Campal Oncological Center, Madrid, MADRID; 3 San Pablo University. CEU, Madrid, MADRID Vascular invasion of supra-hepatic veins is a major complication of primary liver tumours. The tumoral thrombus, when extended to the vena cava and right atrium, may produce occlusion of the tricuspid valve or pulmonary embolism with sudden cardiac death. The presence of macroscopic vascular infiltration represents an advanced stage of the tumour contraindicating liver transplantation, thus liver resection with thrombectomy is the only therapeutic option in this setting despite the concerns of postoperative liver failure and the dismal results at distance A 62-year-old female without chronic liver disease was referred to our Hospital diagnosed of the tumour located in the right hemi-liver with infiltration of the right hepatic veins and a tumour thrombus extension to the retrohepatic and suprahepatic inferior vena cava and right atrium. Routine blood tests revealed normal hematological results and liver biochemistry function. A computed tomography (CT) scan, Nuclear magnetic resonance (MRI) demonstrated the tliver tumor and the thrombus extending into retrohepatic, supraphrenic-intrapericardial IVC and right atrium. Right portal vein embolization was performed before the surgical resection The video shows the preoperative diagnostic procedures to evaluate the location of the tumor, the vascular invasion and extent of thrombus as well as the surgical procedure. Through a bilateral subcostal laparotomy, right hepatectomy extended to segment IV was performed. The transabdominal and transpericardial approach to the intrapericardial IVC and right atrium avoided median sternotomy. Intravascular trombectomy with intraoperative transesophageal echocardiogram was also performed. Two years after the operation, the patient remains alive and well without evidence of tumoral recurrence. © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

72

Abstracts

V-A.03 EXTENDED LEFT HEPATECTOMY WITH COMPLEX PORTAL VEIN RECONSTRUCTION AND IN SITU COLD PERFUSION OF THE LIVER FOR HILAR CHOLANGIOCARCINOMA G. Sapisochin, R. Smoot, J. Qu, A. Fung, J. B. Conneely, P. Kelly, G. Tait, L. Hotoyan, P. D. Greig, I. D. McGilvray Department Of Surgery. Toronto General Hospital. University Of Toronto, Toronto, ONTARIO Tumors of the biliary confluence can pose significant technical challenges when attempting to safely achieve negative margins. Major vascular involvement precludes standard extended resection, often reducing therapeutic options to transplantation or non-resective modalities. We present a case of a left-dominant Klatskin tumor in a young patient, involving the middle hepatic vein and abutting the portal vein bifurcation. Significantly, the portal inflow to the Right Posterior Section comprised of segment VI and accessory segment VI/VII portal vein branches, which would require separate reconstruction if resection was to be considered. We performed and extended left hepatectomy with resection of the portal bifurcation and reconstruction of both branches of the portal vein. In order to protect the remnant liver during the planned vascular reconstruction, in-situ cold perfusion of the liver with preservation solution was utilised. The accompanying video is comprised of real-time operative video. This complex case illustrates the facility of employing advanced transplantation techniques to expand therapeutic options in the oncology setting.

V-A.04 TRANSTHORACIC PORT PLACEMENT INCREASES SAFETY OF TOTAL LAPAROSCOPIC POSTERIOR SECTIONECTOMY L. Schwarz1, T. A. Aloia1, C. Eng2, G. J. Chang1, J. Vauthey1, C. H. Conrad1 1 UT MD Anderson Cancer Center, Department Of Surgical Oncology, Houston, TEXAS; 2UT MD Anderson Cancer Center, Department Of GI Medical Oncology, Houston, TEXAS Background: An anatomic posterior sectionectomy is infrequently performed due to the challenges of controlling the right posterior portal pedicle (RPPP) while preserving the anterior pedicle (RAPP), difficulty of visualizing the drainage of the right hepatic vein into the IVC, and the potential for significant blood loss during the caval and hepatovenous dissection. Patient: A 62-year-old woman with 3 liver metastases to SVI and SVII from sigmoid colon cancer underwent 5 cycles of neoadjuvant chemotherapy with FOLFOX and 4 of bevacizumab with a type I response. She underwent a “Primary First” robotic low anterior rectosigmoid resection followed by a laparoscopic posterior sectionectomy. Technique: The patient was placed in a Modified French Position. A transthoracic trocar was placed for optimal laparoscopic visualization and access of the superior retrohepatic IVC and drainage of the right hepatic vein into IVC. Intraoperative ultrasound was crucial to assess tumor © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

location, define transection plane and preserve flow to RAPP before division of RPPP. The parenchymal transection follows an oblique angle and exposes the right hepatic vein. Conclusion: Transthoracic port placement augments the safety of the dissection along the IVC inferiorly and the right hepatic vein superiorly due to direct visualization. Also, it provides a direct instrument-to-target axis without the typical fulcrum of dissecting the postero/superior liver. Laparoscopic ultrasound is critical to confirm preserved flow to the RPPP and guide the parenchymal transection. Liver volumetry should be obtained prior to surgery to determine adequate future liver remnant if conversion to a right lobectomy becomes necessary.

V-A.05 TOTALLY LAPAROSCOPIC TWO STAGE HEPATECTECTOMY WITH PORTAL VEIN EMBOLIZATION W. Kuo1,3, C. U. Corvera1,2 University Of California, San Francisco, San Francisco, CA; 2VA Medical Center San Francisco, San Francisco, CA; 3Chiayi Christian Hospital, Chia-yi City, TAIWAN

1

54 year-old woman with synchronous metastatic rectal cancer to the liver. Imaging showed several hepatic tumors distributed in the both lobes of the liver; R > L. A two stage procedure with right PVE was planned. This video shows a limited partial left hepatectomy that was followed on POD #1 with an immediate right PVE. Six weeks later, a right hepatectomy was completed. Stage 1. Using three 5 mm ports and a 12 mm periumbilical port, a partial segment # 2 liver resection and wedge resection of segment #3 was done using the bipolar cautery device and vascular stapler. Stage 2. Six weeks after PVE, she was returned to the operating room for R hepatectomy. The hepatic veins were exposed. Hilar dissection began by exposing the RHA lying posterior to the common hepatic duct. It was clipped and divided. The main PV was dissected until the bifurcation was identified. A caudate hepatotomy was done to allow isolation and division of the main RPV by a single firing of a stapler. The line of demarcation was marked using electrocautery and the parenchymal transection was started. As the liver was opened, the right bile duct was transected intrahepatically. Coursing branches of the middle and right hepatic veins were divided intrahepatically using a stapler. The parenchymal transection plane was aligned along the anterior surface IVC until the liver was completely divided. The RHV was identified and divided and the right hemi-liver was mobilized from diaphragm. The specimen was extracted via a low transverse incision.

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

SUNDAY, MARCH 15, 2015, 7:30AM–8:30AM VIDEO B – PANCREAS/BILIARY V-B.01 INTRAOPERATIVE AIR LEAK TEST PREVENTS POST-HEPATECTOMY BILE LEAK R. W. Day, W. R. Burns, C. Conrad, J. Vauthey, T. A. Aloia Department Of Surgical Oncology, The University Of Texas MD Anderson Cancer Center, Houston, TEXAS Background: After major liver resection, the incidence of bile leak is reported at 4–10%. We have previously reported that an intraoperative Air Leak Test (ALT) reduces the rate of bile leak by five-fold with no additional morbidity or mortality. Purpose: Demonstrate the steps of an intraoperative ALT following major liver resection. Results: A 29 year-old man presented with T3N1M1 rectal cancer and synchronous liver metastases. There were two right liver metastases, one at the right portal junction necessitating right hepatectomy. After neoadjuvant FOLFOX and bevacizumab chemotherapy, chemoradiotherapy to his primary rectal cancer, and abdominoperineal resection, he presented for liver resection. His liver metastases had radiographically responded with CEA decrease from 8.5 to 2.3 ng/ mL. The functional liver remnant volume was calculated to be 31%. The video briefly describes key components of the right hepatectomy including extrahepatic vascular inflow control, two-surgeon technique for hepatic parenchymal transection, and intrahepatic stapled division of the right bile duct. It then details the sequential steps of the ALT including cannulation of the cystic duct, Doppler of intrahepatic vascular flow, manual occlusion of the bile duct, injection of air under ultrasound visualization to confirm patency of the biliary confluence, and repeated injection of air to identify and oversew occult open bile ducts. At the completion of the ALT, the patient had no air leaks and no drain was placed. He experienced an uneventful recovery. Conclusion: For patients undergoing major hepatectomy, intraoperative ALT is a rapid, safe, and inexpensive means to decrease the incidence of post-operative bile leak.

V-B.02 MINIMALLY INVASIVE PARTIAL RIGHT HEPATECTOMY WITH TRANSDIAPHRAGMATIC RIGHT LOWER LOBE LUNG RESECTION FOR COLORECTAL METASTASES R. W. Day1, R. J. Mehran2, N. De Rosa1, C. Conrad1, T. L. Moon3, J. Vauthey1, T. A. Aloia1 1 Department Of Surgical Oncology, The University Of Texas MD Anderson Cancer Center, Houston, TEXAS; 2 Department Of Thoracic And Cardiovascular Surgery, The University Of Texas MD Anderson, Houston, TEXAS; 3 Department Of Anesthesiology And Perioperative Medicine, The University Of Texas MD Anderson, Houston, TEXAS

73

both sites of metastases are resected. The main obstacle to simultaneous resection is the additional morbidity of a thoracic incision. We have recently published an experience with open hepatectomy and transdiaphragmatic lung resection, demonstrating the feasibility and safety of this approach. Purpose: In an effort to further improve outcomes we investigated the ability to perform the simultaneous procedure using minimally invasive techniques. Results: A 40 year-old woman with pT1N0 colon cancer status post low anterior resection represented with an elevated carcinoembryonic antigen (CEA = 4.5 ng/mL). Imaging demonstrated three liver lesions, biopsy positive for metastasis and one right lower lobe pulmonary lesion. Preoperatively, she received three cycles of FOLFOX/ bevacizumab treatment and one cycle of FOLFOX alone with radiological and biochemical response (CEA = 2.5 ng/ mL). In this video we demonstrate the key components of her hand-assisted, laparoscopic simultaneous multifocal liver and lung tumor resection including the use of a rubber band retraction technique, resection of a disappeared 3 mm lesion, and the transdiaphragmatic lung resection with EndoStitch™ diaphragm closure. Final pathology confirmed margin negative resection of all 4 tumors. The patient was discharged from the hospital on POD #4 with no medical or surgical complications and a rapid return to normal function, as measured by a validated quality of life assessment tool. Conclusion: In carefully selected patients with metastatic colorectal cancer, minimally invasive transdiaphragmatic approaches to lung resection simultaneous with liver resection are technically feasible and safe to perform.

V-B.03 LAPAROSCOPIC MAJOR ARTERIAL RESECTION DURING PANCREATICODUODENECTOMY B. Franssen, M. Kendrick Mayo Clinic, Rochester, MN With recent improvements in multimodality treatment for borderline and locally advanced pancreatic cancer, major vascular resection during pancreatic surgery has become more frequent. Although portal vein and SMV resection in both open and laparoscopic pancreaticoduodenectomy (PD) has been previously described, the feasibility of major arterial resection in laparoscopic pancreatectomy has not been established. While major arterial resection remains controversial, it is performed in our center in very select patients in order to achieve an R0 resection during PD. The aim of this video presentation is to demonstrate technical aspects and plausibility of a common hepatic artery resection and reconstruction in combination with a portal vein resection in a patient undergoing laparoscopic PD with a radiographically stable ductal adenocarcinoma diagnosed two years prior to surgery.

Background: Patients with metastatic colorectal disease to the liver and lungs have improved long-term outcomes when HPB 2015, 17 (Suppl. 1), 1–81

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

74

Abstracts

V-B.04 MODIFIED ROBOTIC APPLEBY PROCEDURE WITH LEFT NEPHRECTOMY AND ADRENALECTOMY FOR LOCALLY ADVANCED DISTAL PANCREATIC CANCER M. Radomski, A. Zureikat, H. Zeh, M. Hogg University Of Pittsburgh, Pittsburgh, PA Patients presenting with Stage III pancreatic adenocarcinoma involving the distal pancreas and celiac axis have historically been treated by nonsurgical means. However, almost a third of these patients will die without metastatic disease suggesting that aggressive surgical resection involving visceral vascular structures may offer a survival advantage. Appleby first described resection of the celiac axis for gastric cancer. We have extended this technique to include stage III distal pancreatic adenocarcinoma involving the celiac axis without evidence of metastatic disease via a combined laparoscopic and robotic approach. We describe a robotic distal pancreatectomy, splenectomy, left nephrectomy, and left adrenalectomy in a 65 year old female patient with pancreatic adenocarcinoma involving the distal pancreas without evidence of distant metastatic disease. She previously underwent six cycles neoadjuvant chemotherapy with gemcitabine and abraxane in which she had a good response with normalizing CA19-9 values. Her postoperative course was complicated by gastric ischemia which was treated with short course of total parenteral nutrition, and a portal vein thrombus treated with subcutaneous low molecular weight heparin. Her pathology showed a 4.0 cm moderately differentiated adenocarcinoma with negative margins and twenty-six lymph nodes negative for disease. She was discharged home on postoperative day 25 and started on adjuvant chemotherapy with gemcitabine and abraxane on postoperative day 76.

V-B.05 ROBOTIC PANCREATICODUODENECTOMY WITH PORTAL VEIN RESECTION AND PATCH VENOPLASTY M. Girgis, M. Hogg, H. Zeh, A. Zureikat University Of Pittsburgh School Of Medicine, Pittsburgh, PA Introduction: Robotic pancreaticoduodenectomy is safe and feasible for resectable pancreatic head adenocarcinoma. Vascular resection during pancreaticoduodenectomy for borderline resectable PDAC may be associated with slightly increased morbidity but provides good oncologic outcomes. We describe a robotic pancreaticoduodenectomy with planned portal vein resection and patch venoplasty in a patient with borderline resectable pancreas cancer. Methods: This patient’s operation was recorded by the Da Vinci Surgical System. The file footage was compiled and edited. The patient’s consent was obtained. Results: This is the case of a 57-year-old patient with locally advanced pancreas cancer. He underwent 6 cycles of neoadjuvant chemotherapy. Restaging CT scan showed down-staging of the tumor to borderline resectable status. The procedure was completed robotically in 452 minutes with an estimated blood loss of 900 ml. A large 4 cm partial © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

portal vein resection was performed with reconstruction using bovine pericardium in the form of a patch venoplasty. Final pathology revealed a 2.9 cm tumor with 4/37 nodes positive. He recovered uneventfully from the surgery and was discharged from the hospital on postoperative day 6. Within 45 days of surgery, he had initiated adjuvant chemotherapy. Conclusions: We describe a robotic portal vein resection and reconstruction with patch venoplasty for borderline resectable PDAC. We were able to complete the procedure safely and in a timely manner with minimal morbidity.

SUNDAY, MARCH 15, 2015, 8:30AM–10:30AM LONG ORAL J – LIVER PERIOPERATIVE/TECHNIQUES LO-J.01 ELEVATED INR INDEPENDENTLY PREDICTS MAJOR COMPLICATIONS FOLLOWING MAJOR HEPATECTOMY J. N. Leal, T. P. Kingham, P. J. Allen, R. P. DeMatteo, W. R. Jarnagin, M. I. D’Angelica Memorial Sloan Kettering Cancer Center, New York, NY Background: The impact of major hepatectomy on hemostasis is poorly characterized and accuracy of INR as an indicator of coagulation status in the peri-operative setting is unknown. This study aimed to characterize posthepatectomy hemostatic profiles and to evaluate subsequent impact on clinical outcomes. Methods: All patients undergoing major hepatectomy between 2001–2012 were identified from a prospective database. Clinicopathologic and blood product transfusion data were extracted. Post-operative hemostatic profiles were analyzed with respect to plasma transfusion and major complications. Results: 961 patients were identified. The mean number of Couinaud segments resected was 4.3(+/−0.85).Postoperatively peak INR was 1.50 +/− 0.26, platelet nadir 166.1 +/− 61.7 K/mcL, and hemoglobin nadir 9.5 +/− 1.34 g/ dL. INR ≥1.8 occurred in 152 (16%) patients. Major bleeding complication rates in the INR ≥1.8 group were not different from INR <1.8 group (0.7% vs. 0.8%, p = 0.90). Alternatively, major complication rates were significantly higher (37% INR ≥1.8 vs. 17% INR <1.8, p < 0.0001). Among the INR ≥1.8 group, plasma was transfused in 66% of cases and was not associated with a significant change in major complication rates (40% in peak INR ≥1.8 with plasma vs. 30% in peak INR ≥1.8 without plasma, p = 0.30). On multivariate analysis, including standard variables, peak INR ≥1.8 was independently associated with major complications (OR 2.48, 95% CI 1.64–3.74, p < 0.0001) while FFP transfusion was not (OR 1.35 95% CI 0.90–2.05, p = 0.152). Conclusions: Elevated INR following major hepatectomy does not increase the risk of post operative hemorrhage but rather portends increased overall morbidity that is not mitigated by FFP transfusion. HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

LO-J.02 EARLY INCREASES IN POSTOPERATIVE SERUM PHOSPHOROUS AND CREATININE LEVELS ARE ASSOCIATED WITH MORTALITY FOLLOWING MAJOR HEPATECTOMY G. Herbert1, W. Jarnagin1, R. DeMatteo1, P. Allen1, M. D’Angelica1, M. Gonen1, K. Prussing2, T. P. Kingham1 1 Memorial Sloan Kettering Cancer Center, New York, NY; 2 University Of Minnesota Medical School, Duluth, MN Introduction: Post-hepatectomy liver failure (PHLF) is defined by the International Study Group of Liver Surgery as a bilirubin >2.9 mg/dL and INR >1.7 on or after postoperative day 5 (the 50/50 criteria), and greatly increases the risk of death after hepatectomy. Other than pre-existing liver disease and small future liver remnant, few patient factors or early post-operative indicators identify patients at elevated risk of PHLF. Methods: We reviewed demographics, comorbidities, operative procedures, and post-operative laboratory trends of patients undergoing major hepatectomy (>= 3 Couinaud segments) for malignancy from 1998–2013 at our institution. These factors were compared between patients dying within 90 days of surgery, survivors meeting the 50/50 criteria, and all remaining survivors. Results: 1536 patients underwent major hepatectomy during the time period. The majority were performed for metastatic colorectal cancer, with an average of 4.24 resected segments. 49 (3.2%) died within 90 days of surgery. 47 (3.1%) patients met ISGLS criteria for PHLF, of whom 15 died within 90 days of surgery. Operative blood loss was twice as high in patients dying within 90 days as compared to survivors. In spite of greater perioperative resuscitation (achieving equivalent urine output) when compared to survivors, non-survivors had significantly higher creatinine and phosphorous levels on POD 1. Predicted residual liver volume did not differ between patients dying within 90 days and survivors who met the 50/50 criteria (32.1% vs 31.1%). Conclusions: An increase in both creatinine and phosphorous on POD 1 is an early indicator of patients at risk for PHLF.

HPB 2015, 17 (Suppl. 1), 1–81

75

LO-J.03 EFFECTS OF BLOOD TRANSFUSION ON PERIOPERATIVE AND LONG-TERM OUTCOMES AFTER MAJOR HEPATECTOMY FOR METASTATIC COLORECTAL CANCER: A MULTI-INSTITUTIONAL STUDY OF 456 PATIENTS L. M. Postlewait1, M. H. Squires1, D. A. Kooby1, S. M. Weber2, C. R. Scoggins3, K. Cardona1, C. S. Cho2, R. C. Martin3, E. Winslow2, S. K. Maithel1 1 Division Of Surgical Oncology, Emory University, Atlanta, GA; 2Division Of Surgical Oncology, University Of Wisconsin, Madison, WI; 3Division Of Surgical Oncology, University Of Louisville, Louisville, KY Background: Data on the prognostic implications of blood transfusion at the time of major hepatectomy (≥3 segments) for colorectal cancer metastases are conflicting. Our aim was to assess the association of perioperative transfusion with postoperative complications and disease-specific survival (DSS). Methods: Patients who underwent major hepatectomy for metastatic colorectal cancer from 2000–2010 at three US academic institutions were included. 30-day mortalities and patients who died of unknown cause were excluded from survival analyses. Transfusion was analyzed based on timing and volume of transfusion. Results: Of 456 patients, 140 (30.7%) received blood transfusions. Perioperative transfusion was associated with extended hepatectomy (40.8%vs27.9%; p = 0.020), increased tumor size (5.7 vs 4.2 cm; p < 0.001), and increased estimated blood loss (917 vs 390 mL; p < 0.001). There were no differences between patients with or without transfusion regarding age, ASA class, margin, number of lesions, cirrhosis, or lymphovascular invasion. On multivariate analysis, perioperative transfusion was associated with major complications (HR3.14; 95%CI: 1.83–5.39; p < 0.001) and 90-day readmission (HR2.14; 95%CI: 1.18–3.89; p = 0.012). 388 patients were included in survival analyses; median follow-up was 38.8 mos. Perioperative transfusion was not associated with DSS; however, patients who received blood postoperatively had decreased DSS (37.4 vs 42.7 mos;p = 0.044). Increasing volume of transfusion at any time was associated with shortened DSS (Perioperative: 0 units: 42.1 vs 1–2 units: 39.6 vs ≥3 units: 37.4 mo, p = 0.046; © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

76

Abstracts

Postoperative: 0 units: 42.7 vs 1–2 units: 38.6 vs ≥3 units: 27.2 mo, p = 0.031; Figure). Transfusion, regardless of timing and volume, was not independently associated with worsened DSS on multivariate analyses. Conclusion: Perioperative blood transfusion with major hepatectomy for colorectal cancer metastases is independently associated with increased complications and readmission but not reduced disease-specific survival. Judicious use of blood transfusion in the perioperative period is warranted.

Final pathologic diagnoses were SLC (n = 73), BCA (n = 15), BCAC (n = 3) and other primary liver cancers (n = 3). On multivariate analysis, preoperative imaging features such as solitary lesion, septations and solid component were associated with a final diagnosis of BCT (p < 0.05; Table). However, one-third of SLC presented with septations or solitary lesions. Five patients (23.8%) were initially misdiagnosed as SLC and 24 patients (32.9%) were misdiagnosed as BCT of which 15 patients unnecessarily underwent resection. Fenestration and frozen section (FS) was performed in 47 patients (diagnostic accuracy = 100%), impacting surgical management in 12 patients (12.7%) by avoiding (n = 8) or mandating (n = 4) liver resection. No patients with BCT (n = 8), submitted to fenestration and FS followed by resection, developed recurrence. Conclusion: Worrisome imaging features are associated with a higher risk of BCT. However, one-third of patients with SLC present with such imaging characteristics. In the absence of a strong suspicion of malignancy, fenestration and FS should be considered prior to complete resection.

LO-J.04 THE ACCURACY OF PREOPERATIVE IMAGING AND INTRAOPERATIVE FROZEN SECTION IN THE MANAGEMENT OF HEPATIC CYSTS A. Doussot1, B. Groot Koerkamp1, P. J. Allen1, R. P. De Matteo1, J. Shia2, T. P. Kingham1, W. R. Jarnagin1, S. R. Gerst3, M. I. D’Angelica1 1 Department Of Surgery Memorial Sloan Kettering Cancer Center, New York, NY; 2Department Of Pathology Memorial Sloan Kettering Cancer Center, New York, NY; 3 Department Of Radiology Memorial Sloan Kettering Cancer Center, New York, NY Introduction: Biliary cystic tumors (BCT) (biliary cystadenoma (BCA) and cystadenocarcinoma (BCAC)) warrant complete resection. Simple liver cysts (SLC), however, require only fenestration if symptomatic. Distinguishing between BCT and SLC with preoperative imaging is not well studied. Methods: All patients undergoing surgery for a preoperative diagnosis of SLC or BCT were included. Perioperative clinical and imaging features from radiology reports were analyzed for their ability to predict the final diagnosis. Results: Ninety four patients underwent fenestration (n = 53), enucleation (n = 6) or liver resection (n = 35) for suspected SLC (n = 54), BCA (n = 32) and BCAC (n = 8). © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

LO-J.05 THE IMPACT OF PORTAL PEDICLE CLAMPING ON SURVIVAL FROM COLORECTAL LIVER METASTASES IN THE CONTEMPORARY ERA OF LIVER RESECTION: A MATCHED COHORT STUDY M. E. Tsang1, P. J. Karanicolas1,2, R. Habashi1, E. Cheng1, S. Hanna1,2, N. G. Coburn1,2, C. H. Law1,2, J. Hallet2 1 Division Of General Surgery, University Of Toronto, Toronto, ON; 2Sunnybrook Health Sciences Centre – Odette Cancer Centre, Toronto, ON Introduction: Portal pedicle clamping (PPC) reduces bleeding during hepatectomy, but may impact micro-metastases’ HPB 2015, 17 (Suppl. 1), 1–81

Abstracts growth through ischemia-reperfusion injury. We sought to examine the association between PPC and long-term survival following hepatectomy for colorectal liver metastases (CRLM). Methods: We conducted a matched cohort study using our prospective hepatectomy database to identify all patients undergoing hepatectomy for CRLM from 2003 to 2013. Cohorts were selected based on use of PPC, with 1 : 1 matching for age (5-year increments), time period (2003–2007 vs. 2007–2013), and Clinical Risk Score (0 to 5 scale). Primary outcome was overall survival (OS). Conditional logistic and Cox regression analyses determined odds ratios (OR) and hazard ratios (HR). Results: Of 481 hepatectomies for CRLM, 187 (39%) patients underwent PPC. 110 pairs of patients were matched in the cohorts, and the remainder excluded. Peri-operative chemotherapy (p = 0.183), major hepatectomy (>= 3 segments) (p = 0.345) or resection status (R0 vs. R1–2) (p = 0.132) did not differ. 30-day major morbidity (OR 0.73; p = 0.332) and mortality (OR 2.4; p = 0.100) were not significantly associated with PPC. Median follow-up was 35 (range: 0–130) months. When adjusting for extent of resection, blood loss, and operative time, no significant difference was observed in OS (HR 1.76; p = 0.129) for PPC, with 5-year OS of 59.2% (95%CI: 54.0–59.2%) for PPC and 62.3% (95%CI: 61.8–67.3%) without PPC (log-rank p = 0.415). Excluding 90-day deaths did not substantially alter the results (log-rank p = 0.930). Conclusions: PPC was not associated with a significant difference in OS in patients undergoing hepatectomy for CRLM. It does not appear to adversely affect oncologic outcomes.

HPB 2015, 17 (Suppl. 1), 1–81

77

LO-J.06 HEPATO-PANCREATECTOMY: HOW MORBID? RESULTS FROM THE NATIONAL SURGICAL QUALITY IMPROVEMENT PROGRAM T. B. Tran, M. M. Dua, D. A. Spain, B. C. Visser, J. A. Norton, G. A. Poultsides Stanford University School Of Medicine, Stanford, CA Background: Simultaneous resection of both the liver and pancreas carries significant complexity. The objective of this study is to investigate perioperative outcomes following synchronous hepatectomy and pancreatectomy (SHP). Methods: The American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database was queried to identify patients who underwent SHP. Resections were defined as follows: “
© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

78

Abstracts

LO-J.07 PORTAL VEIN EMBOLISATION AND APPLICATION OF HAEMATOPOIETIC STEM CELLS IN PRIMARILY NON-RESECTABLE COLORECTAL LIVER METASTASES FOR INSUFFICIENT FUTURE LIVER REMNANT VOLUME V. Treska1,2 1 University Hospital, Pilsen, PILSEN; 2School Of Medicine, Pilsen, PILSEN Background: Insufficient future liver remnant volume (FLRV) is the cause of the low resectability of colorectal liver metastases (CLMs).Portal vein embolisation (PVE) with the application of autologous haematopoietic stem cells (HSCs) is option for increasing resectability of CLMs. The aim of our study was to compare the effect of PVE only and PVE with application of HSCs in patients with primarily non-resectable CLMs due to insufficient FLRV. Methods: In prospective, randomised study PVE with the application of HSCs was used in 14 patients (GI). The control group (GII) consisted of 14 patients in whom only PVE was performed. We evaluated FLRV growth, CLM volume growth, median survival and progression-free survival (PFS). Results: In all GI patients sufficient FLRV growth occurred within three weeks. In the first two weeks, FLRV increased in most of the patients (p < 0.006). In 13 (92.9%) of the GII patients, optimum FLVR growth was observed within three weeks following PVE (p < 0.002). More rapid FLVR growth was observed in the GI patients (p < 0.01). CLM volume was significantly increased in the GII (p < 0.0005) and also GI (p < 0.006) at the time of liver resection. There was no significant difference in the growth of the CLM volume between the groups (p < 0.17). The median survival of the GI and GII patients was 7.4 and 6.8 months and the two-year PFS was 27 and 22% (n.s.), respectively. Conclusion: We can conclude that PVE with HSC application is a promising method for effectively stimulating FLVR growth in primarily non-resectable CLMs.

LO-J.08 ENHANCED ULTRASOUND WITH NAVIGATION LEADS TO IMPROVED LIVER LESION IDENTIFICATION AND NEEDLE PLACEMENT R. C. Martin University Of Louisville, Louisville, KY Background: The aim of this study was to evaluate whether 3-dimensional (3D) enhanced ultrasound could increase the accuracy and efficiacy for liver tumor identification and needle placement. Methods: A prospective study of 30 surgeons of various training level where evaluated lesion identification success and accuracy of needle placement. All surgeons were evaluated for time(seconds) to identify the liver lesions and placement of needles after review of 3 phase CT scan of the liver and using standard B-mode ultrasound and then enhanced 3D ultrasound (E-3DUS). Results: Participants included 10 HPB surgeons, 5 Surgical Fellows, 10 PGY-4&5 Surgical Residents and 5 PGY-3 resi© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

dents. Liver lesions were correctly identified in 73% of the cases using B-mode ultrasound alone and 100% in E-3DUS. The Mean time to identification is B-mode was 51.9 seconds (SD +/− 37.1), which was significantly longer than with E-3DUS (time 17.9 seconds, SD +/− 10.7)(p = 0.002). There was significant improvement in time to lesion identification using E-3DUS across all training levels (Figure, *-P < 0.002). There was also a significant reduction in time for accurate needle placement across all training levels (mean reduction of 60%, within enhanced accuracy (p = 0.001) Conclusions: E-3DUS significantly enhances lesion identification regardless of size and enhances needle accuracy for all surgeons. This adjunctive system should be considered for both training as well as for all complex liver tumor ablations.

MO-A.02 BRANCHED DNA ENHANCED ALBUMIN RNA IN SITU HYBRIDIZATION PROVIDES A DEFINITIVE DIAGNOSIS FOR INTRAHEPATIC CHOLANGIOCARCINOMA C. R. Ferrone, D. T. Ting, M. Shahid, I. T. Konstantinidis, L. Goyal, N. Bardeesey, D. Borger, K. D. Lillemoe, A. X. Zhu, M. N. Rivera, V. Deshpande Massachusetts General Hospital, Boston, MA USA Purpose: Intrahepatic cholangiocarcinoma (ICC) is often a diagnosis of exclusion after patients have undergone numerous studies and procedures. Currently there is no histopathologic or immunohistochemical marker for ICC. Albumin expression is restricted to the liver. Since liver parenchymal cells are derived from a common embryonic progenitor, the aim of our study was to determine if albumin is a potential biomarker for ICC utilizing a novel and highly sensitive RNA in situ hybridization (ISH) platform. Methods: Modified branched DNA probes were developed for albumin RNA ISH. A total of 467 patient samples of primary and metastatic lesions were evaluated. Results: Of the 467 samples, 43 were ICCs, 42 were hepatocellular carcinomas and 332 were non-hepatic carcinomas including tumors arising from the perihilar region, bile duct, pancreas, stomach, esophagus, colon, lung, breast, ovary, endometrium, kidney, and urinary bladder. Albumin RNA ISH was highly sensitive for cancers of liver origin, staining positive in 42 of 43 ICCs (99%) and 42 HCCs (100%). Perihilar and distal bile duct carcinomas as well as carcinomas arising at other sites were negative for albumin. Notably, 6 of 27 (22%) intrahepatic tumors previously diagnosed as carcinomas of undetermined origin were positive for albumin. Conclusions: Albumin RNA ISH is a sensitive and highly specific diagnostic tool to distinguish ICC from metastatic adenocarcinoma to the liver or carcinoma of unknown origin. The specificity of albumin RNA ish could replace the extensive diagnostic work up patients undergo to confirm the diagnosis.

HPB 2015, 17 (Suppl. 1), 1–81

Abstracts

MO-A.03 HIGH HLA CLASS I ANTIGEN EXPRESSION IN COMBINATION WITH LOW PD-L1 EXPRESSION AS A FAVOURABLE PROGNOSTIC BIOMARKER IN INTRAHEPATIC CHOLANGIOCARCINOMA 1

1

2

1

F. Sabbatino , V. Villani , J. H. Yearley , L. Cai , V. Deshpande3, I. T. Konstantinidis1, S. P. Nota4, Y. Wang1, A. X. Zhu5, L. Goyal5, D. T. Ting5, N. M. El-Bardeesy5, T. S. Hong6, K. K. Tanabe7, C. Moon2, S. Ferrone1,4, K. D. Lillemoe1, C. R. Ferrone1 1 Department Of Surgery, Massachusetts General Hospital, Boston, MA; 2Merck Research Laboratories, Palo Alto, CA; 3Department Of Pathology, Massachusetts General Hospital, Boston, MA; 4Department Of Orthopedic Surgery, Massachusetts General Hospital, Boston, MA; 5 Cancer Center, Massachusetts General Hospital, Boston, MA; 6Department Of Radiation Oncology, Massachusetts General Hospital, Boston, MA; 7Department Of Surgical Oncology, Massachusetts General Hospital, Boston, MA Introduction: For immune checkpoint molecule-specific monoclonal-antibodies (mAbs) to be effective patients must be able to mount an immune response to their tumor. This requires a functional HLA class I antigen-processing machinery in the targeted tumor cells. We evaluated i) lymphocyte infiltrate as a measure of a patient’s immune response to their intrahepatic cholangiocarcinoma (ICC) ii) HLA class I antigen expression iii) expression of the checkpoint molecules programmed cell death 1 (PD-1) and its primary ligand programmed cell death ligand 1 (PD-L1). Methods: Clinicopathologic data for 30 patients undergoing resection for ICC was collected. Tumors were immunohistochemically stained with CD8-, CD4-, HLA-, PD-1- and PD-L1-specific mAbs. Results: Median age was 64 years, 47% had stage II disease and median follow-up was 28.1 months. All tumors had lymphocyte infiltrates. CD8s in the fibrous septa (FS) between tumor lobules (TL) was higher than within TL (mean 104 vs 12, P < 0.0001), but not for CD4s. CD8s and CD4s in the FS and TL correlated with HLA expression (P < 0.05). CD8s were decreased, while CD4s were increased if HLA was down-regulated. PD-1 expression was present in 83% of cases and highest on T cells in the FS. CD8s/CD4s, HLA, and PD-1/PD-L1 expression in the tumor microenvironment did not correlate with survival. In contrast high HLA expression and low PD-L1 expression was associated with a longer overall survival (P = 0.04). Conclusions: Increased PD-L1 expression and HLA class I antigen defects provide ICC with an escape mechanism from immune recognition. This provides rationale for implementing antibodies to checkpoint molecules for ICC.

79

MO-A.04 PIVOTAL ROLE OF INTRAPLATELET VEGF-A DURING ANGIOGENIC PHASE OF LIVER REGENERATION AFTER PARTIAL HEPATECTOMY IN HUMANS B. Aryal1, T. Shimizu2, J. Kadono1, A. Furoi3, M. Yamakuchi2, M. Inoue1, T. Komokata4, T. Hashiguchi2, Y. Imoto1 1 Cardiovascular And Gastroenterological Surgery,Kagoshima University, Kagoshima, KYUSHU Japan; 2Laboratory And Vascular Medicine, Kagoshima University, Kagoshima, KYUSHU Japan; 3Kirishima Medical Center, Kirishima, KYUSHU Japan; 4Kagoshima Medical Center, Kagoshima, KYUSHU Japan The ability of remnant liver to regenerate plays a decisive role in morbidity and mortality after liver resection. Molecular mechanism of liver regeneration (LR) has still remained as a major concern in modulating the strategies in order to improve the outcome after resection. Two phases of LR have been proposed lately; inductive phase marked by rapid proliferation of hepatocytes followed by a delayed angiogenic phase with proliferation of non-parenchymal cells. Several studies have already reported the beneficial role of platelet in LR. We introduce platelet as a central player in the angiogenic phase of LR after partial hepatectomy in human. Materials and Methods: Twenty patients with diagnosed hepatocellular carcinoma; eligible for hepatectomy were enrolled in the study. Serum, plasma and intra-platelet VEGF-A (vascular endothelial growth factor-A) along with the major mitogens were monitored both before and after one month of operation. Result: The major serum VEGF-A pool is attributed to the platelet, with plasma containing scarcer amount or no VEGF-A. The change in serum VEGF-A following liver resection strongly correlated with the altered platelet count. Serum VEGF-A was found to be significantly elevated after a month of partial hepatectomy; intra-platelet VEGF (IPVEGF) showed the similar trend whereas no significant difference was observed in the level of major mitogens including hepatocyte growth factor. The soluble VEGF receptors as important anti-angiogenic markers were down regulated during the on-going angiogenic phase of LR. Conclusion: IP-VEGF in angiogenic phase might clue to anovel therapeutic target to accelerate LR after hepatectomy.

MO-A.05 PROGNOSTIC FACTORS OF RECURRED HEPATOCELLULAR CARCINOMA AFTER PRIMARY LIVER RESECTION J. Jeong, K. Suh, K. Lee, N. Yi Seoul National University College Of Medicine, Seoul, SEOUL Background: Resection is the treatment of choice for hepatocellular carcinoma (HCC). However, recurrence is still a major problem after primary resection and prognostic factors after recurrence is not well known. So, we investigated prognostic factors of recurred HCC after primary resection to determine who had an aggressive tumor biology.

HPB 2015, 17 (Suppl. 1), 1–81

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

80

Abstracts

Methods: We reviewed the patients who underwent surgical resection for HCC between 2005 and 2011. Of a total 941 patients who underwent resection, 589 (62.5%) patients had recurrences. Among them, 141 patients with extrahepatic recurrences or positive margins were excluded. Finally, we analyzed 448 patients who had intrahepatic recurrences after curative resections. Results: The median duration between primary resection and recurrence was 9.0 (0.2–90.0) months. At recurrence, 353 (78.8%) patients had HCCs within Milan’s criteria. After recurrence, 5-year disease-free survival rate was 12.8% and 5-year survival rate was 66.1%. >5 cm of primary tumor size, <1 year of duration to recurrence, >200 of AFP level at recurrence, and beyond Milan’s criteria at recurrence were revealed as independent poor prognostic factors associated with disease-free survival. Whereas, only Milan’s criteria and CTP class at recurrence were identified as independent prognostic factors associated with overall survival. Conclusion: For recurred HCC after primary resection, factors at recurrence including primary tumor size are significantly associated with disease-free survival. Milan’s criteria and CTP class at recurrence are associated with overall survival. Therefore, it should be considered when deciding the second-line treatment option for recurred HCC after primary resection.

MO-B.01 GOING GREEN BEYOND THE BORDERS: EXPANDING THE IRE KILL ZONE WHILE CONSERVING ELECTRICITY S. C. Agle, Y. Li, C. A. Doughtie, C. R. Scoggins, R. C. Martin University Of Louisville, Louisville, KY USA Introduction: Irreversible electroporation (IRE) is an ablation technique in which short, high-voltage pulses are applied to tissue to irreversibly permeabilize the cell membrane resulting in apoptosis. The objective of this study was to determine if a secondary zone of reversible electroporation can be used to enhance tumor cell death. Methods: Five nude mice with orthotopic human pancreatic adenocarcinoma (PANC1) were in each group: gemcitabine (15 mg/kg) alone, IRE alone, gemcitabine plus IRE and a control group. IRE was set up in a 2-probe array around the tumor to produce 2,000 volts/cm at a rate of 90 pulses per minute for a total of 90 pulses. Pancreas tissue slides were measured for hypoxia via the hypoxia-inducible factor (HIF1a) and apoptosis via TUNEL staining. The IRE tissue was obtained from the lateral probe margin adjacent to the ablation zone and the gemcitabine only tissue was taken from the tumor edge. Results: The rate of hypoxia (Th area %) in the tissue treated with IRE alone or combined with gemcitabine were similar (45.2 ± 8.6 and 68.0 ± 15.3) and over 3 times lower (p < 0.1) than the groups receiving only gemcitabine or no treatment (218.8 ± 29.2 and 351.8 ± 28.4). IRE plus gemcitabine group had an apoptotic index of 22.6% ± 7.7 which was more than 4 times higher than the IRE alone group (4.6 ± 2.7, p < 0.01). © 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

Conclusion: In addition to tumor cell death within the irreversible ablation zone the use of IRE could potentially reduce local recurrence by allowing for increased drug administration to the cells on the periphery of the ablation zone.

MO-B.02 THE POST-SIR-SPHERES SURGERY STUDY (P4S): ANALYSIS OF OUTCOMES FOLLOWING HEPATIC RESECTION OR TRANSPLANTATION IN 101 PATIENTS PREVIOUSLY TREATED WITH SELECTIVE INTERNAL RADIATION THERAPY (SIRT) F. Pardo1, M. Schoen2, L. Rheun-Chuan3, D. M. Manas4, D. R. Jeyarajah5, G. Katsanos 6, G. A. Maleux7, B. Sangro8 1 Hepatobiliary, Pancreatic And Transplant Surgery, Pamplona, NAVARRA; 2Surgery, Kalrsruhe, BADEN-WÜRTTEMBERG; 3Interventional Radiology, Taipei, BEITOU DISTRICT,; 4Hepatobiliary And Transplant Surgery, Newcastle, TYNE AND WEAR; 5 Surgical Oncology, Dallas, TX; 6Hepatobiliary And Transplant Surgery, Brussels, BRUXELLES; 7Radiology, Leuven, FLEMISH BRABANT; 8Liver Unit, Pamplona, NAVARRA SIRT (or radioembolisation) is primarily used as palliative treatment for inoperable primary or metastatic liver tumours, and as bridge-to-liver transplantation in hepatocellular carcinoma (HCC). There have been reports of down-sizing to surgical resection/transplantation but there are no robust studies of safety outcomes. P4S is an international, multicentre, retrospective study to assess outcomes associated with liver resection or transplantation following SIRT using yttrium-90 resin microspheres (SIR-Spheres; Sirtex). Primary endpoints were peri-operative and 90-day postoperative morbidity and mortality. Analysis used standard statistical methods. Data were captured on SIRT, surgery (between 08/1998–05/2014) and follow-up on 101 patients with either primary liver cancer (HCC: 48.5%; cholangiocarcinoma: 6.9%) or secondary hepatic metastases from HPB 2015, 17 (Suppl. 1), 1–81

Abstracts colorectal (30.7%), neuroendocrine (6.9%) and other cancers (6.9%). Resection was performed in 72 patients (71.3%; 40.6% major, 30.7% minor) and liver transplantation in 29 (28.7%). Mean interval between first SIRT and first hepatic surgery was 8.7 months. Clavien-Dindo grade 3+ peri-/postoperative complications were: liver failure: 7 (6.9%); woundspecific: 4 (4.0%); cardiovascular: 0 (0%); pulmonary: 8 (7.9%); renal-specific: 2 (2.0%); other: 15 (14.9%). Cumulative 90-day all-cause mortality from first hepatic surgery was 4 (4.0%). These 4 cases were all trisectionectomies (colorectal: 3; cholangiocarcinoma: 1) and typically had ≥1 prior chemotherapy line, pre-surgical co-morbidities and suffered post-hepatectomy multi-organ failure including liver failure. Future liver remnant was targeted with SIRT in 1 of the 4 cases. The safety profile of post-SIRT resection and transplantation appears similar to that previously reported for hepatic surgery. No deaths appear to be directly related to SIRT.

MO-B.03 #PANCSM INITIAL EXPERIENCE WITH CREATION OF A TWITTER-BASED MONTHLY ONLINE PANCREATIC CANCER CHAT COMMUNITY N. J. Gusani1, L. Coker3, M. H. Katz5, D. Reidy-Lagunes4, P. Bloomston2 1 Program For Liver, Pancreas, & Foregut Tumors; Penn State Hershey Cancer Institute, Hershey, PENNSYLVANIA USA; 2Surgical Oncology; Ohio State University, Columbus, OH USA; 3Marketing & Communications; The Ohio State University Comprehensive Cancer Center, Columbus, OH USA; 4Division Of Gastrointestinal Oncology; Memorial Sloan-Kettering Cancer Center, New York, NY USA; 5Department Of Surgical Oncology; The University Of Texas MD Anderson Cancer, Houston, TX USA

81

services like Twitter often used for real-time communication in pre-arranged “chats” around a topic or disease. The authors created a monthly Twitter-based 1-hour chat (#PancSM) to help increase collaboration and interaction among patients, family/caregivers, advocates, health care providers, researchers, and charitable organizations. We report our initial experience. Methods: #PancSM chats are held on the first Thursday of each month, 9–10pm ET. A moderator introduces predetermined topics and discussion points, facilitating interaction with targeted questions. The chat progresses with freeform discussion among participants and is united and made searchable by use of the hashtag #PancSM in all tweets. Metrics for each chat session and participants were derived using online tools (Simply Measured, Followerwonk). Results: The first three monthly #PancSM chats were lively, with 125–262 tweets per session and 9–21 tweets per minute. The chats had a maximal reach of 32,608, with a maximal total exposure of 288,027 impressions. Tweeters participating in #PancSM chats initially included mostly physicians, but by the third chat, advocacy groups and patient caregivers were taking an important role. Conclusions: A Twitter-based monthly Pancreatic Cancer chat has the potential to reach large numbers of pancreatic cancer community members and to discuss issues important to many stakeholders. This forum is an innovative education medium to help improve understanding and patient outcomes in pancreatic cancer.

Background: Pancreatic Cancer (PancCa) is a highly lethal disease about which many myths and much nihilism exists. Social Media (SM) is a novel medium for communication and interaction across time and space, with microblogging

HPB 2015, 17 (Suppl. 1), 1–81

© 2015 The Authors HPB © 2015 Americas Hepato-Pancreato-Biliary Association

Copyright of HPB: The Official Journal of the International Hepato Pacreato Biliary Association is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.