Resection after neoadjuvant therapy for locally advanced, “unresectable” pancreatic cancer

Resection after neoadjuvant therapy for locally advanced, ‘‘unresectable’’ pancreatic cancer Oliver Strobel, MD,a Viktoria Berens,a Ulf Hinz, MSc,a Werner Hartwig, MD,a Thilo Hackert, MD,a Frank Bergmann, MD,b J€ urgen Debus, MD, PhD,c Dirk J€ ager, MD,d Markus W. B€ uchler, MD,a and a Jens Werner, MD, Heidelberg, Germany

Background. For pancreatic cancer, complete macroscopic resection in combination with chemotherapy is the only potentially curative treatment. Many patients present with locally advanced cancers deemed unresectable. We sought to assess the results of exploration after neoadjuvant therapy for locally advanced possibly unresectable pancreatic cancer. Methods. From a prospective database, all consecutive patients undergoing operation from October 2001 to December 2009 after neoadjuvant therapy for locally advanced pancreatic cancer were identified. Main criteria for ‘‘unresectability’’ were infiltration of the celiac axis or superior mesenteric artery. Resection rates, perioperative results, and survival were analyzed. Results. Of 257 patients, 199 (77.4%) had received neoadjuvant chemoradiation, and 58 (22.6%) chemotherapy only. Of 257 patients, 120 (46.7%) underwent successful resection, whereas 137 patients underwent exploration only; 47 (39.2%) multivisceral and 45 (37.5%) vascular resections (12 arterial reconstructions) were performed. There were 6 (5%) ypT0 neoplasms, 36 (30.0%) R0, 61 (50.8%) R1, and 16 (13.3%) R2 resections. The median follow-up of surviving patients (n = 22) was 22 months. Median postoperative survival was greater after resection (12.7 months) than after exploration alone (8.8 months; P < .0001). Median postoperative survival was 24.6 months after R0, 11.9 months after R1, and 8.9 months after R2 resection. The 3-year survival rate after R0 resection was 24%. To determine survival after start of neoadjuvant therapy, 3.7 months (median) have to be added. Conclusion. In locally advanced, unresectable pancreatic cancer, R0/R1 resections can be achieved in up to 40% of patients who undergo operation after neoadjuvant therapy. In these cases, survival rates are similar to those observed for initially resectable pancreatic cancer. (Surgery 2012;152:S33-42.) From the Department of Surgery,a the Institute of Pathology,b the Department of Radiation Oncology,c and the Department of Medical Oncology & Nationales Centrum f€ u r Tumorerkrankungen,d University Hospital Heidelberg, Heidelberg, Germany

DESPITE THE PROGRESS IN OPERATIVE AND MEDICAL THERAPY, PANCREATIC CANCER remains the 4th leading cause of cancer-related deaths in Western countries.1,2 Median survival rates of 20–24 months can only be expected after resection in combination with chemotherapy or chemoradiation, usually administered in an adjuvant setting.3-5 Unfortunately, only about 20% of patients have neoplasms amenable to resection at the time of diagnosis; 50–60% of patients present with metastases that preclude resection, whereas 30% of these cancers are unresectable secondary to locally advanced disease.6 The latter Accepted for publication May 11, 2012. Reprint requests: Jens Werner, MD, Department of Surgery, University of Heidelberg, Im Neuenheimer Feld 110, 69120 Heidelberg, Germany. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2012 Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2012.05.029

group of patients may benefit from a neoadjuvant treatment, which may induce downstaging of the cancer and secondary resectability. Over the last 2 decades, pancreatic resections have become increasingly save with mortality rates well below 5% in specialized centers.7-9 Consequently, the indications and the extent of operations have been expanded. In contrast to arterial resections, resections of the portal or superior mesenteric veins, as well as multivisceral resections, are now performed regularly in specialized centers.10-12 Therefore, the criteria for local resectability have changed; currently only neoplasms infiltrating the celiac trunk or the superior mesenteric artery are considered truly unresectable. Some authors classify pancreatic cancers with abutment of the visceral arteries, with infiltration of the hepatic artery, or with short venous occlusion as ‘‘borderline-resectable.’’13-15 Infiltration of neighboring organs (colon, stomach) is usually not SURGERY S33

S34 Strobel et al

regarded as a reason for irresectability.10 However, an internationally accepted consensus defining criteria for local resectability does not exist because of differing criteria between countries and across institutions. The concept of neoadjuvant therapy has been assessed in multiple small trials as well as in 2 recent meta-analyses,6,16 and was found to be of potential benefit for the group of patients with locally advanced, potentially unresectable pancreatic cancer. The available data has to be interpreted with caution for several reasons: (i) Small studies are inherently subject to publication bias; (ii) the criteria for resectability were nonuniform among studies or not clearly defined; (iii) the metaanalyses included studies assessing neoadjuvant therapy for primarily resectable and borderline resectable neoplasms; and (iv) the meta-analyses included studies published as early as in the 1960s that do not reflect the current standard of care. Our aim was to assess resection rates, perioperative outcome, survival, and factors associated with survival after neoadjuvant therapy and exploration for locally advanced, potentially unresectable pancreatic cancer in a large series of 257 consecutive patients treated in a single, high-volume center of pancreatic surgery. PATIENTS AND METHODS Patients. This study was approved by the local ethics committee and is based on a prospective database of 1,901 cases of pancreatic surgery performed for cancer at our institution between October 2001 and December 2009. In all, 257 consecutive patients (13.5%) underwent exploration for resection after neoadjuvant therapy for locally advanced potentially unresectable pancreatic cancer. Patient characteristics are summarized in Table I. Definite criteria for unresectability at our institution were involvement of the celiac trunk or the superior mesenteric artery or a diffuse infiltration of the mesenteric root below the first branches of the superior mesenteric vessels, established by contact of tumor with arteries over a length of $2 cm and of >1808 as assumed by contrast-enhanced computed tomography (CT)/magnetic resonance imaging or during operative exploration. For those patients who were referred to Heidelberg after initial evaluation elsewhere, the exact criteria for unresectability were thoroughly reevaluated at the first presentation at our institution. If we considered the neoplasms to be resectable, the patient underwent reexploration at our institution. There were 181 patients (70.4%) treated initially at our institution, and 76 (29.6%) referred to our institution after

Surgery September 2012

exploration and neoadjuvant treatment or initiation of a palliative treatment elsewhere. The exact criteria of unresectability before neoadjuvant treatment are specified in Table I. Survival was analyzed separately for all patients operated after neoadjuvant treatment and for the subgroup fulfilling our definite criteria for unresectability. Patients who were referred to us from other regions underwent neoadjuvant therapy close to home and were referred back to us for exploration if there was no evidence of metastatic progression during neoadjuvant treatment. Importantly, our analysis includes all patients who underwent exploration for resection after neoadjuvant therapy at our institution but does not include patients who were not operated on after initiation of a neoadjuvant treatment. Neoadjuvant therapy. At presentation at our institution, therapy was defined in a multidisciplinary setting. For locally advanced pancreatic cancer without evidence of systemic disease, we usually recommend neoadjuvant chemoradiation which is performed in the Department of Radiooncology at the University of Heidelberg. The present series also includes patients treated in cooperation with extraregional radiooncologists or medical oncologists, as well as patients in whom palliative chemotherapy had been initiated. This cohort results in some heterogeneity in the neoadjuvant treatment. 199 patients (77.4%) received neoadjuvant chemoradiation (RCHT) and 58 patients (22.6%) underwent neoadjuvant chemotherapy only (CHT). Patients with RCHT received a median dose of 50.4 Gy (interquartile range, 45.0–54.0) plus gemcitabine monotherapy (n = 119; 46.3%), gemcitabine/cetuximab (n = 37; 14.4%), other gemcitabine-based combinations (n = 23; 8.9%), or a 5-fluorouracil–based therapy (n = 16; 6.2%). Patients with CHT alone received gemcitabine monotherapy (n = 31; 12.1%), other gemcitabinebased combinations (n = 20; 7.8%), or a 5-fluorouracil–based chemotherapy (n = 7; 2.7%). Restaging was performed 4–6 weeks after completion of RCHT and 2–3 weeks after CHT, respectively. Restaging always consisted of contrastenhanced CT of the abdomen and an x-ray of the chest. CT of the thorax and positron emission tomography (PET) were used for lesions suspicious for metastases. Additionally, serum levels of carbohydrate antigen 19-9 (CA19-9) and carcinoembryonic antigen (CEA) were determined before neoadjuvant therapy and at the time of restaging. Operative approach. Operative exploration was performed if restaging showed no evidence of local disease progression or systemic disease.

Strobel et al S35

Surgery Volume 152, Number 3S

Table I. Patient characteristics and criteria for primary unresectability Parameter

Total (n = 257)

Gender, n (%) Female 111 (43.2) Male 146 (56.8) Age (yrs)* 61.0 (54.3–67.0) ASA score, n (%)y I 5 (1.9) II 107 (41.6) III 108 (42.0) IV 4 (1.6) Previous extrapancreatic malignancy 36 (14.0) Definite criteria for primary 161 (62.6) unresectabilityz Infiltration of celiac trunk 66 (25.7) Infiltration of superior mesenteric 86 (33.5) artery Diffuse infiltration of mesenteric root 46 (17.9) Extended criteria for primary 75 (29.2) unresectabilityz Infiltration of hepatic artery 22 (8.6) Extended venous infiltration 39 (15.2) Infiltration of neighboring organs 7 (2.7) Suspected metastasesx 24 (9.3) No exact criteria given (exploration) 21 (8.2) Initial unresectability diagnosed by Surgical exploration 117 (45.5) Cross-sectional imaging only 140 (55.5) Unresectability established before neoadjuvant therapy in Heidelberg 181 (70.4) Referring hospital only 76 (29.6)

Resections (n = 120)

Explorations (n = 137)

58 (48.3) 62 (51.7) 60.2 (54.0–66.9)

53 (38.7) 84 (61.3) 61.9 (54.7–67.0)

2 46 46 2 20 80

(1.7) (38.3) (38.3) (1.7) (16.7) (66.7)

3 61 62 2 16 81

(2.2) (44.5) (45.3) (1.5) (11.7) (59.1)

32 (26.7) 41 (34.2)

34 (24.8) 45 (32.8)

24 (20.0) 30 (25.0)

22 (16.1) 45 (32.9)

7 16 3 9 10

15 23 4 15 11

(5.8) (13.3) (2.5) (7.5) (8.3)

(10.9) (16.8) (2.9) (10.9) (8.0)

55 (45.8) 65 (54.2)

62 (45.3) 75 (54.7)

89 (74.2) 31 (25.8)

92 (67.2) 45 (32.8)

*Median (interquartile range). yNot available in 33. zSeveral criteria in some patients. xUnclear liver or peritoneal lesions on cross-sectional imaging. ASA, American Society for Anesthesiology.

Importantly, the lack of a local radiologic response to neoadjuvant therapy precluded exploration only in those neoplasms incasing the superior mesenteric artery. In patients undergoing operative exploration, any further approach was based on intraoperative findings (Fig 1). In all patients, a thorough exploration of the abdominal cavity, including intraoperative ultrasonography of the liver, was performed to detect peritoneal or liver metastases. If metastatic disease was confirmed by intraoperative frozen section, the operation was concluded or a palliative bypass procedure was performed. If no metastases were detected, the local situation was assessed extensively as the first step of the operation using the artery first approach.17 Soft tissue or scar tissue around the arteries was biopsied and analyzed by frozen section. After neoadjuvant

therapy, a layer between arterial wall and surrounding tissue often allowed to ‘‘peel off’’ suspected tumor/scar tissue from the arteries. Neoplasms with direct arterial infiltration (confirmed by frozen section) were deemed locally unresectable in most cases. Only rarely was an arterial resection performed. We aimed to avoid R2 resections by using the artery first approach.17 In cases without arterial infiltration, a resection was performed. Venous and multivisceral resections were performed whenever necessary for macroscopically complete removal of the neoplasm. In case of suspected residual disease after resection and in some patients with unresectability, intraoperative radiation therapy (IORT) with a dose of 10–15 Gy was administered (Fig 1). IORT was usually combined with neoadjuvant or adjuvant RCHT.

S36 Strobel et al

Surgery September 2012

Fig 1. Therapeutic approach depending on intraoperative findings. IORT, Intraoperative radiation therapy; w/o, without.

Data acquisition. Parameters recorded prospectively in our database included age, gender, tumor type and stage, serum CA19-9 and CEA values, criteria for primary unresectability, previous operations, regimen of neoadjuvant therapy, type of operation, use of IORT, intraoperative blood loss, operative time, postoperative hospital stay, morbidity, and 30-day mortality. Operative morbidity included anastomotic leak, pancreatic fistula (grades A–C),18 chylus fistula, intra-abdominal hemorrhage, abscess, peritonitis, and wound infections. Medical morbidity included delayed gastric emptying, cardiopulmonary complications, urinary tract infections, and liver failure. Follow-up. All patients were followed until their last oncologic follow-up examination or until death. Follow-up data was acquired in our outpatient care unit, by external oncologic follow-up examinations, or by an additional phone interview of patients, relatives, or general practitioners in the summer of 2011. In case of death, the date of death was recorded. Statistical analysis. SAS software (Release 9.1, SAS Institute, Inc, Cary, NC) was used. The quantitative parameters age, follow-up time, operative time, blood loss, and duration of hospital stay, as well as serum values of CEA and CA19-9 are expressed as median values with interquartile ranges. In all graphs, overall survival is defined as the time from the date of the operation to either death from any cause or last follow-up. The survival rates were assessed using the Kaplan–Meier method. Patients alive at the last follow-up were censored. The 2- and 3-year survival rates and the

median survival time are presented. The log-rank test was used to compare survival curves of subgroups. Two sided P-values were computed. RESULTS Intraoperative findings and operative procedures. Of 257 patients undergoing exploration after neoadjuvant therapy, a resection was performed in 120 (46.7%) patients, whereas 137 (53.3%) patients underwent exploration only (Fig 1). In 84 of those patients who were explored (61.3%), exploration revealed metastatic disease that was not detected in preoperative imaging. In the other 53 patients in whom exploration revealed no metastases, frozen sections revealed persisting, direct, arterial infiltration indicating local unresectability; 10 of these patients received IORT as an individual treatment approach. In addition, 53 patients in whom the neoplasm could be resected (44.2% of resections) IORT was applied after resection. A detailed description of the operative procedures is given in Table II. In patients with resection, most neoplasms were located in the head of the pancreas, mirrored by a high percentage of pancreatoduodenectomies (60% of all resections). For complete macroscopic resection, extended resection was necessary frequently, as reflected by 39% multivisceral resections and 38% vascular resections with reconstruction. Arterial resections with reconstruction were performed in 10%. In 66% of explorations, only biopsies were taken, whereas in 34%, an additional bypass procedure was performed.

Strobel et al S37

Surgery Volume 152, Number 3S

Table II. Operative procedures and intraoperative details Resection procedures Classical pancreatoduodenectomy Pylorus preserving pancreatoduodenectomy Distal pancreatectomy Total pancreatectomy Multivisceral resection Vascular resection with reconstruction Arterial Venous Intraoperative radiation therapy Operation time (min)* Blood loss (mL)*

Total (n = 120) 16 (13.3%) 56 (46.7%) 30 18 47 45 12 39 53 403 800

(25.0%) (15.0%) (39.2%) (37.5%) (10.0%) (32.5%) (44.2%) (335–490) (500–1,200)

Explorations

Total (n = 137)

Biopsies only Laparotomy

91 (66.4%) 87 (63.5%)

Laparoscopy Bypass procedure Double bypass Hepaticojejunostomy only Gastrojejunostomy only

4 46 22 14 10

(2.9%) (33.6%) (16.1%) (10.2%) (7.3%)

16 (11.7%) 128 (83–190) 150 (50–300)

*Median (interquartile range).

Pathologic findings. The final histology revealed pancreatic ductal adenocarcinoma in 231 (89.9%) patients, pancreatic ductal adenocarcinoma originating from intraductal papillary mucinous neoplasia in 5 patients (1.9%), ampullary adenocarcinoma in 5 patients (1.9%), and adenocarcinoma of the distal bile duct in 1 patient (0.4%). In 15 patients who underwent exploration only, the final pathologic diagnosis was peritoneal or liver metastases of adenocarcinoma without further differentiation. For resected neoplasms, final pathology revealed a complete pathologic response (ypT0) in 6 (5.0%) patients, ypT1/2 in 4 (3.3%), ypT3 in the majority of 90 (75.0%) patients, and ypT4 in 20 (16.7%) patients. Of patients with resections, 59 patients (49.2%) had no lymph node involvement (ypN0), whereas 59 patients showed involvement of peripancreatic lymph nodes (ypN1). Reflecting the decision making for resection, the majority of 111 (92.5%) patients had no distant metastases; however, in 9 (7.5%) patients, metastases were diagnosed (ypM1). Of these patients, 3 had positive interaortocaval lymph nodes, in 2 patients single liver metastases were resected, and in 4 patients frozen sections of peritoneal lesions were diagnosed primarily as tumor free, but definitive pathology including immunohistochemistry revealed adenocarcinoma. Importantly, final pathology revealed tumor-free margins (R0) in 42 patients (35.0% of resections; n = 6 ypT0 included), an R1 situation was found in 61 (50.8%) patients. In 16 (13.3%) patients, intraoperative and pathologic findings together were interpreted as R2. In 1 (0.8%) case with complete macroscopic resection, the resected specimen did not allow for a conclusive R classification (Rx). Perioperative results. An overview of perioperative outcomes is given in Table III. The median

duration of postoperative hospital stay was 12 days after resection and 8 days after exploration. Reflecting the extent of the intervention, operation-related morbidity was 33% after resection, but only 9% after exploration. Mortality was 4.2% after resections and associated with arterial resections and IORT. Only 1 patient died after an exploration with double bypass. After resection, the predominant complications were intra-abdominal abscesses or chylus fistulae in 10% of patients, respectively, as well as postoperative hemorrhage in 7.5%. Interestingly, pancreatic fistulae were observed rarely (3 patients; 2.5% of 102 patients at risk). Delayed gastric emptying was observed in 8.3% of patients. The reoperation rate after resections was 17.5% (n = 10). Twenty-five patients (9.7%) underwent interventional drainage of intra-abdominal fluid collections. Survival analysis and prognostic factors. A total of 14 patients (5.4%) were lost to follow-up and were excluded from the survival analysis; 22 patients were still alive after a median follow-up of 22 (range, 15–36) months in survivors. The median survival of all patients was 13.8 months after the start of neoadjuvant therapy and 10.1 months after operation. The median difference of 3.7 months corresponds with the duration of neoadjuvant therapy plus the interval to operation and has to be added to postoperative survival data if compared with survival data for palliative chemoradiation/chemotherapy or upfront operation. Median survival was significantly greater after resection compared to exploration (13 vs 9 months; P < .0001). This finding is most likely due to intraoperative selection of patients who did not have metastatic disease at the time of operation/resection. Survival of patients without metastases was still significantly greater after resection

S38 Strobel et al

Surgery September 2012

Table III. Perioperative results Resection Nonmultivisceral (n = 73; 60.8%)

Multivisceral and arterial (n = 47; 39.2%)

Total (n = 137)

Biopsy only (n = 91; 66.4%)

Bypass (n = 46; 33.6%)

12 (9–19)

12 (10–17)

12 (9–20)

8 (6–11)

7 (6–10)

10 (7–13)

57 (47.5%) 39 (32.5%)

31 (42%) 22 (30%)

26 (55%) 17 (36%)

23 (16.8%) 13 (9.5%)

12 (13%) 7 (8%)

11 (24%) 6 (13%)

35 (29.2%) 21 (17.5%) 17 (14.2%)

19 (26%) 12 (16%) 7 (10%)

16 (34%) 9 (19%) 10 (21%)

16 (11.7%) 6 (4.4%) 8 (5.8%)

8 (9%) 2 (2%) 3 (3%)

8 (17%) 4 (9%) 5 (11%)

5 (4.2%)

1 (1%)

4 (9%)

1 (0.7%)

0 (0%)

1 (2%)

Total (n = 120) Postoperative hospital stay (d)* Total morbidity Operation-related morbidity Medical morbidity Reoperation Interventional drainage Mortality

Exploration

*Median (interquartile range).

compared to those patients who were not resected owing to persisting local unresectability (Fig 2, A). ‘‘Long-term’’ survival was only observed after resection. Survival after resection was associated with the type of resection (R status; Fig 2, B) as well as with tumor stages (as derived from yTNM classifications; Fig 2, C). Strikingly, patients with R0 resection had a median postoperative survival of 25 months and a median survival of 28 months after initiation of the neoadjuvant therapy. In patients with complete radiologic response and (y)UICC stage I tumors, median postoperative survival was 25 months and the 3-year survival rate was 38%. In patients with exploration only, survival was significantly shorter in metastatic compared with locally advanced disease (11 months without metastases vs 8 months with metastases; P = .0004). Overall postoperative survival in all patients was associated with preoperative (post neoadjuvant therapy) CA19-9 serum levels (Fig 2, D) as well as with preoperative CEA levels (not shown). Between patients treated with neoadjuvant RCHT or CHT only, there was neither a difference with regard to resection rates nor in survival in the entire cohort, or in any of the subgroups analyzed. In the group with resection, patients with additional IORT had a lesser survival compared with patients without IORT (12 vs 18 months; P = .04). This observation most likely mirrors an effect of patient selection, because patients with suspected residual disease were more likely to receive IORT. In contrast, those patients who were only explored owing to local unresectability tended to live longer after exploration with additional IORT (15 months with IORT) compared with those in whom IORT was not performed (10 months without IORT; P = .0902). The criteria for unresectability have important implications for patient selection as well as for the

interpretation of survival results and the potential value of neoadjuvant therapy. Some of our patients had been referred after initial treatment elsewhere. There was no difference in overall survival of patients initially treated in Heidelberg and patients in whom unresectability was first diagnosed elsewhere. This finding was true for both patients who were resected and those who were only explored. We also performed a subgroup analysis for the 161 (62.2%) patients with definite criteria of unresectability (infiltration of the celiac trunk and the superior mesenteric artery). Survival of this subgroup did not differ from survival of patients with extended criteria in the overall cohort (median survival 11 vs 9 months, respectively; P = .53), nor in the subgroups with resection (median survival 15 vs 10 months, respectively; P = .82), or exploration (median survival 9 vs 9 months, respectively; P = .95). The subgroup of patients with definite criteria for primary unresectability still had a significantly greater median survival after resection if compared to exploration (persistence of local unresectability) (Fig 3, A). Survival after resection still tended to be associated with the status of resection (Fig 3, B). DISCUSSION Resection combined with systemic chemotherapy or chemoradiation remains the only treatment with potential cure or ‘‘long term’’ survival in pancreatic cancer. About 30–40% of patients are diagnosed with locally advanced, unresectable tumors.6 In this study, we assessed the results of resection after neoadjuvant therapy in patients diagnosed initially as locally advanced, unresectable pancreatic cancer. In the treatment of locally advanced, clinically unresectable pancreatic cancer, there have been

Strobel et al S39

Surgery Volume 152, Number 3S

% 100

% 100

Patients without metastases:

Status of resection:

Survival Distribution Function

Survival Distribution Function

Exploration: N=51; 10.6 months median survival; 13.2% 2-year survival rate Resection: N=107; 12.7 months median survival; 18.3% 3-year survival rate

75

P=0.025

50

25

0 0

12

36 48 60 Time after operation (months)

72

0/IA/IB: N=8; 25 months median survival; 38% 3-year survival rate IIA: N=41; 23 months median survival; 25% 3-year survival rate IIB: N=40; 11 months median survival; 10% 3-year survival rate III: N=18; 4 months median survival; 14% 2-year survival rate IV: N=8; 10 months median survival; 17% 2-year survival rate

P<0.03

50

25

0

0

12

P=0.19

50

25

0

12

% 100

UICC stages after resection:

75

75

0

84

24

36 48 60 Time after operation (months)

72

84

Preoperative CA19-9 levels [U/ml]: <37: N=69; 14 months median survival; 12% 3-year survival rate 37-<400: N=110; 10.2 months median survival; 7.8% 3-year survival rate ≥400: N=48; 6 months median survival; 3% 2-year survival rate

Survival Distribution Function

Survival Distribution Function

% 100

24

R0: N=41; 24.6 months median survival; 23.7% 3-year survival rate R1: N=57; 11.9 months median survival; 12.9% 3-year survival rate R2: N=16; 8.9 months median survival; 14.1% 2-year survival rate

75

P<0.0001

50

25

0 24

36 48 60 Time after operation (months)

72

84

0

12

24

36 48 60 Time after operation (months)

72

84

Fig 2. Overall postoperative survival (Kaplan–Meier estimates). (A) Survival of patients without metastases after resection versus exploration for local unresectability. (B) Survival after resection dependent on the status of resection (R status). (C) Survival after resection dependent on tumor stages (as derived from yTNM classification). (D) Influence of preoperative serum CA19-9 levels on postoperative survival. Survival data are calculated starting from the date of the operation. To estimate survival from the beginning of neoadjuvant treatment, a median of 3.7 months has to be added. Patients alive at the last follow-up are censored (I).

phase II trials evaluating palliative chemoradiation and randomized, controlled trials comparing chemoradiation with chemotherapy alone,19,20 but so far there are no published trials comparing palliative therapy with a strategy of neoadjuvant therapy followed by exploration. A recent systematic review suggests that neoadjuvant therapy may be of benefit for patients with locally advanced pancreatic cancer and that up to one third of patients with primarily borderline resectable or clinically unresectable neoplasms would be expected to have resectable tumors after neoadjuvant therapy with comparable survival to patients with initially resectable tumors.6 A systematic review of concurrent studies on neoadjuvant chemoradiation for primarily unresectable pancreatic cancer found very

variable resection rates (8–64%) and a surprisingly high rate of R0 resections (median, 88%).21 It remains unclear if resection margins in these studies were analyzed according to current standards.22,23 The survival rates of patients treated with chemoradiation in a neoadjuvant setting compares favorably with those reported after palliative chemoradiation alone.21 These systematic reviews, however, include studies that lack a clear definition of criteria of local unresectability or included patients with "borderline" resectable neoplasms, potentially affecting both resection rates and survival data. Moreover, the inclusion of relatively small retrospective studies may result in an overestimation of resection rates and survival data owing to publication bias for positive results.

S40 Strobel et al

% 100

Only patients with definite criteria for primary unresectability without metastases: Exploration: N=36; 11 months median survival; 9% 2-year survival rate Resection: N=72; 15 months median survival; 21% 3-year survival rate

Survival Distribution Function

Survival Distribution Function

% 100

Surgery September 2012

75

P<0.004

50

25

Only patients with definite criteria for primary unresectability Status of resection: R0: N=25; 28 months median survival; 31% 3-year survival rate R1: N=37; 13 months median survival; 13% 3-year survival rate R2: N=14; 13 months median survival; 16% 3-year survival rate

75 P=0.19

50

25

0

0 0

12

24 36 Time after surgery (months)

48

60

0

12

24 36 Time after operation (months)

48

60

Fig 3. Overall postoperative survival (Kaplan–Meier estimates) of the subgroup of patients with definite criteria for primary unresectability. (A) Survival of patients without metastases after resection versus exploration for local unresectability. (B) Survival after resections dependent on the status of resection (R status).

To the best of our knowledge, the current report on 257 patients is the largest study assessing the results of exploration for resection after neoadjuvant therapy for patients diagnosed initially as clinically unresectable owing to locally advanced cancer. A limitation of our study is that the cohort only includes patients who underwent exploration after neoadjuvant therapy. The referral pattern of the patients did not allow us to recognize and follow-up the total cohort of patients entering neoadjuvant therapy. Therefore, the number of patients who did not complete the neoadjuvant therapy or developed systemic disease progression and did not undergo exploration remains unknown. Indeed, 97 patients in the present study were also included in a study led by our radiation oncologists focusing on the efficiency of a gemcitabine-based chemoradiation protocol for neoadjuvant treatment in locally advanced, unresectable pancreatic cancer. In this study, the overall resection rate of all patients receiving neoadjuvant chemoradiation (n = 215) was 26%.24 The strength of our study is that the large number of patients provides solid data on the extent of operation necessary as well as the perioperative outcome of operation after neoadjuvant therapy for initially unresectable pancreatic cancer. Furthermore, it provides the opportunity for a detailed survival analysis and for the identification of prognostic parameters that allow us to better define which patients may actually benefit from exploration after neoadjuvant therapy. This information is useful for operative decision making as

well as for the design of future randomized, controlled trials. The operative procedures performed show clearly that pancreatic resections after neoadjuvant therapy for locally advanced, initially ‘‘unresectable’’ neoplasms are technically demanding and often involve vascular and multivisceral resections. This fact is also reflected by the morbidity and mortality rates that are in an acceptable range given the extent of operation, but are greater in comparison with standard resections.7-10 Together, these data recommend that exploration after neoadjuvant therapy for locally advanced pancreatic cancer should only be performed in experienced (high-volume) centers of pancreatic surgery that regularly perform multivisceral and venous resections. Even with extended resections, the exploration rate of 53.3% in the present study remains relatively high. Identification of peritoneal or liver metastases not found by preoperative restaging with cross-sectional imaging techniques were the predominant reason for unresectability, accounting for 61.3% of explorations. As known from primarily resectable pancreatic cancer, the preoperative identification of small liver or peritoneal metastases remains challenging by all available diagnostic tools, including PET scans and staging laparoscopy.25-27 In our study, postoperative survival was significantly associated with preoperative (post neoadjuvant therapy) CA19-9 levels, similar to what has been observed for initially resectable pancreatic cancer.28 In the future, tumor markers such as CA19-9 may be useful as one parameter

Strobel et al S41

Surgery Volume 152, Number 3S

to select patients for exploration. Persisting local unresectability accounted for 39% of explorations. Current techniques of cross-sectional imaging, including PET-CT, cannot reliably distinguish vital cancer from scar tissue. In our experience, soft tissue around arteries observed in restaging examinations after neoadjuvant therapy often represents scar tissue without viable tumor. Refinement of the preoperative diagnostic workup to lessen the rate of exploration without resection has to be one focus of future studies. The morbidity and mortality of exploration is relatively low, suggesting that exploration should be offered whenever restaging after neoadjuvant therapy reveals a chance for resection. Our overall cohort, including patients after both resection or exploration only, has a median overall survival of 10 months after operation and of 14 months after start of neoadjuvant therapy. These survival data compare favorably with the outcome reported for palliative chemotherapy or chemoradiation of 6–11 months.19,29 Because our study does not include a control group of patients undergoing palliative treatment, we cannot prove that neoadjuvant therapy and resection is superior to palliative treatment; however, several results derived from the subgroup analyses of resected patients suggest strongly that subsets of patients do benefit from this approach. First, median survival in patients with localized disease is significantly greater after resection compared with exploration and ‘‘long-term’’ survival of 4 years and more is observed only in the group with resection (Fig 2, A). Second, after resection, survival is associated with the state of resection (R status) and with tumor stage (Fig 2, B and C). Similar to initially resectable pancreatic cancer, this points to a benefit of resection per se. After R0 resections, we observed a median postoperative survival of 25 months and a median survival calculated from the beginning of therapy of 28 months, which is much more than one would expect from palliative therapy. As suggested by Gillen et al,6 the survival results in the resected group were comparable with those reported after upfront resection of initially resectable pancreatic cancers. A subgroup analysis in our cohort revealed no difference in resection rate and survival in patients treated with neoadjuvant chemoradiation or chemotherapy alone. The effect of IORT remains unclear. The lesser survival of patients with IORT in the subgroup with resection is caused likely by selection of patients with suspected positive margins for IORT. In contrast, the greater survival of patients with IORT for locally unresectable disease

without metastases may suggest a potential benefit of IORT in this subgroup. Our study suggests that neoadjuvant therapy may result in effective downstaging of locally advanced, initially unresectable pancreatic cancer in a subset of patients. If resection can be performed after neoadjuvant therapy, survival is comparable with what is described for resection and adjuvant therapy for initially resectable pancreatic cancer. Thus, patients with locally advanced pancreatic cancer should be considered not only to receive palliative treatment, but should undergo reevaluation for potential operative exploration based on restaging results or they should be included in neoadjuvant protocols. Several key questions remain to be answered in future studies. Randomized, controlled trials are needed to assess whether neoadjuvant therapy and exploration is superior to chemoradiation or chemotherapy alone. Furthermore, the optimal protocol for neoadjuvant treatment needs to be defined. To make it even more complicated, it is likely that only subsets of patients may benefit from neoadjuvant protocols and exploration or from specific protocols. Therefore, the identification of better parameters for patient selection will be important. In the future, molecular markers may help to distinguish patients with localized disease from patients at high risk of systemic progression and help to select them for chemoradiation or systemic chemotherapy. REFERENCES 1. American Cancer Society. Cancer facts & figures. Atlanta: American Cancer Society; 2010. 2. Krebs in Deutschland 2005/2006. Haufigkeiten und Trends. 7. Ausgabe. Robert Koch-Institut (Hrsg) und die Gesellschaft der epidemiologischen Krebsregister in Deutschland e. V. (Hrsg). Berlin. 2010. 3. Neoptolemos JP, Stocken DD, Friess H, Bassi C, Dunn JA, Hickey H, et al. A randomized trial of chemoradiotherapy and chemotherapy after resection of pancreatic cancer. N Engl J Med 2004;350:1200-10. 4. Neoptolemos JP, Stocken DD, Bassi C, Ghaneh P, Cunningham D, Goldstein D, et al. Adjuvant chemotherapy with fluorouracil plus folinic acid vs gemcitabine following pancreatic cancer resection: a randomized controlled trial. JAMA 2010;304:1073-81. 5. Oettle H, Post S, Neuhaus P, Gellert K, Langrehr J, Ridwelski K, et al. Adjuvant chemotherapy with gemcitabine vs observation in patients undergoing curative-intent resection of pancreatic cancer: a randomized controlled trial. JAMA 2007;297:267-77. 6. Gillen S, Schuster T. Meyer Zum BC, Friess H, Kleeff J. Preoperative/neoadjuvant therapy in pancreatic cancer: a systematic review and meta-analysis of response and resection percentages. PLoS Med 2010;7:e1000267. 7. Buchler MW, Wagner M, Schmied BM, Uhl W, Friess H, Z’graggen K. Changes in morbidity after pancreatic

S42 Strobel et al

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

resection: toward the end of completion pancreatectomy. Arch Surg 2003;138:1310-4. Cameron JL, Riall TS, Coleman J, Belcher KA. One thousand consecutive pancreaticoduodenectomies. Ann Surg 2006;244:10-5. Fernandez-del Castillo C, Rattner DW, Warshaw AL. Standards for pancreatic resection in the 1990s. Arch Surg 1995;130:295-9. Hartwig W, Hackert T, Hinz U, Hassenpflug M, Strobel O, Buchler MW, et al. Multivisceral resection for pancreatic malignancies: risk-analysis and long-term outcome. Ann Surg 2009;250:81-7. Muller SA, Hartel M, Mehrabi A, Welsch T, Martin DJ, Hinz U, et al. Vascular resection in pancreatic cancer surgery: survival determinants. J Gastrointest Surg 2009;13:784-92. Tseng JF, Tamm EP, Lee JE, Pisters PW, Evans DB. Venous resection in pancreatic cancer surgery. Best Pract Res Clin Gastroenterol 2006;20:349-64. Katz MH, Pisters PW, Evans DB, Sun CC, Lee JE, Fleming JB, et al. Borderline resectable pancreatic cancer: the importance of this emerging stage of disease. J Am Coll Surg 2008;206:833-46. Stokes JB, Nolan NJ, Stelow EB, Walters DM, Weiss GR, de Lange EE, et al. Preoperative capecitabine and concurrent radiation for borderline resectable pancreatic cancer. Ann Surg Oncol 2011;18:619-27. Evans DB, Erickson BA, Ritch P. Borderline resectable pancreatic cancer: definitions and the importance of multimodality therapy. Ann Surg Oncol 2010;17:2803-5. Assifi MM, Lu X, Eibl G, Reber HA, Li G, Hines OJ. Neoadjuvant therapy in pancreatic adenocarcinoma: A metaanalysis of phase II trials. Surgery 2011;150:466-73. Weitz J, Rahbari N, Koch M, Buchler MW. The ‘‘artery first’’ approach for resection of pancreatic head cancer. J Am Coll Surg 2010;210:e1-4. Bassi C, Dervenis C, Butturini G, Fingerhut A, Yeo C, Izbicki J, et al. Postoperative pancreatic fistula: an international study group (ISGPF) definition. Surgery 2005;138: 8-13. Cardenes HR, Moore AM, Johnson CS, Yu M, Helft P, Chiorean EG, et al. A phase II study of gemcitabine in combination with radiation therapy in patients with localized,

Surgery September 2012

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

unresectable, pancreatic cancer: a Hoosier Oncology Group Study. Am J Clin Oncol 2011;34:460-5. Chauffert B, Mornex F, Bonnetain F, Rougier P, Mariette C, Bouche O, et al. Phase III trial comparing intensive induction chemoradiotherapy (60 Gy, infusional 5-FU and intermittent cisplatin) followed by maintenance gemcitabine with gemcitabine alone for locally advanced unresectable pancreatic cancer. Definitive results of the 2000-01 FFCD/ SFRO study. Ann Oncol 2008;19:1592-9. Morganti AG, Massaccesi M, La Torre G, Caravatta L, Piscopo A, Tambaro R, et al. A systematic review of resectability and survival after concurrent chemoradiation in primarily unresectable pancreatic cancer. Ann Surg Oncol 2010;17: 194-205. Esposito I, Kleeff J, Bergmann F, Reiser C, Herpel E, Friess H, et al. Most pancreatic cancer resections are R1 resections. Ann Surg Oncol 2008;15:1651-60. Verbeke CS, Leitch D, Menon KV, McMahon MJ, Guillou PJ, Anthoney A. Redefining the R1 resection in pancreatic cancer. Br J Surg 2006;93:1232-7. Habermehl D, Kessel K, Welzel T, Hof H, Abdollahi A, Bergmann F, et al. Neoadjuvant chemoradiation with Gemcitabine for locally advanced pancreatic cancer. Radiat Oncol 2012;7:28. Pakzad F, Groves AM, Ell PJ. The role of positron emission tomography in the management of pancreatic cancer. Semin Nucl Med 2006;36:248-56. Schachter PP, Avni Y, Shimonov M, Gvirtz G, Rosen A, Czerniak A. The impact of laparoscopy and laparoscopic ultrasonography on the management of pancreatic cancer. Arch Surg 2000;135:1303-7. Vollmer CM, Drebin JA, Middleton WD, Teefey SA, Linehan DC, Soper NJ, et al. Utility of staging laparoscopy in subsets of peripancreatic and biliary malignancies. Ann Surg 2002; 235:1-7. Kondo N, Murakami Y, Uemura K, Hayashidani Y, Sudo T, Hashimoto Y, et al. Prognostic impact of perioperative serum CA 19-9 levels in patients with resectable pancreatic cancer. Ann Surg Oncol 2010;17:2321-9. Conroy T, Desseigne F, Ychou M, Bouche O, Guimbaud R, Becouarn Y, et al. FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. N Engl J Med 2011;364:1817-25.