- Email: [email protected]
Is pancreaticoduodenectomy with mesentericoportal venous resection safe and worthwhile?
The American Journal of Surgery 182 (2001) 120 –129
Is pancreaticoduodenectomy with mesentericoportal venous resection safe and worthwhile? Philippe Bachellier, M.D., Hiroshi Nakano, M.D., Ph.D., Elie Oussoultzoglou, M.D., Jean-Christophe Weber, M.D., Karim Boudjema, M.D., Ph.D., Philippe Wolf, M.D., Ph.D., Daniel Jaeck, M.D., Ph.D., F.R.C.S.* Centre de Chirurgie Visce´rale et de Transplantation, Hoˆpital Universitaire de Hautepierre, Avenue Molie`re, 67098 Strasbourg Cedex, France. Manuscript received November 10, 2000; revised manuscript May 14, 2001
Abstract Background: Whether or not superior mesentericoportal venous resection (SM-PVR) associated with pancreaticoduodenectomy (PD) is safe and worthwhile has not been fully confirmed. The aim of the present study was to investigate results of this surgical procedure performed for pancreatic head and periampullary neoplasms. Methods: As a first analysis, postoperative morbidity and mortality after PD with (n ⫽ 31) or without SM-PVR (n ⫽ 119) were investigated in 150 patients with pancreatic head and periampullary neoplasms. As a second analysis, rates of margin-negative resection and survival after SM-PVR (n ⫽ 21) and without SM-PVR (n ⫽ 66) were compared in 87 patients with pancreatic ductal adenocarcinoma of the pancreatic head. In these patients undergoing SM-PVR (n ⫽ 21), survival rate was investigated in patients who did (n ⫽ 13) and did not (n ⫽ 8) undergo a margin-negative resection. Results: In the first analysis, duration of surgery and volume of blood transfused perioperatively were higher in patients undergoing SM-PVR. However, mortality, morbidity rates, and mean hospital stay did not differ between patients who did undergo SM-PVR (31 patients, 3.2%, 48.4%, and 22.2 days, respectively) and who did not (119 patients, 2.5%, 47.1%, 25.9 days, respectively). No postoperative death occurred in the recent part of the present study, since 1994, in patients undergoing SM-PVR. In the second analysis of pancreatic ductal adenocarcinoma, rates of margin-negative resection and 2-year survival did not significantly differ between patients who did and did not undergo SM-PVR (62% and 22%, respectively, versus 73% and 24%). In patients undergoing SM-PVR, survival rate was significantly higher for patients undergoing a margin-negative resection (n ⫽ 13) than for patients undergoing a macroscopic or microscopic margin-positive resection (n ⫽ 8, 2-year survival ⫽ 57.1% versus 0%, P ⬍0.05). Conclusion: PD combined with SM-PVR can be performed safely. This surgical procedure is followed by a promising survival rate and can be recommended in order to obtain a margin-negative resection; however, candidates for SM-PVR should be carefully selected. © 2001 Excerpta Medica, Inc. All rights reserved. Keywords: Pancreatic cancer; Pancreaticoduodenectomy; Mesentericoportal venous resection
Pancreaticoduodenectomy (PD) is the only potential cure for ductal adenocarcinoma of the pancreatic head or periampullary tumors [1– 4]. Postoperative mortality rates reported currently for PD range from 0% to 8% [2–5]. In addition, 5-year survival rates in patients with pancreatic ductal adenocarcinoma undergoing pancreatic resection ranged from 7% to 25% [1– 4]. However, survival rates after pancreatic resection remain very poor when the surgical margin is positive for cancer [6 – 8]. In an attempt to obtain * Corresponding author. Tel.: ⫹33-388-12-7258; fax: ⫹33-388-127286. E-mail address: [email protected]
a negative surgical margin, the superior mesenteric vein (SMV), superior mesenteric-portal vein (SM-PV) confluence, portal vein (PV), and the retropancreatic arteries are sometimes resected [4,8 –19]. Morbidity rates after this additional superior mesentericoportal venous resection (SMPVR) range from 26% to 55% [11,14,16,17]. However, SM-PVR has not been universally accepted because it has yet to be made clear that this procedure can be performed with acceptable morbidity and mortality even if it increases the rate of the margin-negative resection [11,13–15]. In addition, whether SM-PVR improves long-term survival has not been fully confirmed. In the present study, we investigated morbidity and mor-
0002-9610/01/$ – see front matter © 2001 Excerpta Medica, Inc. All rights reserved. PII: S 0 0 0 2 - 9 6 1 0 ( 0 1 ) 0 0 6 8 6 - 9
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129 Table 1 Indications for pancreaticoduodenectomy with or without SM-PVR
Pancreatic ductal adenocarcinoma Cholangiocarcinoma Neuroendocrine carcinoma Pancreatic cystic neoplasm Ampullary adenocarcinoma Ampullary adenoma Duodenal adenocarcinoma
PD or TP without vascular resection (n ⫽ 119)
PD or TP with SM-PVR (n ⫽ 31)
Total (n ⫽ 150)
66
21
87
11 8
2 6
13 14
12
1
13
15
1
16
4 3
0 0
4 3
PD ⫽ pancreaticoduodenectomy; TP ⫽ total pancreatectomy; SMPVR ⫽ superior mesentericoportal vein resection.
tality after PD with and without SM-PVR in patients with ductal adenocarcinoma of the pancreatic head and periampullary neoplasms. In addition, in an attempt to determine whether this procedure is justified in the treatment of malignant diseases, we investigated the rates of the marginnegative resection and cumulative survival after PD associated with SM-PVR in patients with ductal adenocarcinoma of the pancreatic head.
Patients and methods First analysis: all patients undergoing pancreatectomy From January 1990 through March 1999, 187 patients underwent PD or total pancreatectomy (TP) in our department. One hundred fifty patients (86 men and 64 women) with ductal adenocarcinoma of the pancreatic head or periampullary neoplasms were selected for the current analysis among these 187 patients (Table 1). The mean age of the patients was 61.4 ⫾ 1.0 years (range 27 to 85). Among the 150 patients, 116 (77.3%) underwent PD (including 6 patients who underwent a pylorus-preserving procedure) and 34 (22.7%) underwent TP. Among the 116 patients undergoing PD, 52 underwent reconstruction with pancreaticojejunostomy from 1990 through 1994. Since 1994, 64 patients underwent pancreaticogastrostomy with the method of Delcore et al [20] in which 2 cm of pancreatic remnant is telescoped into the gastric lumen, without stenting in the main pancreatic duct. Postoperative complications, particularly delayed gastric emptying and pancreatic fistula, were diagnosed according to the criteria of Yeo et al [21]. Pancreatic fistula was diagnosed either when more than 50 mL of drainage fluid with an amylase concentration more than threefold higher than the normal upper limit in serum was collected on postoperative day 10 or later, or
121
when pancreatic anastomotic disruption was demonstrated with radiographic imaging. Second analysis: patients with ductal adenocarcinoma of the pancreatic head Eighty-seven patients (50 men and 37 women) had ductal adenocarcinoma of the pancreatic head. Their mean age was 62.9 ⫾ 1.2 years (range 27 to 84). Among these 87 patients, 53 (60.9%) underwent PD with the standard Whipple procedure, 1 (1.1%) underwent pylorus-preserving PD, and 12 (13.8%) underwent TP, all without vascular resection. The remaining 21 patients (24.l%) underwent PD or TP with SM-PVR according to our institutional criteria, as described below. The clinicopathological stages [22] were stage 0 in 1 patient, stage I in 7 patients, stage II in 17 patients, stage III in 34 patients, stage IVa in 19 patients, and stage IVb in 9 patients. Among the 19 patients presenting with stage IVa, 14 underwent SM-PVR with histologically proven invasion of the venous wall. The other 5, belonging to the no SMPVR group, had tumoral deposit along the SM-PV, and hence SM-PVR was not performed because additional deposits were found around the celiac axis and the SMA. Therefore, the staging (stage IVa) of these 5 patients was defined according to the operative findings and classified as R2 operations. Among the 9 patients with stage IVb, 3 had one resectable peritoneal nodule, 5 had one resectable liver metastase, and 1 had paraaortic positive lymph nodes that were not detected intraoperatively with intraoperative frozen examination. There were 3 patients with stage IVb in the SM-PVR group. Two of them presented with a liver metastasis and another had para-aortic lymph node metastases not detected with intraoperative frozen examination. Pathological lymph node metastasis was detected in 53 (60.9%) of the 87 patients. A margin-negative resection was performed in 61 patients (70.1%). Postoperative adjuvant chemoradiation therapy was administered to 52 of 87 patients, including 12 (57.1%) of 21 patients who underwent SM-PVR and 40 (60.6%) of 66 patients who did not. Criteria and methods for performing SM-PVR and lymphadenectomy Spiral computed tomography (Somatom Plus, Siemens, Erlanger, Germany) was used since 1991, and preoperative abdominal angiography was routinely performed. A standardized preoperative imaging including spiral computed tomography, ultrasonography, abdominal angiography, and magnetic resonance cholangiopancreatography was routinely applied to all patients since 1995. Endoscopic retrograde cholangiopancreatography and endoscopic ultrasonography were selectively performed. To be considered preoperatively, all patients were required to fulfill the following criteria for SM-PVR: (1) absence of extrapancreatic
122
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129
Fig. 1. Intraoperative views. A. Pancreatic head tumor and involvement of the portal vein. B. En-bloc pancreaticoduodenectomy with a segmental portal venous resection. C. Reconstruction of the portal vein. PV ⫽ portal vein; SpV ⫽ splenic vein; SMV ⫽ superior mesenteric vein; T ⫽ tumor; P ⫽ remnant pancreas; 3 ⫽ venous anastomosis.
disease, (2) no evidence of involvement of the superior mesenteric artery (SMA) and celiac axis, and (3) a patent SM-PV [15]. Total obstruction of the SM-PV with collateral veins shown in the preoperative imaging was considered as a contraindication for SM-PVR, as shown by Nakao et al [23], but unilateral and bilateral narrowings of the SM-PV were accepted as an indication for SM-PVR. PD was selected as the standard procedure. However, TP with splenectomy was performed when examination of frozen sections revealed that pancreatic resection margin was positive for tumor or when the pancreatic body or tail was found to contain multifocal tumor. Patients were principally required to fulfill the following intraoperative criteria for SM-PVR: (1) absence of liver metastasis, (2) absence of peritoneal dissemination, (3) absence of lymph node in-
volvement of the para-aortic area or more distant regions, (4) absence of involvement of the SMA, common hepatic artery, and celiac axis, and (5) evidence of tumor adherence to the SM-PV. The 3 patients with stage IVb of the SMPVR group were highly selected cases. They underwent PD or TP with SM-PVR associated with resection of one liver metastasis. This aggressive approach appeared justified only because of the young age of these 3 patients (27, 42, and 45 years old). An en-bloc resection of the SM-PV associated with PD was mainly performed. All mobilizations needed for PD or TP were performed before SM-PVR. During lymphadenectomy the portal vein and hepatic artery in the hepatoduodenal ligament were dissected until the porta hepatis. In the infrapancreatic area, the SMV was dissected. Progressively
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129 Table 2 Reconstruction methods after SM-PVR in 15 patients undergoing PD Types of SM-PV involvement
Types of reconstructions
Number of patients
Involvement of PV Involvement of SM-PV confluence (ⱕ6 cm)
End-to-end anastomosis End-to-end anastomosis with splenic vein reimplantation SMV-PV resectionreconstruction with saphenous vein graft and splenic vein reimplantation End-to-end anastomosis Lateral venous suture
4 5
Involvement of SM-PV confluence (⬎6 cm)
Involvement of SMV Lateral involvement of SM-PV
1
3 2
Patients undergoing total pancreatectomy were excluded. PD ⫽ pancreaticoduodenectomy; PV ⫽ portal vein; SMV ⫽ superior mesenteric vein; SM-PV ⫽ superior mesentericoportal vein; SM-PVR ⫽ superior mesentericoportal vein resection.
the gastrocolic trunk and the pancreaticoduodenal collaterals were ligated. The first jejunal vein insertion at the right posteroinferior side of the SMV was identified. Anterior and posterior lymph nodes dissection was continued until identifying SMA and celiac axis. The uncinate process was freed from the SMA as much as possible before SM-PVR. After establishing that a full mobilization of the SM-PV could not be performed (Fig. 1), vascular clamps were carefully placed at the upper and lower vascular ends without any routine clamping of the SMA. SMA clamping was performed when the portal vein reconstruction could not be achieved in less than 60 minutes. The upper and lower ends of the SM-PV were cut, and an en-bloc SM-PVR associated with PD was completed. Vascular reconstruction was then performed with a continuous suture using 5-0 Prolene (Ethicon, Brussels, Belgium). An en-bloc SM-PVR associated with PD was performed in 25 cases; a SM-PV resection was performed after PD in 6 cases. Among the 31 patients undergoing SM-PVR, segmental resection of the SM-PV was performed in 27 patients and Table 3 Profile of 31 patients with SM-PVR and 119 patients without SM-PVR
Male/female Age (years) Hospital stay (days) Duration of operation (minutes) Perioperative blood transfusion requirement (units)
SM-PVR (n ⫽ 31)
No SM-PVR (n ⫽ 119)
P value
15/16 62.2 ⫾ 2.0 22.2 ⫾ 1.4 501.2 ⫾ 17.1
71/48 61.2 ⫾ 1.2 25.9 ⫾ 1.1 412.7 ⫾ 6.5
0.354* 0.663† 0.111† ⬍0.0001†
2.0 ⫾ 0.2
⬍0.0001†
5.6 ⫾ 0.8
* Chi-square test. † Unpaired Student’s t test. SM-PVR ⫽ superior mesentericoportal vein resection.
123
lateral venous resection was achieved in 4 patients. Segmental resection of the PV was performed in 7 patients, of the confluence of the SM-PV in 13 patients, and of the SMV in 7 patients. The portal tract was reconstructed with a direct end-to-end anastomosis in 25 patients. Such venous reconstruction with a end-to-end anastomosis could be obtained, in our series, with as much as 6-cm venous segment resection. Larger venous resection needed a SMV-PV reconstructive bypass with a saphenous vein graft in 1 patient and a Gore-tex (Medical Products, Flagstaff, AZ) graft in another. The portal tract was reconstructed without use of a portosystemic shunt in all patients but 1, in whom a portosystemic shunt as described by Nakao et al [24] was used. This shunt was used because of occurrence of splanchnic venous congestion during clamping. A simple venous suture was performed in the 4 patients undergoing lateral venous resection. In patients undergoing PD with SM-PVR (n ⫽ 15), resection of the SM-PV confluence was performed in 6 patients. The splenic vein was reconstructed in all 6 patients. Methods of venous resection-reconstruction in patients undergoing PD with SM-PVR (n ⫽ 15) are shown in Table 2. Regional lymph nodes as defined by the Japanese classification [22], ie, nodes around the common hepatic artery, nodes around the celiac trunk, nodes in the hepatoduodenal ligament, and posterior and anterior pancreaticoduodenal nodes, were routinely dissected in patients with periampullary malignancies. When intraoperative pathological diagnosis showed involvement of these lymph nodes, a complete dissection of lymph nodes around the SMA and a lymph node sampling of the para-aortic area were performed. An extended lymph node dissection around the aorta as reported by Yeo et al [25] was not performed in the present study. Histopathological examination According to the Japanese histopathological classification of curative resection, the resection margin (pancreatic resection margin, resection margin of the bile duct, and peritumoral margin) should be 5 mm free from the cut surface (no exposure of the tumor at the cut surface). All pathologic specimens were re-reviewed by a pathologist to determine the presence or absence of tumor deposit on surgical margins (pancreatic transection line, bile duct transection line, duodenum, stomach, and peripancreatic soft tissue), as well as extent of local invasiveness (invasion into duodenum, bile duct, or peripancreatic soft tissue), histologic differentiation, lymph node metastases, and involvement of the SM-PVR. Statistical methods Data are expressed as mean ⫾ standard error of mean. All statistical analyses, including Student’s t test, the chi-
124
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129
Table 4 Postoperative complications in patients undergoing resection with or without SM-PVR Postoperative complications*
Sepsis Urinary infection Wound infection Pulmonary complications Pulmonary embolism Thrombophlebitis Pancreatic fistula Postoperative hemorrhage from operative field Gastric ulcer Liver ischemia Neurological complications Biliary stenosis Intraabdominal abscess Superior mesenteric venous infarction Intestinal occlusion Delayed gastric emptying Total number of patients with complications Number of patients undergoing reoperation
All patients
Patients with pancreatic ductal adenocarcinoma
SM-PVR (n ⫽ 31)
No SM-PVR (n ⫽ 119)
SM-PVR (n ⫽ 21)
No SM-PVR (n ⫽ 66)
5 0 1 5 1 3 1 (1/15 ⫽ 6.7%) 0 0 0 1 0 1 1 0 5 15 (48.4%) 3 (9.7%)
9 9 2 10 6 3 14 (14/101 ⫽ 13.9%)† 9 4 1 4 1 8 0 2 14 56 (47.1%)‡ 15 (12.6%)
3 0 0 3 1 1 0 0 0 0 0 0 0 1 0 2 8 (38.1%) 2 (9.5%)
1 6 0 5 0 1 4 5 3 0 3 1 2 0 1 7 30 (45.5%)†† 10 (15.2%)
* All listed postoperative complications are shown when a patient suffered multiple complications. † P ⫽ 0.44, chi-square test. ‡ P ⫽ 0.77, chi-square test. †† P ⫽ 0.55, chi-square test. SM-PVR ⫽ superior mesentericoportal vein resection.
square test, survival analysis, and stepwise multifactorial regression analysis, were performed with the computer software package Statview IV (Abacus Concepts, Inc., Berkeley, California). Rates of cumulative survival were calculated with the Kaplan-Meier method, and differences were analyzed with the log-rank test. A difference was considered significant when the P value was less than 0.05. In the stepwise multifactorial regression analysis, data were considered significant when the F test value was more than 4.0.
The length of hospitalization did not differ significantly between the two groups (Table 3). The rate of postoperative complications did not differ significantly (P ⫽ 0.77) between patients undergoing SMPVR (48.4%) and those (47.1%) not undergoing vascular resection (Table 4). However, the frequency of pancreatic fistula among patients undergoing PD was slightly lower, Table 5 Postoperative complications necessitating relaparotomy
Results
Indications for reoperation
SM-PVR (n ⫽ 31)
First analysis: early surgical outcomes in all patients undergoing PD
Intraabdominal abscess Superior mesenteric venous infarction Operative field hemorrhage due to pancreatic fistula Other operative field hemorrhage* Digestive hemorrhage Pancreatic fistula Stenosis or leakage of gastrojejunostomy Liver ischemia due to celiac axis compression Stenosis of hepaticojejunostomy Total
2 1
0 0
0
3
0 0 0 0
2 1 3 4
0
1
In the 150 patients, the mean operation time was 428.2 ⫾ 6.6 minutes, and the volume of red blood cells transfused perioperatively was 2.7 ⫾ 0.2 units. Following the criteria of Yeo et al [21], 71 patients (47.3%) were judged to have developed postoperative complications, and 4 patients (2.7%) died within 30 days after surgery. The mean hospital stay was 25.2 ⫾ 0.9 days. Among the 31 patients who underwent SM-PVR, PD was performed in 15 patients and TP in 16 patients. Operation time was significantly longer and perioperative blood transfusion was performed significantly more in patients undergoing SM-PVR (Table 3). Portal flow was clamped for 25.1 ⫾ 2.2 minutes in patients undergoing SM-PVR.
0 3 (9.7%)
No SM-PVR (n ⫽ 119)
1 15 (12.6%)†
* The other causes of the operative field hemorrhage include postoperative bleeding and a hemorrhage caused by biliary leakage. † P ⫽ 0.66, chi-square test. SM-PVR ⫽ superior mesentericoportal vein resection.
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129
125
Table 6 Profile of patients with ductal adenocarcinoma of the pancreatic head undergoing SM-PVR and those not undergoing SM-PVR SM-PVR (n ⫽ 21) Male/female Age (years) Operative method Total pancreataectomy Pancreaticoduodenectomy Method of pancreatic anastomosis Pancreaticogastrostomy Pancreaticojejunostomy Duration of operation (min) Perioperative red blood cell transfusion requirement (units) Tumor size (mm) Histopathological type Well-differentiated ductal adenocarcinoma Moderately differentiated ductal adenocarcinoma Poorly differentiated ductal adenocarcinoma Histopathological stage (22) 0 I II III IVa IVb Carcinoembryonic antigen (CEA) CA19-9 Clinicopathological curability Margin-negative resection Margin-positive resection Hospital stay after surgery (days) Postoperative complication Absent Present
No SM-PVR (n ⫽ 66)
P value
10/11 60.4 ⫾ 2.5
39/27 62.7 ⫾ 1.3
0.356* 0.405†
10 11
12 54
0.007*
5 6 503.1 ⫾ 18.4 5.5 ⫾ 1.0 47.9 ⫾ 3.8 5 15 1
27 27 410.4 ⫾ 8.0 1.8 ⫾ 0.2 41.0 ⫾ 2.5 22 36 8
0.741* ⬍0.0001† ⬍0.0001† 0.164† 0.141*
0 0 5 4 9 3 7.0 ⫾ 2.0 2231.9 ⫾ 807.1
1 7 12 30 10 6 10.8 ⫾ 1.8 1182.6 ⫾ 408.9
0.117*
13 8 20.3 ⫾ 1.4
48 18 23.9 ⫾ 1.2
0.345*
13 8
36 30
0.314† 0.314†
0.112† 0.554*
* Chi-square test. † Unpaired Student t test. SM-PVR ⫽ superior mesentericoportal vein resection.
but not significantly so, in patients undergoing SM-PVR (1 of 15 ⫽ 6.7%, rate of normal texture of pancreas ⫽ 48.4%) than in those not undergoing vascular resection (14 of 101 ⫽ 13.9%, P ⫽ 0.44, rate of normal texture of pancreas ⫽ 54.6%, Table 4). The frequency of a second laparotomy to treat postoperative complications did not differ between the two groups (9.7% of patients undergoing SMPVR and 12.6% of patients not undergoing vascular resection, P ⫽ 0.46). However, no patient required reoperation for pancreatic fistula or bleeding complications in patients undergoing SM-PVR, whereas 9 patients required reoperation owing to these complications among patients not undergoing SM-PVR (Table 5). Among patients undergoing SM-PVR, 1 died of mesenteric venous infarction 4 days after surgery. The venous reconstruction was performed without any difficulties or a kinking, and the cause of venous thrombosis was not clear. The mortality rate did not differ between patients undergoing SM-PVR (3.2%) and patients not undergoing vascular resection (2.5%). In addition, no postoperative death occurred in patients undergoing SM-PVR in the recent series since 1994 (n ⫽ 24).
Stepwise regression multivariate analysis showed that normal parenchyma of the remnant pancreas was the only significant risk factor for postoperative complications (F value ⫽ 5.4); SM-PVR itself was not a risk factor (F value ⫽ 0.1). Second analysis: survival and other outcomes of patients with ductal adenocarcinoma of the pancreatic head Among the 87 patients with ductal adenocarcinoma of the pancreatic head, duration of surgery was 431.2 ⫾ 8.7 minutes, and the volume of perioperative red blood cell transfusion was 2.6 ⫾ 0.3 units. Thirty-eight patients (43.6%) had postoperative complications, and 2 patients (2.3%) died within 30 days after surgery. Mean hospital stay was 23.1 ⫾ 1.0 days. The male-to-female ratio, age, type of surgery, clinical stage, and serum concentrations of preoperative carcinoembryonic antigen, and CA19-9 did not differ between patients undergoing SM-PVR and those not undergoing vascular resection (Table 6). The rate of the margin-negative resection was slightly but not significantly higher among patients
126
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129
Fig. 2. Cumulative survival curves of patients with ductal adenocarcinoma of the pancreatic head undergoing pancreaticoduodenectomy with or without superior mesentericoportal venous resection (SM-PVR). Survival of patients undergoing SM-PVR (n ⫽ 21, solid line) is similar to that of patients not undergoing SM-PVR (n ⫽ 66, dashed line, P ⫽ 0.48). Oneand two-year survival rates are 53.6% and 21.5% in patients undergoing SM-PVR and 51.1% and 24.3% in patients not undergoing SM-PVR.
not undergoing vascular resection (72.7%, 48 of 66 patients) than among those undergoing SM-PVR (61.9%, 13 of 21 patients, P ⫽ 0.35). Histopathologic examination of specimens from the 21 patients with suspected venous infiltration by cancer showed intimal infiltration in 7, medial infiltration in 4, and adventitial infiltration in 3. Therefore, one third of patients (n ⫽ 7) undergoing SM-PVR did not have histological involvement of the SM-PV. However, the involvement had
been grossly suspected because of close adhesion between tumor and venous wall. The rate of postoperative complications was slightly but not significantly lower in patients undergoing SM-PVR (38.1%) than in patients not undergoing vascular resection (45.5%, Table 4). Hospitalization after surgery was slightly but not significantly shorter in patients undergoing SM-PVR (20.3 ⫾ 1.4 days) than in those not undergoing vascular resection (23.9 ⫾ 1.2 days; P ⫽ 0.11). Among the 21 patients undergoing SM-PVR, 1 (4.7%) died of mesenteric venous infarction 4 days after surgery. One patient (1.5%) died of erosive hemorrhage caused by pancreatic fistula 7 days after surgery among the 66 patients not undergoing vascular resection. The mean follow-up period was 12.1 months (range 1 to 48). Ten of the 21 patients undergoing SM-PVR and 40 of the 66 patients not undergoing vascular resection died of cancer progression. Of the 2 patients undergoing SM-PVR and arterial resection, 1 is still alive 5 months after surgery and the other died of cancer recurrence 6 months after surgery. Median survival of patients (including those who died perioperatively) undergoing SM-PVR and of those not undergoing SM-PVR did not differ significantly (369 days versus 374 days, P ⫽ 0.48, Fig. 2). One-year and 2-year survival rates were 53.6% and 21.5% in patients undergoing SM-PVR and 51.1% and 24.3% in patients not undergoing vascular resection. Among patients undergoing SM-PVR, survival rate did not significantly differ between patients with histological venous infiltration (n ⫽ 14) and those without infiltration (n ⫽ 7; 2-year survival 34.9% versus 0%). Among patients who underwent a histologically marginnegative resection, median survival did not differ between patients undergoing SM-PVR (503 days) and those not undergoing resection (417 days; P ⫽ 0.141). Among the 13 patients undergoing SM-PVR and a margin-negative resection, 6 (46.2%) had recurrent disease, which was locoregional in each case. Survival rate for patients undergoing a margin-negative resection (n ⫽ 13) was significantly higher than that for patients undergoing macroscopic or microscopic margin positive resection (n ⫽ 8; 2-year survival, 57.1% versus 0%, P ⫽ 0.042, Fig. 3).
Comments
Fig. 3. Cumulative survival curves of patients undergoing a margin-negative resection (R0) and of those undergoing a macroscopically (R1) or microscopically (R2) margin-positive resection among patients undergoing superior mesentericoportal venous resection (SM-PVR). Survival rate is significantly higher for patients undergoing a margin-negative resection (n ⫽ 13, solid line) than for those undergoing a margin-positive resection (n ⫽ 8, dashed line, 2-year survival 57.1% versus 0%, P ⫽ 0.042).
Our results showed that PD with SM-PVR can be safely performed in selected cases. Duration of operation was longer and perioperative blood transfusion requirement was greater in patients undergoing SM-PVR than in patients not undergoing SM-PVR. However, postoperative morbidity, mortality, and hospitalization after SM-PVR did not differ between the two groups of patients. In addition, no postoperative death occurred in patients undergoing SM-PVR in the recent series, since 1994, in contrast to 1 death before 1994, suggesting a learning curve effect.
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129
127
Table 7 Reported rates of mortality, morbidity, survival, and margin-negative resection in pancreaticoduodenectomy with SM-PVR Authors (references)
Institutions and city names
Sindelar et al [9]
National Cancer Institute, Bethesda Mannheim University, Mannheim Academic Medical Centre, Amsterdam Memorial Sloan-Kettering Cancer Center, New York Munich University, Munich M.D. Anderson Cancer Center, Houston Pisa University, Pisa Tokyo Women’s Medical College, Tokyo Mie University, Tsu Kanazawa University, Kanazawa Nagoya University, Nagoya Keio University, Tokyo
Mortality (within 1 month) (%)
Morbidity (%)
Median survival (months)
Marginnegative resection (%)
20
20
55
12
—
60 (10.7%)* 20 (11.4%)
6.7 15
— 55
12 8
— 15
58 (17.5%)
5
12§
13
73
31 (10.4%)* 31 (41,3%)
0 0
41.9 30
8 22
32 87
27 (12.2)* 172 (69.0%)
7.2 5
42.1 23
— —
— —
30 (41.7%) 72 (50.0%) 104 (78%) 107 (55.7%)
1.0 6.9 8 3.8
15.7 37.2 — 23.1
26.5 — 49
37 (19.6%)
13.5
35.1§
Ishikawa et al [33,34,36]
Hannover Medical School, Hannover Osaka Medical Center, Osaka
19 — 5-year survival, 14%‡ 9.0
35 (70.0%)
5.7
—
Present study
Hautepierre Hospital, Strasbourg
31 (20.7%)*
3.2
3-year survival, 29% 13
Trede et al [4,32] Allema et al [11] Fortner et al [12] and Harrison et al [13] Roder et al [14] Leach et al [15] and Fuhrman et al [16] Mosca et al [37] Imaizumi et al [38] Naganuma et al [39] Nagakawa et al [31] Nakao et al [23,24]† Takahashi et al [17] and Ogata et al [26] Klempnauer et al [35]
Number of patients with SM-PVR and frequency
— 48.4
—
61.9
* Periampullary cancers other than pancreatic cancer were included. † Data of all patients were not documented. ‡ Data of patients with curative resection are shown. § Relapalotomy rate is demonstrated. SM-PVR ⫽ superior mesentericoportal vein resection.
In patients with ductal adenocarcinoma of the pancreatic head, PD or TP with SM-PVR seems justified in order to achieve a R0 resection when a close adhesion between the tumor and the venous wall is the only obstacle for resection. In fact, mortality, mobidity, and survival rates did not differ between patients with SM-PVR and those without SMPVR. In addition, as expected, the present study showed that survival of patients undergoing a margin-negative resection (R0) with a SM-PVR is significantly better than that of patients with microscopic (R1) or macroscopic (R2) margin-positive resection. However, this survival benefit should be carefully considered because the protocols of chemoradiation therapy were not uniformly performed in all patients with pancreatic ductal carcinoma. Currently, postoperative mortality after PD has decreased, ranging from 0% to 8% [2–5]. Surgeons seem to focus on pancreatectomy with vascular resection to obtain a margin-negative resection (R0) [4,8 –19,26 –29]. Surgical outcomes of pancreatectomy combined with vascular resection in patients with pancreatic cancer and periampullary diseases have recently been investigated. Yeo et al [30] reported that SM-PVR was performed in only 4% (26 of 650) of patients undergoing PD, a lower rate than the 20.7% in the present study. In addition, our review of the literature found only 20 studies of PD from 14 institutions that in-
cluded 20 or more patients who underwent SM-PVR (Table 7) [4,9,11–17,23,24,26,31–39]. Among these reports from the 14 institutions, only 6 reported mortality, morbidity, rate of margin-negative resection, and long-term survival. Therefore, a study examining these outcomes is still needed to determine the most appropriate management of tumors involving retropancreatic vessels. In these reports, mortality rate ranged from 0% to 20% and morbidity rate ranged from 12% to 55%. The present results therefore showed acceptable rates of mortality (3.2%) and morbidity (48.4%). Since 1994 we have principally performed PD with pancreaticogastrostomy. In the present study, the splenic vein was reimplanted to the SMV or SM-PV bypass in 6 patients. Ligation of the splenic vein is a simple therapeutic option. However, left-sided portal hypertension can occur with increased rate of gastrointestinal bleeding even from the pancreatic transection line or from the gastric mucosa (“hypertensive gastropathy”). In our practice we recommend, when it is possible, reimplantation of the splenic vein to decrease the risk of side effects of left-sided portal hypertension. Whether survival after SM-PVR is better than that of patients not undergoing SM-PVR should be discussed. Studies from the United States and Europe have reported median survival ranging from 8 to 22 months in patients with pancreatic cancer who had undergone pancreatic re-
128
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129
section with SM-PVR (Table 7) [4,9,11–16,32]. In all reports survival did not differ between patients who did and did not undergo SM-PVR. The present study also showed the same result. It is therefore concluded that SM-PVR itself does not improve survival in patients with pancreatic cancer. Whether survival of patients after SM-PVR is better than that of patients in whom margin-positive resection is performed because SM-PV involvement is not resected should be discussed. Allema et al [11] reported no difference in survival between patients undergoing SM-PVR and those in whom a positive margin at the SM-PV was not resected; they concluded that pancreatic resection with SM-PVR does not improve survival. However, their conclusion may be qualified because the rate of margin-negative resection was only 15% in patients undergoing SM-PVR. Several Japanese authors have reported higher survival rates after pancreatectomy with SM-PVR for patients with pancreatic cancer [17,26,27,31,33]. Takahashi et al [17] have reported a 5-year survival rate of 14% in patients undergoing marginnegative resection, with 6 patients surviving more than 5 years after SM-PVR. Ishikawa et al [33,36] have reported a 3-year survival rate of 29%, with 7 out of 35 patients surviving more than 3 years after SM-PVR. The present study confirmed that survival after SM-PVR in the patients undergoing a margin-negative resection is better than in those not undergoing a margin-negative resection (2-year survival, 57.1% versus 0%). In the present study one third of patients undergoing SM-PVR did not have histological involvement of SM-PV. However, Ishikawa et al [40] recently showed that intraoperative cytology showed positive SM-PV involvement in several patients who underwent PD without evidence of macroscopic involvement of the SM-PV, and that further SM-PVR after PD confirmed cancer cell infiltration of the SM-PV in most of them. These results suggest difficulty for intraoperatively confirming SM-PV involvement. In conclusion, the results of the present study show that combined PD and SM-PVR can be performed safely. However, SM-PVR should be considered for selected patients and performed only in selected institutions. Greater blood loss and longer operative time cannot be ignored, and the safety of this method has to be confirmed by a sufficient number of institutions. Moreover, there is probably a learning curve effect, as we experienced. On the basis of our findings with regard to postoperative morbidity, mortality, rate of margin-negative resection, and survival of patients with pancreatic adenocarcinoma, SM-PVR should be performed only when a margin-negative resection is expected to be achieved. References [1] Nitecki SS, Sarr MG, Colby TV, van Heerden JA. Long-term survival after resection for ductal adenocarcinoma of the pancreas: is it really improving? Ann Surg 1995;221:59 – 66.
[2] Wade TP, El-Ghazzawy AG, Virgo KS, Johnson FE. The Whipple resection for cancer in U.S. Department of Veterans Affairs Hospitals. Ann Surg 1995;221:241– 8. [3] Warshaw AL, Fernandez-del Castillo C. Pancreatic carcinoma. N Engl J Med 1992;326:455– 65. [4] Trede M, Schwall G, Saeger HD. Survival after pancreatoduodenectomy. Ann Surg 1990;211:447–58. [5] Cameron JL, Pitt HA, Yeo CJ, et al. One hundred and forty-five consecutive pancreaticoduodenectomies without mortality. Ann Surg 1993;217:430 – 8. [6] Klempnauer J, Ridder GJ, Bektas H, Pichlmayer R. Surgery for exocrine pancreatic cancer—who are the 5- and 10-year survivors? Oncology 1995;52:353–9. [7] Willet CG, Lewandrowski K, Warshaw AL, Efird J, Compton CC. Resection margins in carcinoma of the head of the pancreas. Implications for radiation therapy. Ann Surg 1993;217:144 – 8. [8] Yeo CJ, Cameron JL, Lillemoe KD, et al. Pancreaticoduodenectomy for cancer of the head of the pancreas: 201 patients. Ann Surg 1995;221:721–33. [9] Sindelar WF. Clinical experience with regional pancreatectomy for adenocarcinoma of the pancreas. Arch Surg 1989;124:127–32. [10] Launois B, Stasik C, Bardaxoglou E, et al. Who benefits from portal vein resection during pancreaticoduodenectomy for pancreatic cancer? World J Surg 1999;23:926 –9. [11] Allema JH, Reinders ME, van Guilik TM, et al. Portal vein resection in patients undergoing pancreatoduodenectomy for carcinoma of the pancreatic head. Br J Surg 1994;81:1642– 6. [12] Fortner JG, Klimstra DS, Scnie RT, Maclean BJ. Tumor size is the primary prognosticator for pancreatic cancer after regional pancreatectomy. Ann Surg 1996;223:147–53. [13] Harrison LE, Klimstra DS, Brennan MF. Isolated portal vein involvement in pancreatic adenocarcinoma. A contraindication for resection? Ann Surg 1996;224:342–9. [14] Roder JD, Stein HJ, Siewert JR. Carcinoma of the periampullary region: who benefits from portal vein resection? Am J Surg 1996; 171:170 –5. [15] Leach SD, Lee JE, Charnsangavej C, et al. Survival following pancreaticoduodenectomy with resection of the superior mesenteric-portal vein confluence for adenocarcinoma of the pancreatic head. Br J Surg 1998;85:611–17. [16] Fuhrman GM, Leach SD, Staley CA, et al. Rationale for en bloc vein resection in the treatment of pancreatic adenocarcinoma adherent to the superior mesenteric-portal vein confluence. Ann Surg 1996;223: 154 – 62. [17] Takahashi S, Ogata Y, Tsuzuki T. Combined resection of the pancreas and portal vein for pancreatic cancer. Br J Surg 1994;81: 1190 –3. [18] Ishikawa O, Ohigashi H, Imaoka S, et al. Preoperative indications for extended pancreatectomy for locally advanced pancreas cancer involving the portal vein. Ann Surg 1992;215:231– 6. [19] Tashiro S, Uchino R, Hiraoka T, et al. Surgical indication and significance of portal vein resection in biliary and pancreatic cancer. Surgery 1991;109:481–7. [20] Delcore R, Thomas JH, Pierce GE, Hermreck AS. Pancreatogastrostomy: a safe drainage procedure after pancreatoduodenectomy. Surgery 1990;108:641–7. [21] Yeo CJ, Cameron JL, Maher MM, et al. A prospective randomized trial of pancreaticogastrostomy versus pancreaticojejunostomy after pancreaticoduodenectomy. Ann Surg 1995;222:580 –92. [22] Kayahara M, Nagakawa T, Ueno K, et al. Surgical strategy for carcinoma of the pancreas head area based on clinicopathologic analysis of nodal involvement and plexus invasion. Surgery 1995; 117:616 –23. [23] Nakao A, Harada A, Nonami T, et al. Clinical significance of portal invasion by pancreatic head carcinoma. Surgery 1995;117:50 –5.
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129 [24] Nakao A, Takagi H. Isolated pancreatectomy for pancreatic head carcinoma using catheter bypass of the portal vein. Hepatogastroenterology 1993;40:426 –9. [25] Yeo CJ, Cameron JL, Sohn TA, et al. Pancreaticoduodenectomy with or without extended retroperitoneal lymphadenectomy for periampullary adenocarcinoma. Comparison of morbidity and mortality and short-term outcome. Ann Surg 1999;229:613–24. [26] Ogata Y, Hishinuma S, Takahashi S, et al. Indication and results of pancreatectomy with combined resection of vessels for adenocarcinoma of the pancreas. Nippon Geka Gakkai Zasshi 1997;98:615– 21. [27] Nakao A, Harada A, Nonami T, et al. Regional vascular resection using catheter bypass procedure for pancreatic cancer. Hepatogastroenterology 1995;42:734 –9. [28] Luttges J, Vogel I, Menke M, et al. The retroperitoneal resection margin and vessel involvement are important factors determining survival after pancreaticoduodenectomy for ductal adenocarcinoma of the head of the pancreas. Virchows Arch 1998;433:237– 42. [29] Civello IM, Frontera D, Viola G, et al. Extensive resection in pancreatic cancer: review of the literature and personal experience. Hepatogastroenterology 1998;45:1877– 83. [30] Yeo CJ, Cameron JL, Sohn TA, et al. Six hundred fifty consecutive pancreaticoduodenectomies in the 1990s: pathology, complications, and outcomes. Ann Surg 1997;226:248 – 60. [31] Nagakawa T, Konishi I, Ueno K, et al. Extended radical pancreatectomy for carcinoma of the head of the pancreas. Hepatogastroenterology 1998;45:849 –54.
129
[32] Trede M. Vascular problems and techniques associated with pancreatectomy and regional pancreatectomy. In: Trede M, Carter DC, editors. Surgery of the pancreas. 2nd ed. New York: Churchill Livingstone, 1997, p 505–16. [33] Ishikawa O, Ohhigashi H, Sasaki Y, et al. Practical usefulness of lymphatic and connective tissue clearance for the carcinoma of the pancreas head. Ann Surg 1988;208:215–20. [34] Ishikawa O. Surgical technique, curability and postoperative quality of life in an extended pancreatectomy for adenocarcinoma of the pancreas. Hepatogastroenterology 1996;43:320 –5. [35] Klempnauer J, Ridder GJ, Bektas H, Pichlmayr R. Extended resections of ductal pancreatic cancer—impact on operative risk and prognosis. Oncology 1996;53:47–53. [36] Ishikawa O, Ohhigashi H, Sasaki Y, et al. Extended pancreatectomy and liver perfusion chemotherapy for resectable adenocarcinoma of the pancreas. Digestion 1999;60(suppl 1):135– 8. [37] Mosca F, Giulianotti PC, Balestracci T, et al. Long-term survival in pancreatic cancer: pylorus-preserving versus Whipple pancreatoduodenectomy. Surgery 1997;122:553– 66. [38] Imaizumi T, Hanyu F, Harada N, et al. Extended radical Whipple resection for cancer of the pancreatic head: operative procedure and results. Dig Surg 1998;15:299 –307. [39] Naganuma T, Isaji S, Kawarada Y. Staging and extended resection for pancreatic cancer. Pancreas 1998;16:355– 62. [40] Ishikawa O, Ohhigashi H, Sasaki Y, et al. Intraoperative cytodiagnosis for detecting a minute invasion of the portal vein during pancreaticoduodenectomy for adenocarcinoma of the pancreatic head. Am J Surg 1998;175:477– 81.
Is pancreaticoduodenectomy with mesentericoportal venous resection safe and worthwhile? Philippe Bachellier, M.D., Hiroshi Nakano, M.D., Ph.D., Elie Oussoultzoglou, M.D., Jean-Christophe Weber, M.D., Karim Boudjema, M.D., Ph.D., Philippe Wolf, M.D., Ph.D., Daniel Jaeck, M.D., Ph.D., F.R.C.S.* Centre de Chirurgie Visce´rale et de Transplantation, Hoˆpital Universitaire de Hautepierre, Avenue Molie`re, 67098 Strasbourg Cedex, France. Manuscript received November 10, 2000; revised manuscript May 14, 2001
Abstract Background: Whether or not superior mesentericoportal venous resection (SM-PVR) associated with pancreaticoduodenectomy (PD) is safe and worthwhile has not been fully confirmed. The aim of the present study was to investigate results of this surgical procedure performed for pancreatic head and periampullary neoplasms. Methods: As a first analysis, postoperative morbidity and mortality after PD with (n ⫽ 31) or without SM-PVR (n ⫽ 119) were investigated in 150 patients with pancreatic head and periampullary neoplasms. As a second analysis, rates of margin-negative resection and survival after SM-PVR (n ⫽ 21) and without SM-PVR (n ⫽ 66) were compared in 87 patients with pancreatic ductal adenocarcinoma of the pancreatic head. In these patients undergoing SM-PVR (n ⫽ 21), survival rate was investigated in patients who did (n ⫽ 13) and did not (n ⫽ 8) undergo a margin-negative resection. Results: In the first analysis, duration of surgery and volume of blood transfused perioperatively were higher in patients undergoing SM-PVR. However, mortality, morbidity rates, and mean hospital stay did not differ between patients who did undergo SM-PVR (31 patients, 3.2%, 48.4%, and 22.2 days, respectively) and who did not (119 patients, 2.5%, 47.1%, 25.9 days, respectively). No postoperative death occurred in the recent part of the present study, since 1994, in patients undergoing SM-PVR. In the second analysis of pancreatic ductal adenocarcinoma, rates of margin-negative resection and 2-year survival did not significantly differ between patients who did and did not undergo SM-PVR (62% and 22%, respectively, versus 73% and 24%). In patients undergoing SM-PVR, survival rate was significantly higher for patients undergoing a margin-negative resection (n ⫽ 13) than for patients undergoing a macroscopic or microscopic margin-positive resection (n ⫽ 8, 2-year survival ⫽ 57.1% versus 0%, P ⬍0.05). Conclusion: PD combined with SM-PVR can be performed safely. This surgical procedure is followed by a promising survival rate and can be recommended in order to obtain a margin-negative resection; however, candidates for SM-PVR should be carefully selected. © 2001 Excerpta Medica, Inc. All rights reserved. Keywords: Pancreatic cancer; Pancreaticoduodenectomy; Mesentericoportal venous resection
Pancreaticoduodenectomy (PD) is the only potential cure for ductal adenocarcinoma of the pancreatic head or periampullary tumors [1– 4]. Postoperative mortality rates reported currently for PD range from 0% to 8% [2–5]. In addition, 5-year survival rates in patients with pancreatic ductal adenocarcinoma undergoing pancreatic resection ranged from 7% to 25% [1– 4]. However, survival rates after pancreatic resection remain very poor when the surgical margin is positive for cancer [6 – 8]. In an attempt to obtain * Corresponding author. Tel.: ⫹33-388-12-7258; fax: ⫹33-388-127286. E-mail address: [email protected]
a negative surgical margin, the superior mesenteric vein (SMV), superior mesenteric-portal vein (SM-PV) confluence, portal vein (PV), and the retropancreatic arteries are sometimes resected [4,8 –19]. Morbidity rates after this additional superior mesentericoportal venous resection (SMPVR) range from 26% to 55% [11,14,16,17]. However, SM-PVR has not been universally accepted because it has yet to be made clear that this procedure can be performed with acceptable morbidity and mortality even if it increases the rate of the margin-negative resection [11,13–15]. In addition, whether SM-PVR improves long-term survival has not been fully confirmed. In the present study, we investigated morbidity and mor-
0002-9610/01/$ – see front matter © 2001 Excerpta Medica, Inc. All rights reserved. PII: S 0 0 0 2 - 9 6 1 0 ( 0 1 ) 0 0 6 8 6 - 9
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129 Table 1 Indications for pancreaticoduodenectomy with or without SM-PVR
Pancreatic ductal adenocarcinoma Cholangiocarcinoma Neuroendocrine carcinoma Pancreatic cystic neoplasm Ampullary adenocarcinoma Ampullary adenoma Duodenal adenocarcinoma
PD or TP without vascular resection (n ⫽ 119)
PD or TP with SM-PVR (n ⫽ 31)
Total (n ⫽ 150)
66
21
87
11 8
2 6
13 14
12
1
13
15
1
16
4 3
0 0
4 3
PD ⫽ pancreaticoduodenectomy; TP ⫽ total pancreatectomy; SMPVR ⫽ superior mesentericoportal vein resection.
tality after PD with and without SM-PVR in patients with ductal adenocarcinoma of the pancreatic head and periampullary neoplasms. In addition, in an attempt to determine whether this procedure is justified in the treatment of malignant diseases, we investigated the rates of the marginnegative resection and cumulative survival after PD associated with SM-PVR in patients with ductal adenocarcinoma of the pancreatic head.
Patients and methods First analysis: all patients undergoing pancreatectomy From January 1990 through March 1999, 187 patients underwent PD or total pancreatectomy (TP) in our department. One hundred fifty patients (86 men and 64 women) with ductal adenocarcinoma of the pancreatic head or periampullary neoplasms were selected for the current analysis among these 187 patients (Table 1). The mean age of the patients was 61.4 ⫾ 1.0 years (range 27 to 85). Among the 150 patients, 116 (77.3%) underwent PD (including 6 patients who underwent a pylorus-preserving procedure) and 34 (22.7%) underwent TP. Among the 116 patients undergoing PD, 52 underwent reconstruction with pancreaticojejunostomy from 1990 through 1994. Since 1994, 64 patients underwent pancreaticogastrostomy with the method of Delcore et al [20] in which 2 cm of pancreatic remnant is telescoped into the gastric lumen, without stenting in the main pancreatic duct. Postoperative complications, particularly delayed gastric emptying and pancreatic fistula, were diagnosed according to the criteria of Yeo et al [21]. Pancreatic fistula was diagnosed either when more than 50 mL of drainage fluid with an amylase concentration more than threefold higher than the normal upper limit in serum was collected on postoperative day 10 or later, or
121
when pancreatic anastomotic disruption was demonstrated with radiographic imaging. Second analysis: patients with ductal adenocarcinoma of the pancreatic head Eighty-seven patients (50 men and 37 women) had ductal adenocarcinoma of the pancreatic head. Their mean age was 62.9 ⫾ 1.2 years (range 27 to 84). Among these 87 patients, 53 (60.9%) underwent PD with the standard Whipple procedure, 1 (1.1%) underwent pylorus-preserving PD, and 12 (13.8%) underwent TP, all without vascular resection. The remaining 21 patients (24.l%) underwent PD or TP with SM-PVR according to our institutional criteria, as described below. The clinicopathological stages [22] were stage 0 in 1 patient, stage I in 7 patients, stage II in 17 patients, stage III in 34 patients, stage IVa in 19 patients, and stage IVb in 9 patients. Among the 19 patients presenting with stage IVa, 14 underwent SM-PVR with histologically proven invasion of the venous wall. The other 5, belonging to the no SMPVR group, had tumoral deposit along the SM-PV, and hence SM-PVR was not performed because additional deposits were found around the celiac axis and the SMA. Therefore, the staging (stage IVa) of these 5 patients was defined according to the operative findings and classified as R2 operations. Among the 9 patients with stage IVb, 3 had one resectable peritoneal nodule, 5 had one resectable liver metastase, and 1 had paraaortic positive lymph nodes that were not detected intraoperatively with intraoperative frozen examination. There were 3 patients with stage IVb in the SM-PVR group. Two of them presented with a liver metastasis and another had para-aortic lymph node metastases not detected with intraoperative frozen examination. Pathological lymph node metastasis was detected in 53 (60.9%) of the 87 patients. A margin-negative resection was performed in 61 patients (70.1%). Postoperative adjuvant chemoradiation therapy was administered to 52 of 87 patients, including 12 (57.1%) of 21 patients who underwent SM-PVR and 40 (60.6%) of 66 patients who did not. Criteria and methods for performing SM-PVR and lymphadenectomy Spiral computed tomography (Somatom Plus, Siemens, Erlanger, Germany) was used since 1991, and preoperative abdominal angiography was routinely performed. A standardized preoperative imaging including spiral computed tomography, ultrasonography, abdominal angiography, and magnetic resonance cholangiopancreatography was routinely applied to all patients since 1995. Endoscopic retrograde cholangiopancreatography and endoscopic ultrasonography were selectively performed. To be considered preoperatively, all patients were required to fulfill the following criteria for SM-PVR: (1) absence of extrapancreatic
122
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129
Fig. 1. Intraoperative views. A. Pancreatic head tumor and involvement of the portal vein. B. En-bloc pancreaticoduodenectomy with a segmental portal venous resection. C. Reconstruction of the portal vein. PV ⫽ portal vein; SpV ⫽ splenic vein; SMV ⫽ superior mesenteric vein; T ⫽ tumor; P ⫽ remnant pancreas; 3 ⫽ venous anastomosis.
disease, (2) no evidence of involvement of the superior mesenteric artery (SMA) and celiac axis, and (3) a patent SM-PV [15]. Total obstruction of the SM-PV with collateral veins shown in the preoperative imaging was considered as a contraindication for SM-PVR, as shown by Nakao et al [23], but unilateral and bilateral narrowings of the SM-PV were accepted as an indication for SM-PVR. PD was selected as the standard procedure. However, TP with splenectomy was performed when examination of frozen sections revealed that pancreatic resection margin was positive for tumor or when the pancreatic body or tail was found to contain multifocal tumor. Patients were principally required to fulfill the following intraoperative criteria for SM-PVR: (1) absence of liver metastasis, (2) absence of peritoneal dissemination, (3) absence of lymph node in-
volvement of the para-aortic area or more distant regions, (4) absence of involvement of the SMA, common hepatic artery, and celiac axis, and (5) evidence of tumor adherence to the SM-PV. The 3 patients with stage IVb of the SMPVR group were highly selected cases. They underwent PD or TP with SM-PVR associated with resection of one liver metastasis. This aggressive approach appeared justified only because of the young age of these 3 patients (27, 42, and 45 years old). An en-bloc resection of the SM-PV associated with PD was mainly performed. All mobilizations needed for PD or TP were performed before SM-PVR. During lymphadenectomy the portal vein and hepatic artery in the hepatoduodenal ligament were dissected until the porta hepatis. In the infrapancreatic area, the SMV was dissected. Progressively
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129 Table 2 Reconstruction methods after SM-PVR in 15 patients undergoing PD Types of SM-PV involvement
Types of reconstructions
Number of patients
Involvement of PV Involvement of SM-PV confluence (ⱕ6 cm)
End-to-end anastomosis End-to-end anastomosis with splenic vein reimplantation SMV-PV resectionreconstruction with saphenous vein graft and splenic vein reimplantation End-to-end anastomosis Lateral venous suture
4 5
Involvement of SM-PV confluence (⬎6 cm)
Involvement of SMV Lateral involvement of SM-PV
1
3 2
Patients undergoing total pancreatectomy were excluded. PD ⫽ pancreaticoduodenectomy; PV ⫽ portal vein; SMV ⫽ superior mesenteric vein; SM-PV ⫽ superior mesentericoportal vein; SM-PVR ⫽ superior mesentericoportal vein resection.
the gastrocolic trunk and the pancreaticoduodenal collaterals were ligated. The first jejunal vein insertion at the right posteroinferior side of the SMV was identified. Anterior and posterior lymph nodes dissection was continued until identifying SMA and celiac axis. The uncinate process was freed from the SMA as much as possible before SM-PVR. After establishing that a full mobilization of the SM-PV could not be performed (Fig. 1), vascular clamps were carefully placed at the upper and lower vascular ends without any routine clamping of the SMA. SMA clamping was performed when the portal vein reconstruction could not be achieved in less than 60 minutes. The upper and lower ends of the SM-PV were cut, and an en-bloc SM-PVR associated with PD was completed. Vascular reconstruction was then performed with a continuous suture using 5-0 Prolene (Ethicon, Brussels, Belgium). An en-bloc SM-PVR associated with PD was performed in 25 cases; a SM-PV resection was performed after PD in 6 cases. Among the 31 patients undergoing SM-PVR, segmental resection of the SM-PV was performed in 27 patients and Table 3 Profile of 31 patients with SM-PVR and 119 patients without SM-PVR
Male/female Age (years) Hospital stay (days) Duration of operation (minutes) Perioperative blood transfusion requirement (units)
SM-PVR (n ⫽ 31)
No SM-PVR (n ⫽ 119)
P value
15/16 62.2 ⫾ 2.0 22.2 ⫾ 1.4 501.2 ⫾ 17.1
71/48 61.2 ⫾ 1.2 25.9 ⫾ 1.1 412.7 ⫾ 6.5
0.354* 0.663† 0.111† ⬍0.0001†
2.0 ⫾ 0.2
⬍0.0001†
5.6 ⫾ 0.8
* Chi-square test. † Unpaired Student’s t test. SM-PVR ⫽ superior mesentericoportal vein resection.
123
lateral venous resection was achieved in 4 patients. Segmental resection of the PV was performed in 7 patients, of the confluence of the SM-PV in 13 patients, and of the SMV in 7 patients. The portal tract was reconstructed with a direct end-to-end anastomosis in 25 patients. Such venous reconstruction with a end-to-end anastomosis could be obtained, in our series, with as much as 6-cm venous segment resection. Larger venous resection needed a SMV-PV reconstructive bypass with a saphenous vein graft in 1 patient and a Gore-tex (Medical Products, Flagstaff, AZ) graft in another. The portal tract was reconstructed without use of a portosystemic shunt in all patients but 1, in whom a portosystemic shunt as described by Nakao et al [24] was used. This shunt was used because of occurrence of splanchnic venous congestion during clamping. A simple venous suture was performed in the 4 patients undergoing lateral venous resection. In patients undergoing PD with SM-PVR (n ⫽ 15), resection of the SM-PV confluence was performed in 6 patients. The splenic vein was reconstructed in all 6 patients. Methods of venous resection-reconstruction in patients undergoing PD with SM-PVR (n ⫽ 15) are shown in Table 2. Regional lymph nodes as defined by the Japanese classification [22], ie, nodes around the common hepatic artery, nodes around the celiac trunk, nodes in the hepatoduodenal ligament, and posterior and anterior pancreaticoduodenal nodes, were routinely dissected in patients with periampullary malignancies. When intraoperative pathological diagnosis showed involvement of these lymph nodes, a complete dissection of lymph nodes around the SMA and a lymph node sampling of the para-aortic area were performed. An extended lymph node dissection around the aorta as reported by Yeo et al [25] was not performed in the present study. Histopathological examination According to the Japanese histopathological classification of curative resection, the resection margin (pancreatic resection margin, resection margin of the bile duct, and peritumoral margin) should be 5 mm free from the cut surface (no exposure of the tumor at the cut surface). All pathologic specimens were re-reviewed by a pathologist to determine the presence or absence of tumor deposit on surgical margins (pancreatic transection line, bile duct transection line, duodenum, stomach, and peripancreatic soft tissue), as well as extent of local invasiveness (invasion into duodenum, bile duct, or peripancreatic soft tissue), histologic differentiation, lymph node metastases, and involvement of the SM-PVR. Statistical methods Data are expressed as mean ⫾ standard error of mean. All statistical analyses, including Student’s t test, the chi-
124
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129
Table 4 Postoperative complications in patients undergoing resection with or without SM-PVR Postoperative complications*
Sepsis Urinary infection Wound infection Pulmonary complications Pulmonary embolism Thrombophlebitis Pancreatic fistula Postoperative hemorrhage from operative field Gastric ulcer Liver ischemia Neurological complications Biliary stenosis Intraabdominal abscess Superior mesenteric venous infarction Intestinal occlusion Delayed gastric emptying Total number of patients with complications Number of patients undergoing reoperation
All patients
Patients with pancreatic ductal adenocarcinoma
SM-PVR (n ⫽ 31)
No SM-PVR (n ⫽ 119)
SM-PVR (n ⫽ 21)
No SM-PVR (n ⫽ 66)
5 0 1 5 1 3 1 (1/15 ⫽ 6.7%) 0 0 0 1 0 1 1 0 5 15 (48.4%) 3 (9.7%)
9 9 2 10 6 3 14 (14/101 ⫽ 13.9%)† 9 4 1 4 1 8 0 2 14 56 (47.1%)‡ 15 (12.6%)
3 0 0 3 1 1 0 0 0 0 0 0 0 1 0 2 8 (38.1%) 2 (9.5%)
1 6 0 5 0 1 4 5 3 0 3 1 2 0 1 7 30 (45.5%)†† 10 (15.2%)
* All listed postoperative complications are shown when a patient suffered multiple complications. † P ⫽ 0.44, chi-square test. ‡ P ⫽ 0.77, chi-square test. †† P ⫽ 0.55, chi-square test. SM-PVR ⫽ superior mesentericoportal vein resection.
square test, survival analysis, and stepwise multifactorial regression analysis, were performed with the computer software package Statview IV (Abacus Concepts, Inc., Berkeley, California). Rates of cumulative survival were calculated with the Kaplan-Meier method, and differences were analyzed with the log-rank test. A difference was considered significant when the P value was less than 0.05. In the stepwise multifactorial regression analysis, data were considered significant when the F test value was more than 4.0.
The length of hospitalization did not differ significantly between the two groups (Table 3). The rate of postoperative complications did not differ significantly (P ⫽ 0.77) between patients undergoing SMPVR (48.4%) and those (47.1%) not undergoing vascular resection (Table 4). However, the frequency of pancreatic fistula among patients undergoing PD was slightly lower, Table 5 Postoperative complications necessitating relaparotomy
Results
Indications for reoperation
SM-PVR (n ⫽ 31)
First analysis: early surgical outcomes in all patients undergoing PD
Intraabdominal abscess Superior mesenteric venous infarction Operative field hemorrhage due to pancreatic fistula Other operative field hemorrhage* Digestive hemorrhage Pancreatic fistula Stenosis or leakage of gastrojejunostomy Liver ischemia due to celiac axis compression Stenosis of hepaticojejunostomy Total
2 1
0 0
0
3
0 0 0 0
2 1 3 4
0
1
In the 150 patients, the mean operation time was 428.2 ⫾ 6.6 minutes, and the volume of red blood cells transfused perioperatively was 2.7 ⫾ 0.2 units. Following the criteria of Yeo et al [21], 71 patients (47.3%) were judged to have developed postoperative complications, and 4 patients (2.7%) died within 30 days after surgery. The mean hospital stay was 25.2 ⫾ 0.9 days. Among the 31 patients who underwent SM-PVR, PD was performed in 15 patients and TP in 16 patients. Operation time was significantly longer and perioperative blood transfusion was performed significantly more in patients undergoing SM-PVR (Table 3). Portal flow was clamped for 25.1 ⫾ 2.2 minutes in patients undergoing SM-PVR.
0 3 (9.7%)
No SM-PVR (n ⫽ 119)
1 15 (12.6%)†
* The other causes of the operative field hemorrhage include postoperative bleeding and a hemorrhage caused by biliary leakage. † P ⫽ 0.66, chi-square test. SM-PVR ⫽ superior mesentericoportal vein resection.
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129
125
Table 6 Profile of patients with ductal adenocarcinoma of the pancreatic head undergoing SM-PVR and those not undergoing SM-PVR SM-PVR (n ⫽ 21) Male/female Age (years) Operative method Total pancreataectomy Pancreaticoduodenectomy Method of pancreatic anastomosis Pancreaticogastrostomy Pancreaticojejunostomy Duration of operation (min) Perioperative red blood cell transfusion requirement (units) Tumor size (mm) Histopathological type Well-differentiated ductal adenocarcinoma Moderately differentiated ductal adenocarcinoma Poorly differentiated ductal adenocarcinoma Histopathological stage (22) 0 I II III IVa IVb Carcinoembryonic antigen (CEA) CA19-9 Clinicopathological curability Margin-negative resection Margin-positive resection Hospital stay after surgery (days) Postoperative complication Absent Present
No SM-PVR (n ⫽ 66)
P value
10/11 60.4 ⫾ 2.5
39/27 62.7 ⫾ 1.3
0.356* 0.405†
10 11
12 54
0.007*
5 6 503.1 ⫾ 18.4 5.5 ⫾ 1.0 47.9 ⫾ 3.8 5 15 1
27 27 410.4 ⫾ 8.0 1.8 ⫾ 0.2 41.0 ⫾ 2.5 22 36 8
0.741* ⬍0.0001† ⬍0.0001† 0.164† 0.141*
0 0 5 4 9 3 7.0 ⫾ 2.0 2231.9 ⫾ 807.1
1 7 12 30 10 6 10.8 ⫾ 1.8 1182.6 ⫾ 408.9
0.117*
13 8 20.3 ⫾ 1.4
48 18 23.9 ⫾ 1.2
0.345*
13 8
36 30
0.314† 0.314†
0.112† 0.554*
* Chi-square test. † Unpaired Student t test. SM-PVR ⫽ superior mesentericoportal vein resection.
but not significantly so, in patients undergoing SM-PVR (1 of 15 ⫽ 6.7%, rate of normal texture of pancreas ⫽ 48.4%) than in those not undergoing vascular resection (14 of 101 ⫽ 13.9%, P ⫽ 0.44, rate of normal texture of pancreas ⫽ 54.6%, Table 4). The frequency of a second laparotomy to treat postoperative complications did not differ between the two groups (9.7% of patients undergoing SMPVR and 12.6% of patients not undergoing vascular resection, P ⫽ 0.46). However, no patient required reoperation for pancreatic fistula or bleeding complications in patients undergoing SM-PVR, whereas 9 patients required reoperation owing to these complications among patients not undergoing SM-PVR (Table 5). Among patients undergoing SM-PVR, 1 died of mesenteric venous infarction 4 days after surgery. The venous reconstruction was performed without any difficulties or a kinking, and the cause of venous thrombosis was not clear. The mortality rate did not differ between patients undergoing SM-PVR (3.2%) and patients not undergoing vascular resection (2.5%). In addition, no postoperative death occurred in patients undergoing SM-PVR in the recent series since 1994 (n ⫽ 24).
Stepwise regression multivariate analysis showed that normal parenchyma of the remnant pancreas was the only significant risk factor for postoperative complications (F value ⫽ 5.4); SM-PVR itself was not a risk factor (F value ⫽ 0.1). Second analysis: survival and other outcomes of patients with ductal adenocarcinoma of the pancreatic head Among the 87 patients with ductal adenocarcinoma of the pancreatic head, duration of surgery was 431.2 ⫾ 8.7 minutes, and the volume of perioperative red blood cell transfusion was 2.6 ⫾ 0.3 units. Thirty-eight patients (43.6%) had postoperative complications, and 2 patients (2.3%) died within 30 days after surgery. Mean hospital stay was 23.1 ⫾ 1.0 days. The male-to-female ratio, age, type of surgery, clinical stage, and serum concentrations of preoperative carcinoembryonic antigen, and CA19-9 did not differ between patients undergoing SM-PVR and those not undergoing vascular resection (Table 6). The rate of the margin-negative resection was slightly but not significantly higher among patients
126
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129
Fig. 2. Cumulative survival curves of patients with ductal adenocarcinoma of the pancreatic head undergoing pancreaticoduodenectomy with or without superior mesentericoportal venous resection (SM-PVR). Survival of patients undergoing SM-PVR (n ⫽ 21, solid line) is similar to that of patients not undergoing SM-PVR (n ⫽ 66, dashed line, P ⫽ 0.48). Oneand two-year survival rates are 53.6% and 21.5% in patients undergoing SM-PVR and 51.1% and 24.3% in patients not undergoing SM-PVR.
not undergoing vascular resection (72.7%, 48 of 66 patients) than among those undergoing SM-PVR (61.9%, 13 of 21 patients, P ⫽ 0.35). Histopathologic examination of specimens from the 21 patients with suspected venous infiltration by cancer showed intimal infiltration in 7, medial infiltration in 4, and adventitial infiltration in 3. Therefore, one third of patients (n ⫽ 7) undergoing SM-PVR did not have histological involvement of the SM-PV. However, the involvement had
been grossly suspected because of close adhesion between tumor and venous wall. The rate of postoperative complications was slightly but not significantly lower in patients undergoing SM-PVR (38.1%) than in patients not undergoing vascular resection (45.5%, Table 4). Hospitalization after surgery was slightly but not significantly shorter in patients undergoing SM-PVR (20.3 ⫾ 1.4 days) than in those not undergoing vascular resection (23.9 ⫾ 1.2 days; P ⫽ 0.11). Among the 21 patients undergoing SM-PVR, 1 (4.7%) died of mesenteric venous infarction 4 days after surgery. One patient (1.5%) died of erosive hemorrhage caused by pancreatic fistula 7 days after surgery among the 66 patients not undergoing vascular resection. The mean follow-up period was 12.1 months (range 1 to 48). Ten of the 21 patients undergoing SM-PVR and 40 of the 66 patients not undergoing vascular resection died of cancer progression. Of the 2 patients undergoing SM-PVR and arterial resection, 1 is still alive 5 months after surgery and the other died of cancer recurrence 6 months after surgery. Median survival of patients (including those who died perioperatively) undergoing SM-PVR and of those not undergoing SM-PVR did not differ significantly (369 days versus 374 days, P ⫽ 0.48, Fig. 2). One-year and 2-year survival rates were 53.6% and 21.5% in patients undergoing SM-PVR and 51.1% and 24.3% in patients not undergoing vascular resection. Among patients undergoing SM-PVR, survival rate did not significantly differ between patients with histological venous infiltration (n ⫽ 14) and those without infiltration (n ⫽ 7; 2-year survival 34.9% versus 0%). Among patients who underwent a histologically marginnegative resection, median survival did not differ between patients undergoing SM-PVR (503 days) and those not undergoing resection (417 days; P ⫽ 0.141). Among the 13 patients undergoing SM-PVR and a margin-negative resection, 6 (46.2%) had recurrent disease, which was locoregional in each case. Survival rate for patients undergoing a margin-negative resection (n ⫽ 13) was significantly higher than that for patients undergoing macroscopic or microscopic margin positive resection (n ⫽ 8; 2-year survival, 57.1% versus 0%, P ⫽ 0.042, Fig. 3).
Comments
Fig. 3. Cumulative survival curves of patients undergoing a margin-negative resection (R0) and of those undergoing a macroscopically (R1) or microscopically (R2) margin-positive resection among patients undergoing superior mesentericoportal venous resection (SM-PVR). Survival rate is significantly higher for patients undergoing a margin-negative resection (n ⫽ 13, solid line) than for those undergoing a margin-positive resection (n ⫽ 8, dashed line, 2-year survival 57.1% versus 0%, P ⫽ 0.042).
Our results showed that PD with SM-PVR can be safely performed in selected cases. Duration of operation was longer and perioperative blood transfusion requirement was greater in patients undergoing SM-PVR than in patients not undergoing SM-PVR. However, postoperative morbidity, mortality, and hospitalization after SM-PVR did not differ between the two groups of patients. In addition, no postoperative death occurred in patients undergoing SM-PVR in the recent series, since 1994, in contrast to 1 death before 1994, suggesting a learning curve effect.
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129
127
Table 7 Reported rates of mortality, morbidity, survival, and margin-negative resection in pancreaticoduodenectomy with SM-PVR Authors (references)
Institutions and city names
Sindelar et al [9]
National Cancer Institute, Bethesda Mannheim University, Mannheim Academic Medical Centre, Amsterdam Memorial Sloan-Kettering Cancer Center, New York Munich University, Munich M.D. Anderson Cancer Center, Houston Pisa University, Pisa Tokyo Women’s Medical College, Tokyo Mie University, Tsu Kanazawa University, Kanazawa Nagoya University, Nagoya Keio University, Tokyo
Mortality (within 1 month) (%)
Morbidity (%)
Median survival (months)
Marginnegative resection (%)
20
20
55
12
—
60 (10.7%)* 20 (11.4%)
6.7 15
— 55
12 8
— 15
58 (17.5%)
5
12§
13
73
31 (10.4%)* 31 (41,3%)
0 0
41.9 30
8 22
32 87
27 (12.2)* 172 (69.0%)
7.2 5
42.1 23
— —
— —
30 (41.7%) 72 (50.0%) 104 (78%) 107 (55.7%)
1.0 6.9 8 3.8
15.7 37.2 — 23.1
26.5 — 49
37 (19.6%)
13.5
35.1§
Ishikawa et al [33,34,36]
Hannover Medical School, Hannover Osaka Medical Center, Osaka
19 — 5-year survival, 14%‡ 9.0
35 (70.0%)
5.7
—
Present study
Hautepierre Hospital, Strasbourg
31 (20.7%)*
3.2
3-year survival, 29% 13
Trede et al [4,32] Allema et al [11] Fortner et al [12] and Harrison et al [13] Roder et al [14] Leach et al [15] and Fuhrman et al [16] Mosca et al [37] Imaizumi et al [38] Naganuma et al [39] Nagakawa et al [31] Nakao et al [23,24]† Takahashi et al [17] and Ogata et al [26] Klempnauer et al [35]
Number of patients with SM-PVR and frequency
— 48.4
—
61.9
* Periampullary cancers other than pancreatic cancer were included. † Data of all patients were not documented. ‡ Data of patients with curative resection are shown. § Relapalotomy rate is demonstrated. SM-PVR ⫽ superior mesentericoportal vein resection.
In patients with ductal adenocarcinoma of the pancreatic head, PD or TP with SM-PVR seems justified in order to achieve a R0 resection when a close adhesion between the tumor and the venous wall is the only obstacle for resection. In fact, mortality, mobidity, and survival rates did not differ between patients with SM-PVR and those without SMPVR. In addition, as expected, the present study showed that survival of patients undergoing a margin-negative resection (R0) with a SM-PVR is significantly better than that of patients with microscopic (R1) or macroscopic (R2) margin-positive resection. However, this survival benefit should be carefully considered because the protocols of chemoradiation therapy were not uniformly performed in all patients with pancreatic ductal carcinoma. Currently, postoperative mortality after PD has decreased, ranging from 0% to 8% [2–5]. Surgeons seem to focus on pancreatectomy with vascular resection to obtain a margin-negative resection (R0) [4,8 –19,26 –29]. Surgical outcomes of pancreatectomy combined with vascular resection in patients with pancreatic cancer and periampullary diseases have recently been investigated. Yeo et al [30] reported that SM-PVR was performed in only 4% (26 of 650) of patients undergoing PD, a lower rate than the 20.7% in the present study. In addition, our review of the literature found only 20 studies of PD from 14 institutions that in-
cluded 20 or more patients who underwent SM-PVR (Table 7) [4,9,11–17,23,24,26,31–39]. Among these reports from the 14 institutions, only 6 reported mortality, morbidity, rate of margin-negative resection, and long-term survival. Therefore, a study examining these outcomes is still needed to determine the most appropriate management of tumors involving retropancreatic vessels. In these reports, mortality rate ranged from 0% to 20% and morbidity rate ranged from 12% to 55%. The present results therefore showed acceptable rates of mortality (3.2%) and morbidity (48.4%). Since 1994 we have principally performed PD with pancreaticogastrostomy. In the present study, the splenic vein was reimplanted to the SMV or SM-PV bypass in 6 patients. Ligation of the splenic vein is a simple therapeutic option. However, left-sided portal hypertension can occur with increased rate of gastrointestinal bleeding even from the pancreatic transection line or from the gastric mucosa (“hypertensive gastropathy”). In our practice we recommend, when it is possible, reimplantation of the splenic vein to decrease the risk of side effects of left-sided portal hypertension. Whether survival after SM-PVR is better than that of patients not undergoing SM-PVR should be discussed. Studies from the United States and Europe have reported median survival ranging from 8 to 22 months in patients with pancreatic cancer who had undergone pancreatic re-
128
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129
section with SM-PVR (Table 7) [4,9,11–16,32]. In all reports survival did not differ between patients who did and did not undergo SM-PVR. The present study also showed the same result. It is therefore concluded that SM-PVR itself does not improve survival in patients with pancreatic cancer. Whether survival of patients after SM-PVR is better than that of patients in whom margin-positive resection is performed because SM-PV involvement is not resected should be discussed. Allema et al [11] reported no difference in survival between patients undergoing SM-PVR and those in whom a positive margin at the SM-PV was not resected; they concluded that pancreatic resection with SM-PVR does not improve survival. However, their conclusion may be qualified because the rate of margin-negative resection was only 15% in patients undergoing SM-PVR. Several Japanese authors have reported higher survival rates after pancreatectomy with SM-PVR for patients with pancreatic cancer [17,26,27,31,33]. Takahashi et al [17] have reported a 5-year survival rate of 14% in patients undergoing marginnegative resection, with 6 patients surviving more than 5 years after SM-PVR. Ishikawa et al [33,36] have reported a 3-year survival rate of 29%, with 7 out of 35 patients surviving more than 3 years after SM-PVR. The present study confirmed that survival after SM-PVR in the patients undergoing a margin-negative resection is better than in those not undergoing a margin-negative resection (2-year survival, 57.1% versus 0%). In the present study one third of patients undergoing SM-PVR did not have histological involvement of SM-PV. However, Ishikawa et al [40] recently showed that intraoperative cytology showed positive SM-PV involvement in several patients who underwent PD without evidence of macroscopic involvement of the SM-PV, and that further SM-PVR after PD confirmed cancer cell infiltration of the SM-PV in most of them. These results suggest difficulty for intraoperatively confirming SM-PV involvement. In conclusion, the results of the present study show that combined PD and SM-PVR can be performed safely. However, SM-PVR should be considered for selected patients and performed only in selected institutions. Greater blood loss and longer operative time cannot be ignored, and the safety of this method has to be confirmed by a sufficient number of institutions. Moreover, there is probably a learning curve effect, as we experienced. On the basis of our findings with regard to postoperative morbidity, mortality, rate of margin-negative resection, and survival of patients with pancreatic adenocarcinoma, SM-PVR should be performed only when a margin-negative resection is expected to be achieved. References [1] Nitecki SS, Sarr MG, Colby TV, van Heerden JA. Long-term survival after resection for ductal adenocarcinoma of the pancreas: is it really improving? Ann Surg 1995;221:59 – 66.
[2] Wade TP, El-Ghazzawy AG, Virgo KS, Johnson FE. The Whipple resection for cancer in U.S. Department of Veterans Affairs Hospitals. Ann Surg 1995;221:241– 8. [3] Warshaw AL, Fernandez-del Castillo C. Pancreatic carcinoma. N Engl J Med 1992;326:455– 65. [4] Trede M, Schwall G, Saeger HD. Survival after pancreatoduodenectomy. Ann Surg 1990;211:447–58. [5] Cameron JL, Pitt HA, Yeo CJ, et al. One hundred and forty-five consecutive pancreaticoduodenectomies without mortality. Ann Surg 1993;217:430 – 8. [6] Klempnauer J, Ridder GJ, Bektas H, Pichlmayer R. Surgery for exocrine pancreatic cancer—who are the 5- and 10-year survivors? Oncology 1995;52:353–9. [7] Willet CG, Lewandrowski K, Warshaw AL, Efird J, Compton CC. Resection margins in carcinoma of the head of the pancreas. Implications for radiation therapy. Ann Surg 1993;217:144 – 8. [8] Yeo CJ, Cameron JL, Lillemoe KD, et al. Pancreaticoduodenectomy for cancer of the head of the pancreas: 201 patients. Ann Surg 1995;221:721–33. [9] Sindelar WF. Clinical experience with regional pancreatectomy for adenocarcinoma of the pancreas. Arch Surg 1989;124:127–32. [10] Launois B, Stasik C, Bardaxoglou E, et al. Who benefits from portal vein resection during pancreaticoduodenectomy for pancreatic cancer? World J Surg 1999;23:926 –9. [11] Allema JH, Reinders ME, van Guilik TM, et al. Portal vein resection in patients undergoing pancreatoduodenectomy for carcinoma of the pancreatic head. Br J Surg 1994;81:1642– 6. [12] Fortner JG, Klimstra DS, Scnie RT, Maclean BJ. Tumor size is the primary prognosticator for pancreatic cancer after regional pancreatectomy. Ann Surg 1996;223:147–53. [13] Harrison LE, Klimstra DS, Brennan MF. Isolated portal vein involvement in pancreatic adenocarcinoma. A contraindication for resection? Ann Surg 1996;224:342–9. [14] Roder JD, Stein HJ, Siewert JR. Carcinoma of the periampullary region: who benefits from portal vein resection? Am J Surg 1996; 171:170 –5. [15] Leach SD, Lee JE, Charnsangavej C, et al. Survival following pancreaticoduodenectomy with resection of the superior mesenteric-portal vein confluence for adenocarcinoma of the pancreatic head. Br J Surg 1998;85:611–17. [16] Fuhrman GM, Leach SD, Staley CA, et al. Rationale for en bloc vein resection in the treatment of pancreatic adenocarcinoma adherent to the superior mesenteric-portal vein confluence. Ann Surg 1996;223: 154 – 62. [17] Takahashi S, Ogata Y, Tsuzuki T. Combined resection of the pancreas and portal vein for pancreatic cancer. Br J Surg 1994;81: 1190 –3. [18] Ishikawa O, Ohigashi H, Imaoka S, et al. Preoperative indications for extended pancreatectomy for locally advanced pancreas cancer involving the portal vein. Ann Surg 1992;215:231– 6. [19] Tashiro S, Uchino R, Hiraoka T, et al. Surgical indication and significance of portal vein resection in biliary and pancreatic cancer. Surgery 1991;109:481–7. [20] Delcore R, Thomas JH, Pierce GE, Hermreck AS. Pancreatogastrostomy: a safe drainage procedure after pancreatoduodenectomy. Surgery 1990;108:641–7. [21] Yeo CJ, Cameron JL, Maher MM, et al. A prospective randomized trial of pancreaticogastrostomy versus pancreaticojejunostomy after pancreaticoduodenectomy. Ann Surg 1995;222:580 –92. [22] Kayahara M, Nagakawa T, Ueno K, et al. Surgical strategy for carcinoma of the pancreas head area based on clinicopathologic analysis of nodal involvement and plexus invasion. Surgery 1995; 117:616 –23. [23] Nakao A, Harada A, Nonami T, et al. Clinical significance of portal invasion by pancreatic head carcinoma. Surgery 1995;117:50 –5.
P. Bachellier et al. / The American Journal of Surgery 182 (2001) 120 –129 [24] Nakao A, Takagi H. Isolated pancreatectomy for pancreatic head carcinoma using catheter bypass of the portal vein. Hepatogastroenterology 1993;40:426 –9. [25] Yeo CJ, Cameron JL, Sohn TA, et al. Pancreaticoduodenectomy with or without extended retroperitoneal lymphadenectomy for periampullary adenocarcinoma. Comparison of morbidity and mortality and short-term outcome. Ann Surg 1999;229:613–24. [26] Ogata Y, Hishinuma S, Takahashi S, et al. Indication and results of pancreatectomy with combined resection of vessels for adenocarcinoma of the pancreas. Nippon Geka Gakkai Zasshi 1997;98:615– 21. [27] Nakao A, Harada A, Nonami T, et al. Regional vascular resection using catheter bypass procedure for pancreatic cancer. Hepatogastroenterology 1995;42:734 –9. [28] Luttges J, Vogel I, Menke M, et al. The retroperitoneal resection margin and vessel involvement are important factors determining survival after pancreaticoduodenectomy for ductal adenocarcinoma of the head of the pancreas. Virchows Arch 1998;433:237– 42. [29] Civello IM, Frontera D, Viola G, et al. Extensive resection in pancreatic cancer: review of the literature and personal experience. Hepatogastroenterology 1998;45:1877– 83. [30] Yeo CJ, Cameron JL, Sohn TA, et al. Six hundred fifty consecutive pancreaticoduodenectomies in the 1990s: pathology, complications, and outcomes. Ann Surg 1997;226:248 – 60. [31] Nagakawa T, Konishi I, Ueno K, et al. Extended radical pancreatectomy for carcinoma of the head of the pancreas. Hepatogastroenterology 1998;45:849 –54.
129
[32] Trede M. Vascular problems and techniques associated with pancreatectomy and regional pancreatectomy. In: Trede M, Carter DC, editors. Surgery of the pancreas. 2nd ed. New York: Churchill Livingstone, 1997, p 505–16. [33] Ishikawa O, Ohhigashi H, Sasaki Y, et al. Practical usefulness of lymphatic and connective tissue clearance for the carcinoma of the pancreas head. Ann Surg 1988;208:215–20. [34] Ishikawa O. Surgical technique, curability and postoperative quality of life in an extended pancreatectomy for adenocarcinoma of the pancreas. Hepatogastroenterology 1996;43:320 –5. [35] Klempnauer J, Ridder GJ, Bektas H, Pichlmayr R. Extended resections of ductal pancreatic cancer—impact on operative risk and prognosis. Oncology 1996;53:47–53. [36] Ishikawa O, Ohhigashi H, Sasaki Y, et al. Extended pancreatectomy and liver perfusion chemotherapy for resectable adenocarcinoma of the pancreas. Digestion 1999;60(suppl 1):135– 8. [37] Mosca F, Giulianotti PC, Balestracci T, et al. Long-term survival in pancreatic cancer: pylorus-preserving versus Whipple pancreatoduodenectomy. Surgery 1997;122:553– 66. [38] Imaizumi T, Hanyu F, Harada N, et al. Extended radical Whipple resection for cancer of the pancreatic head: operative procedure and results. Dig Surg 1998;15:299 –307. [39] Naganuma T, Isaji S, Kawarada Y. Staging and extended resection for pancreatic cancer. Pancreas 1998;16:355– 62. [40] Ishikawa O, Ohhigashi H, Sasaki Y, et al. Intraoperative cytodiagnosis for detecting a minute invasion of the portal vein during pancreaticoduodenectomy for adenocarcinoma of the pancreatic head. Am J Surg 1998;175:477– 81.