Preoperative and postoperative chemoradiation strategies in patients treated with pancreaticoduodenectomy for adenocarcinoma of the pancreas.

PURPOSE: The effects of preoperative versus postoperative fluorouracil (5-FU)-based chemotherapy and irradiation on treatment toxicity, duration of treatment, tumor recurrence, and survival were compared in patients who underwent potentially curative therapy for adenocarcinoma of the pancreatic head during a 5-year period. METHODS: From July 1990 to July 1995, 142 patients with localized adenocarcinoma of the pancreatic head deemed resectable on the basis of radiographic images were treated with curative intent using a multimodality approach involving either preoperative or postoperative chemoradiation. Patients with biopsy confirmation of adenocarcinoma and a low-density mass in the pancreatic head identified by computed tomography (CT) received preoperative chemoradiation. Patients without a mass on CT or in whom the preoperative biopsy was negative underwent pancreaticoduodenectomy with planned postoperative chemoradiation. Protocol-based preoperative chemoradiation consisted of external-beam irradiation at a dose of 50.4 Gy (standard fractionation; 1.8 Gy/d, 5 d/wk) or 30 Gy (rapid fractionation; 3 Gy/d, 5 d/wk) combined with continuous infusion 5-FU (300 mg/m2/d, 5 d/wk). Postoperative chemoradiation combined 50.4 Gy of external-beam irradiation (standard fractionation) with continuous-infusion 5-FU. RESULTS: No patient who received preoperative chemoradiation experienced a delay in surgery because of chemoradiation toxicity, but six of 25 eligible patients (24%) did not receive postoperative chemoradiation because of delayed recovery after pancreaticoduodenectomy. No significant differences in toxicities from chemoradiation were observed between groups. Patients treated with rapid-fractionation preoperative chemoradiation had a significantly (P < .01) shorter duration of treatment (median, 62.5 days) compared with patients who received postoperative chemoradiation (median, 98.5 days) or standard-fractionation preoperative chemoradiation (median, 91.0 days). At a median followup of 19 months, no significant differences in survival were observed between treatment groups. No patient who received preoperative chemoradiation and pancreaticoduodenectomy experienced a local recurrence; peritoneal (regional) recurrence occurred in 10% of these patients. Local or regional recurrence occurred in 21% of patients who received pancreaticoduodenectomy and postoperative chemoradiation. CONCLUSION: Delivery of preoperative and postoperative chemoradiation in patients who underwent potentially curative pancreaticoduodenectomy for adenocarcinoma of the pancreatic head resulted in similar treatment toxicity, patterns of tumor recurrence, and survival. Rapid-fractionation preoperative chemoradiation ensured the delivery of all components of therapy to all eligible patients with a significantly shorter duration of treatment than with standard-fractionation chemoradiation given either before or after pancreaticoduodenectomy. Prolonged recovery after pancreaticoduodenectomy prevents the delivery of postoperative adjuvant chemoradiation in up to one fourth of eligible patients.

Relative expression of E-cadherin and type IV collagenase genes predicts disease outcome in patients with resectable pancreatic carcinoma.

We examined the expression level of several genes that regulate distinct steps of metastasis in formalin-fixed, paraffin-embedded, archival specimens of primary human pancreatic carcinomas from patients undergoing curative surgery. The expression of epidermal growth factor receptor, E-cadherin, type IV collagenase [matrix metalloproteinase (MMP) 2 and MMP-9), basic fibroblast growth factor, vascular endothelial growth factor/vascular permeability factor, and interleukin 8 was examined by a colorimetric in situ mRNA hybridization technique. Down-regulation of E-cadherin and up-regulation of type IV collagenase (MMP-9 and MMP-2) at the periphery of the neoplasms (P = 0.0167, 0.0102, and 0.0349, respectively) had significant prognostic value. The ratio of type IV collagenase expression (mean of the expression of MMP-2 and MMP-9) to E-cadherin expression (MMP:E-cadherin ratio) at the periphery of the tumors was significantly higher in patients with recurrent disease (4.7 +/- 2.1) than in patients who were disease free (2.3 +/- 1.7; P = 0.0008). Death from pancreatic cancer was significantly associated with a high MMP:E-cadherin ratio (>3.0) by overall survival analysis (P < 0.0002), whereas a low MMP:E-cadherin ratio (<3.0) was found in seven of eight patients alive 28-64 months after surgery. Multivariate analysis of overall survival showed that the MMP:E-cadherin ratio was a significant independent prognostic factor, whereas stage, nodal metastasis, and histological type were not. These data show that multiparametric analysis for several metastasis-related genes may allow physicians to assess the metastatic potential and hence predict the clinical outcome of individual patients with resectable pancreatic carcinoma.

bcl-2 and p53 expression in resectable pancreatic adenocarcinomas: association with clinical outcome.

The bcl-2 proto-oncogene and the p53 tumor suppressor gene are important determinants of tumor cell susceptibility to apoptosis. bcl-2 and mutant p53 proteins inhibit apoptosis in vitro and can provide prognostic information in certain tumor types. We analyzed bcl-2 and p53 expression in archival pancreatic (n = 35) and ampullary (n = 6) adenocarcinomas, resected for cure, and their relationship to overall survival. Patients were treated with 5-fluorouracil and irradiation either pre- (n = 21) or postoperatively (n = 15); 5 patients received surgery alone. Using specific monoclonal antibodies, cytoplasmic bcl-2 and nuclear p53 proteins were detected in 22 of 40 (55%) and 20 of 37 (54%) tumors, respectively. No relationship was found between bcl-2 and p53 expression. Neither bcl-2 nor p53 correlated with histological response to preoperative chemoradiation. Lymph node involvement predicted poor overall survival (P = 0.02). A trend toward improved survival was seen in well-differentiated (P = 0.08) tumors and in those with increased bcl-2 expression (P = 0.06). p53 expression was not related to clinical outcome. In a multivariate analysis, nodal status was the single most important predictor of overall survival. Of note, the combined variable of bcl-2 expression and histological grade was a stronger prognostic variable than nodal status alone. Unlike nodal status, these features can potentially be evaluated in preoperative biopsy specimens.

Vessel counts and vascular endothelial growth factor expression in pancreatic adenocarcinoma.

Angiogenesis is essential for growth and metastasis of solid malignancies. In several tumours, tumour vessel count and expression of vascular endothelial growth factor (VEGF), a potent angiogenic factor, have been associated with prognosis. To determine if vessel count and VEGF expression are prognostic factors in pancreatic cancer, we examined these parameters in resected tumour specimens from 22 patients who did not receive pre-operative therapy. Paraffin-embedded tumour specimens were immunohistochemically stained for factor VIII (surrogate for vessels) and VEGF. Vessel counts and VEGF expression were evaluated without knowledge of patient outcome. The median follow-up for the entire group had not been reached as of 23.1 months (range 10-69 months). The mean vessel count and VEGF expression were no different between those patients who had recurrences and those who did not. By linear regression analysis, the correlation of VEGF expression with vessel count did not reach statistical significance (P = 0.0685). Survival and time to recurrence were similar in patients with high and low vessel counts and VEGF expression of 1, 2 or 3. Tumour differentiation or lymph node positivity had no effect on either VEGF expression or vessel count. Our data suggest that, in contrast to findings in other solid malignancies, vessel count and VEGF expression are not predictors of survival or recurrence in patients with resectable adenocarcinoma of the pancreas.

Preoperative chemoradiation for adenocarcinoma of the pancreas: excessive toxicity of prophylactic hepatic irradiation.

PURPOSE: In an effort to reduce relapse in the liver and improve survival in patients with potentially resectable adenocarcinoma of the pancreatic head, we combined whole-liver irradiation with our standard preoperative chemoradiation regimen. METHODS AND MATERIALS: Eleven patients with biopsy-proven, potentially resectable adenocarcinoma of the pancreatic head were treated with 50.4 Gy of external beam irradiation to the pancreas (1.8 Gy/day, 5 days/week) and concurrent continuous infusion 5-fluorouracil (300 mg/m2 per day). The liver was treated with 23.4 Gy on Days 8 through 21 (13 fractions; 1.8 Gy/fraction). Patients, who upon restaging with radiography and computed tomography were considered to have resectable tumors, were subsequently taken to surgery. If, at surgery, tumors were resectable, pancreaticoduodenectomy was performed, and 10 Gy of intraoperative electron-beam radiation therapy was delivered to the bed of the resected pancreas. RESULTS: All 11 patients completed chemoradiation. Two treatment-related deaths occurred following chemoradiation, prompting premature termination of the study. Of seven patients taken to surgery, four underwent resection. Seven patients have died of disease, five with liver metastases. CONCLUSIONS: Prophylactic hepatic chemoradiation, as given in this study, was associated with two treatment-related deaths and a higher than expected incidence of subsequent liver metastases. Our data do not support the use of this treatment program in patients with adenocarcinoma of the pancreas.

Nonfunctioning islet cell carcinoma of the pancreas.

BACKGROUND: Nonfunctioning islet cell carcinoma of the pancreas has a variable and often indolent natural history, which has resulted in a wide range of treatment recommendations. To more clearly define the natural history and appropriate treatment of this disease, we reviewed our institutional experience over the last 39 years. METHODS: The records of all patients confirmed to have a nonfunctioning islet cell carcinoma of the pancreas were retrospectively reviewed. Kaplan-Meier life tables were constructed and log-rank comparisons were performed. RESULTS: The 73 patients studied had an overall 5-year actuarial survival rate of 50%. Patients with localized disease at presentation (n = 39) had a significantly higher survival rate (p = 0.03) compared with patients with metastatic disease (n = 34). The 19 patients who underwent a potentially curative resection of the primary tumor had a significantly higher survival rate (p = 0.03) compared with the 20 patients with locally advanced, unresectable, nonmetastatic disease. Nine of these 20 patients died of complications of the primary tumor. In contrast, only 2 of 22 cancer-related deaths in the 34 patients with metastatic disease at diagnosis were due to the primary tumor. CONCLUSIONS: (1) Surgical resection should be performed in patients with resectable nonmetastatic disease. (2) Resection of the primary tumor in the presence of metastatic disease is rarely indicated. (3) Innovative treatment strategies are needed for patients with locally advanced, unresectable, nonmetastatic tumors of the pancreatic head.

Preoperative chemoradiation and pancreaticoduodenectomy for adenocarcinoma of the pancreas.

Chemoradiation prior to pancreaticoduodenectomy ensures that all patients who undergo resection complete multimodality therapy, avoids resection in patients with rapidly progressive disease, and allows radiation therapy to be delivered to well-oxygenated cells before surgical devascularization. Twenty-eight patients with cytologic or histologic proof of localized adenocarcinoma of the pancreatic head received preoperative chemoradiation (fluorouracil, 300 mg/m2 per day, and 50.4 Gy) with the intent of proceeding to resection; all 28 completed this preoperative therapy. Hospital admission because of gastrointestinal toxic effects was required in nine patients, yet no patient experienced a delay in operation. Restaging was performed 4 to 5 weeks after completion of chemoradiation, and five patients were found to have metastatic disease; the 23 patients without evidence of progressive disease underwent laparotomy. At laparotomy, three patients were found to have unsuspected metastatic disease, three patients had unresectable locally advanced disease, and 17 patients were able to undergo pancreaticoduodenectomy. One perioperative death resulted from myocardial infarction, and perioperative complications occurred in three patients. Histologic evidence of tumor cell injury was present in all resected specimens. Our results suggest that pancreaticoduodenectomy can be performed with a low incidence of complications after chemoradiation for localized adenocarcinoma of the pancreas.

The need for standardized pathologic staging of pancreaticoduodenectomy specimens.

A standardized method for pathologic evaluation and staging of pancreaticoduodenectomy (PD) specimens is critical for accurate reporting of the number and location of lymph nodes and margins of resection. We examined the impact of standardized pathologic evaluation (SPE) of PD specimens on the identification of regional lymph nodes and describe our detailed system for the pathologic analysis of the PD specimen. Forty consecutive patients underwent PD for histologically confirmed adenocarcinoma of the pancreatic head between April 1990 and August 1993. Fifteen consecutive specimens were examined before the introduction of the SPE, and 25 consecutive specimens underwent SPE. Resection margins were evaluated by frozen-section analysis, and then the specimen was divided into six regions on an anatomic dissection board for lymph node identification. The 25 specimens examined according to the SPE had a significantly increased number of lymph nodes identified (P = 0.0001) compared with the 15 specimens examined without the SPE. Twelve of the 25 specimens contained positive lymph nodes, 6 of which were confined to the pancreaticoduodenal region. No positive nodes were found in the periaortic region. There were no differences in pathologic variables between patients found to have negative and those with positive regional lymph nodes. SPE of PD specimens provides a method for improved lymph node identification, ensures accurate prospective evaluation of margins of resection, and provides a complete analysis of potentially important pathologic variables. We offer this system as a standardized model for groups engaged in protocol-based clinical research examining innovative multimodality treatment strategies for patients with resectable pancreatic cancer.

Preoperative chemoradiation, pancreaticoduodenectomy, and intraoperative radiation therapy for adenocarcinoma of the pancreatic head.

BACKGROUND: Local recurrence in the bed of the resected pancreas is the most common site of tumor recurrence following a standard pancreaticoduodenectomy (PD) for adenocarcinoma of the pancreatic head. In an attempt to improve local and regional disease control and thereby enhance the quality and length of survival in patients undergoing potentially curative PD, we have used a protocol of preoperative multimodality therapy. PATIENTS AND METHODS: All patients were treated with external-beam radiation (30.0 or 50.4 Gy) and concomitant 5-fluorouracil (300 mg/m2 per day) prior to PD. Electron-beam intraoperative radiation therapy was given to the bed of the resected pancreas before reconstruction. Patients were assessed for recurrence by physical examination, chest roentgenography, and computed tomography scan performed at 3-month intervals following treatment. RESULTS: Thirty-nine patients completed all therapy; 1 perioperative death occurred. Thirty-eight tumor recurrences have been documented in 29 patients at a median of 11 months from the date of diagnosis; 23 patients died of disease. The liver was the most frequent site of recurrence, and liver metastases were a component of treatment failure in 53% of patients. Isolated local or peritoneal recurrences were documented in only 4 patients (11%). The only significant clinical or pathologic variable predictive of local-regional recurrence was a previous laparotomy and intraoperative biopsy. The median survival of all 39 patients was 19 months, and the 4-year actuarial survival rate was 19%. CONCLUSIONS: Preoperative chemoradiation, PD, and electron-beam intraoperative radiation therapy for adenocarcinoma of the pancreatic head have resulted in improved local-regional tumor control, with distant metastatic disease becoming the predominant site of tumor recurrence. Future treatment strategies should incorporate effective multimodality therapy for local-regional disease as demonstrated in this study. Major improvements in overall survival will likely await the development of systemic or regional therapy for liver metastases.

Thin-section contrast-enhanced computed tomography accurately predicts the resectability of malignant pancreatic neoplasms.

A prospective diagnostic study was designed to determine the ability of thin-section contrast-enhanced computed tomography (CT) to predict the resectability of malignant neoplasms of the pancreatic head. Patients with a presumed resectable pancreatic neoplasm referred during a 21-month period were studied with abdominal CT performed at 1.5-mm section thickness and 5-mm slice interval during the bolus phase of intravenous contrast enhancement. CT criteria for resectability included the absence of extrapancreatic disease, no evidence of arterial encasement, and a patent superior mesenteric-portal venous confluence. Of 145 patients evaluated, 42 were considered to have resectable tumors by CT criteria, and 37 (88%) underwent potentially curative pancreaticoduodenectomy. Six patients were found to have a microscopically positive retroperitoneal resection margin; no patient had a grossly positive resection margin. Five (12%) of 42 patients were found at laparotomy to have unresectable, locally advanced or metastatic tumors. Thin-section contrast-enhanced CT is an essential component of the preoperative evaluation for pancreaticoduodenectomy and can prevent needles laparotomy in most patients with locally advanced or metastatic disease.

Reoperative pancreaticoduodenectomy for periampullary carcinoma.

BACKGROUND: We have noted a continued increase in the number of patients referred to our institution for presumed or biopsy-proven periampullary carcinoma following an "exploratory" laparotomy during which tumor resection was not performed. Although previous work has demonstrated the safety of reoperative pancreaticoduodenectomy (PD), the need to avoid nontherapeutic laparotomy in these patients is obvious. In the current study, we sought to determine why PD was not performed at the initial operation. METHODS: Using the prospective pancreatic cancer database, we identified all patients who underwent reoperative PD at our institution between June 1990 and October 1995. Radiologic imaging prior to reoperation was standardized and based on thin-section, contrast-enhanced computed tomography (CT); helical CT was used in more recent cases. Pathologic data were obtained, and initial outside operative reports were reviewed to determine why a PD was not performed at the initial procedure. RESULTS: Twenty-nine patients underwent reoperative PD. Resection was not performed at the initial laparotomy because of the surgeon's assessment of local unresectability (17 patients), lack of a tissue diagnosis of malignancy (9), misdiagnoses (2), and error in intraoperative management (1). In the 17 patients deemed to have unresectable disease, successful reoperative PD required vascular resection in 10. All 10 of these patients had resection with negative microscopic margins of excision. Of the 9 patients who did not have resection owing to diagnostic uncertainty, all 9 had undergone multiple intraoperative biopsies interpreted as negative for malignancy; 6 of 9 had carcinoma confirmed on permanent-section analysis of the biopsy specimens. Four patients suffered major complications from intraoperative large-needle biopsy. CONCLUSIONS: Detailed preoperative imaging and a clearly defined operative plan would have allowed successful resection at the initial operation in 27 of 29 patients who underwent reoperative PD. Avoidable patient morbidity and the cost of unnecessary surgery argue strongly against "exploratory" surgery in patients with presumed periampullary neoplasms.

Survival following pancreaticoduodenectomy with resection of the superior mesenteric-portal vein confluence for adenocarcinoma of the pancreatic head.

BACKGROUND: The survival of patients who underwent pancreaticoduodenectomy with or without en bloc resection of the superior mesenteric-portal vein (SMPV) confluence for adenocarcinoma of the pancreatic head was compared. METHODS: To be considered for surgery, patients were required to fulfil the following computed tomography criteria for resectability: (1) absence of extrapancreatic disease, (2) no evidence of tumour extension to the superior mesenteric artery (SMA) or coeliac axis, and (3) a patent SMPV confluence. Tumour adherence to the superior mesenteric vein (SMV) or SMPV confluence was assessed at operation and en bloc venous resection was performed when necessary to achieve complete tumour extirpation. RESULTS: Seventy-five consecutive patients underwent pancreaticoduodenectomy, 44 without venous resection and 31 with en bloc resection of the SMPV confluence. There were no perioperative deaths in either group; late (more than 6 months) occlusion of the reconstructed SMPV confluence contributed to the death of two patients. Median survival in the 31 patients who required venous resection at the time of pancreaticoduodenectomy was 22 months, and that for the 44 control patients was 20 months (P = 0.25). CONCLUSION: Patients with adenocarcinoma of the pancreatic head who require venous resection during pancreaticoduodenectomy for isolated tumour extension to the SMV or SMPV confluence (in the absence of tumour extension to the SMA or coeliac axis) have a duration of survival no different from that of patients who undergo standard pancreaticoduodenectomy. These data suggest that venous involvement is a function of tumour location rather than an indicator of aggressive tumour biology.

Prospective, randomized trial of octreotide to prevent pancreatic fistula after pancreaticoduodenectomy for malignant disease.

OBJECTIVE: This study was conducted to determine whether the perioperative administration of octreotide decreases the incidence of pancreatic anastomotic leak after pancreaticoduodenectomy for malignancy. SUMMARY BACKGROUND DATA: Three multicenter, prospective, randomized trials concluded that patients who receive octreotide during and after pancreatic resection have a reduction in the total number of complications or a decreased incidence of pancreatic fistula. However, in the subset of patients who underwent pancreaticoduodenectomy for malignancy, either no analysis was performed or no benefit from octreotide could be demonstrated. METHODS: A single-institution, prospective, randomized trial was conducted between June 1991 and December 1995 involving 120 patients who were randomized to receive octreotide (150 microg subcutaneously every 8 hours through postoperative day 5) or no further treatment after pancreaticoduodenectomy for malignancy. The surgical technique was standardized, and the pancreaticojejunal anastomosis was created using the duct-to-mucosa or invagination technique. RESULTS: The two patient groups were similar with respect to patient demographics, treatment variables, and histologic diagnoses. The rate of clinically significant pancreatic leak was 12% in the octreotide group and 6% in the control group (p = 0.23). Perioperative morbidity was 30% and 25%, respectively. Patients who underwent reoperative pancreaticoduodenectomy had an increased incidence of pancreatic anastomotic leak, whereas those who received preoperative chemoradiation had a decreased incidence of pancreatic anastomotic leak. CONCLUSIONS: The routine use of octreotide after pancreaticoduodenectomy for malignancy cannot be recommended.

Rationale for en bloc vein resection in the treatment of pancreatic adenocarcinoma adherent to the superior mesenteric-portal vein confluence. Pancreatic Tumor Study Group.

OBJECTIVE: Tumor invasion of the superior mesenteric-portal vein (SMPV) confluence is often considered a contraindication to pancreaticoduodenectomy for patients with malignant tumors of the pancreas or periampullary region. The authors sought to determine whether pancreaticoduodenectomy with en bloc resection of the SMPV confluence could be safely performed and whether tumors involving the SMPV confluence were associated with pathologic parameters suggesting poor prognosis. SUMMARY BACKGROUND DATA: Several centers have reported high rates of retroperitoneal margin positivity after pancreaticoduodenectomy for tumors of the pancreatic head and periampullary region. Positive-margin or incomplete resection is associated with early tumor recurrence and no survival benefit compared with palliative therapy. Tumor adherence to the lateral of posterior wall of the SMPV confluence often represents the only barrier to complete tumor resection at the time of pancreaticoduodenectomy. METHODS: Data on all patients undergoing pancreaticoduodenectomy for adenocarcinoma of the pancreas or periampullary region over a 3.5-year period were entered prospectively in a pancreatic tumor database. To be considered for surgery, patients were required to fulfill the following computed tomography criteria for resectability: 1) the absence of extrapancreatic disease, 2) no tumor encasement of the superior mesenteric artery or celiac axis, and 3) a patent SMPV confluence. Tumor adherence to the superior mesenteric vein or SMPV confluence was assessed intraoperatively, and en bloc venous resection was performed when necessary to achieve complete tumor extirpation. Data on operative characteristics, morbidity, mortality, tumor size, nodal metastases, margin positivity, perineural invasion, and tumor DNA content were compared for patients who did and did not receive venous resection. RESULTS: Fifty-nine patients underwent pancreaticoduodenectomy, 36 without venous resection and 23 with en bloc resection of the SMPV confluence. No differences in median hospital stay, morbidity, mortality, tumor size, margin positivity, nodal positivity, or tumor DNA content were observed between groups. CONCLUSIONS: When necessary, segmental resection of the SMPV confluence may be performed safely during pancreaticoduodenectomy for periampullary malignant tumors. Tumors invading the SMPV confluence are not associated with histologic parameters suggesting a poor prognosis. Our data suggest that venous involvement is a function of tumor location rather than an indicator of aggressive tumor biology.