Fluorouracil vs Gemcitabine Chemotherapy Before and After Fluorouracil-Based Chemoradiation Following Resection of Pancreatic Adenocarcinoma: A Randomized Controlled Trial

Context Among patients with locally advanced metastatic pancreatic adenocarcinoma, gemcitabine has been shown to improve outcomes compared with fluorouracil.

Objective To determine if the addition of gemcitabine to adjuvant fluorouracil chemoradiation (chemotherapy plus radiation) improves survival for patients with resected pancreatic adenocarcinoma.

Design, Setting, and Participants Randomized controlled phase 3 trial of patients with complete gross total resection of pancreatic adenocarcinoma and no prior radiation or chemotherapy enrolled between July 1998 and July 2002 with follow-up through August 18, 2006, at 164 US and Canadian institutions.

Intervention Chemotherapy with either fluorouracil (continuous infusion of 250 mg/m² per day; n = 230) or gemcitabine (30-minute infusion of 1000 mg/m² once per week; n = 221) for 3 weeks prior to chemoradiation therapy and for 12 weeks after chemoradiation therapy. Chemoradiation with a continuous infusion of fluorouracil (250 mg/m² per day) was the same for all patients (50.4 Gy).

Main Outcome Measures Survival for all patients and survival for patients with pancreatic head tumors were the primary end points. Secondary end points included toxicity.

Results A total of 451 patients were randomized, eligible, and analyzable. Patients with pancreatic head tumors (n = 388) had a median survival of 20.5 months and a 3-year survival of 31% in the gemcitabine group vs a median survival of 16.9 months and a 3-year survival of 22% in the fluorouracil group (hazard ratio, 0.82 [95% confidence interval, 0.65-1.03]; P = .09). The treatment effect was strengthened on multivariate analysis (hazard ratio, 0.80 [95% confidence interval, 0.63-1.00]; P = .05). Grade 4 hematologic toxicity was 1% in the fluorouracil group and 14% in the gemcitabine group (P < .001) without a difference in febrile neutropenia or infection. There were no differences in the ability to complete chemotherapy or radiation therapy (>85%).

Conclusions The addition of gemcitabine to adjuvant fluorouracil-based chemoradiation was associated with a survival benefit for patients with resected pancreatic cancer, although this improvement was not statistically significant.

Trial Registration clinicaltrials.gov Identifier: NCT00003216

Despite potentially curative resection for pancreatic adenocarcinoma, patterns of failure analyses demonstrate a 50% to 85% component of local relapse associated with liver and intra-abdominal failure and a 5-year survival of less than 20%.^1-5 The frequency and pattern of failure make the combination of adjuvant postoperative chemotherapy and radiation an important consideration. Phase 3 trials have demonstrated long-term survival of up to 20% of patients with resected adenocarcinoma of the pancreas treated with adjuvant fluorouracil-based chemoradiation (chemotherapy plus radiation).^6-8 Although these trials have been small, their findings, combined with large institutional results,^9,10 have supported the use of postoperative adjuvant chemoradiation with fluorouracil as a standard of care in the United States for more than 20 years.

The US Food and Drug Administration approved gemcitabine in 1996 as the first chemotherapeutic agent for patients with pancreatic cancer since its approval of fluorouracil nearly 35 years ago. Approval was based on the results of studies on advanced disease.^11,12 In a randomized trial of gemcitabine vs fluorouracil as a first-line therapy in 126 patients with advanced or metastatic adenocarcinoma of the pancreas, the median survival for patients treated with gemcitabine was 5.7 months and 1-year survival was 18% compared with a median survival of 4.4 months and 1-year survival of 2% for patients treated with fluorouracil (P = .003).¹¹ The activity of gemcitabine in advanced pancreatic cancer led to the evaluation of gemcitabine in combination with chemoradiation with fluorouracil in an attempt to improve survival for patients with resected pancreatic adenocarcinoma. This study represents the first US cooperative group adjuvant pancreatic phase 3 trial in 3 decades.

An intergroup trial was conducted by the following US National Cancer Institute–sponsored cooperative groups: the Radiation Therapy Oncology Group (RTOG), the Eastern Cooperative Oncology Group, and the Southwest Oncology Group, inclusive of Canadian affiliates. The RTOG served as the lead group with the trial designation RTOG 97-04.

The eligibility criteria included histologically confirmed adenocarcinoma of the pancreas and gross total tumor resection, confirmed by central review of operative and pathology reports. In addition, postoperative computed tomographic (CT) imaging was required within 3 weeks of randomization to exclude patients who had evidence of persistent or recurrent local disease or developed metastatic disease prior to therapy. Surgical resection margins were categorized as negative, microscopically positive, or unknown (defined as those having no comment regarding margins in the pathology report).

Eligibility requirements also included stages T1 to T4, N0 to N1, and M0 according to the 1997 staging criteria of the American Joint Commission on Cancer (Box).¹³ If there were no identifiable lymph nodes within the resection specimen, the patient was ineligible. Patients were required to have a Karnofsky performance status of 60 or higher; adequate hematologic, renal, and hepatic function as defined by the following: a white blood cell count of 3 × 10³/μL or higher, a platelet count of 100 × 10³/μL or higher, serum bilirubin and creatinine less than 1.5 × the upper limit of institutional normal, a serum aspartate aminotransferase concentration 5 × the upper limit of institutional normal, and a documented caloric intake of more than 1500 kcal/d. Patients with any prior radiotherapy to any site or chemotherapy were ineligible for this study, as were patients with any prior malignancy other than nonmelanoma of the skin or in situ of the cervix. The serum tumor marker CA 19-9 was submitted for central testing and review. Protocol therapy was required to begin 3 to 8 weeks after resection and within 5 days of randomization. All patients required written and informed consent according to institutional and federal guidelines. All institutions were required to have current institutional review board approval on file with their respective group prior to registration of any patients. The trial was routinely monitored for excessive toxicity by the RTOG Data Monitoring Committee, which functions independently of the RTOG.

• Primary tumor

•    T1: Tumor limited to the pancreas and 2 cm or less in greatest dimension.

•    T2: Tumor limited to the pancreas and more than 2 cm in   greatest dimension.

•    T3: Tumor extends directly into  any of the following: duodenum,  bile duct, peripancreatic tissues.

•    T4: Tumor extends directly into  any one of the following: stom  ach, spleen, colon, adjacent largevessels.

• Regional lymph nodes

•    N0: No regional lymph node  metastasis.

•    N1: Regional lymph node   metastasis.

• Distant metastasis

•    M0: No distant metastasis.

•    M1: Distant metastasis.

After undergoing tumor resection, patients were randomly assigned to either fluorouracil (group 1) or gemcitabine (group 2). Randomization was performed 3 to 8 weeks after surgery by a dynamic balancing procedure, which included stratification according to tumor diameter (<3 cm vs ≥3 cm), nodal status (negative vs positive), and surgical margins (negative vs positive vs unknown). Chemotherapy prior to chemoradiation therapy in group 1 consisted of a continuous infusion of 250 mg/m² of fluorouracil per day for 3 weeks. Chemotherapy prior to chemoradiation therapy in group 2 consisted of a 30-minute infusion of 1000 mg/m² of gemcitabine once weekly for 3 weeks. Between 1 and 2 weeks after completion of chemotherapy, chemoradiation was initiated and was the same for both groups (50.4 Gy with a continuous infusion of 250 mg/m² of fluorouracil daily throughout radiation therapy).

Another phase of chemotherapy was initiated 3 to 5 weeks after completion of chemoradiation therapy. Group 1 received 3 months of a continuous infusion of fluorouracil daily [(4 weeks on plus 2 weeks off) × 2]. Group 2 received 3 months of gemcitabine [(3 weeks on plus 1 week off) × 3]. Radiation therapy was delivered in 28 fractions (5 days per week) to the tumor bed and regional nodes. The tumor bed was defined by preoperative CT imaging. Local pancreatic, celiac, mesenteric, periaortic, pancreatic, duodenal, and hepatic portal lymph nodes were included in the radiation therapy fields.¹⁴ After an initial dose of 45 Gy, the final dose of 5.4 Gy was limited to the tumor bed as defined by the preoperative tumor volume. At least 4-MV photons and a minimum 3 to 4 field approach was used. Doses were limited to less than 60% of hepatic volume receiving more than 30 Gy. At least two-thirds of one functioning kidney was spared from radiation therapy fields and the spinal cord was limited to less than 45 Gy. Prospective quality assurance of radiation therapy was required for all. This was inclusive of submission of a preoperative abdominal CT scan and radiation therapy fields to be used for central review and approval prior to the start of chemoradiation.

A follow-up visit was required at 2 to 4 weeks after completion of chemoradiation and prior to the start of the second phase of chemotherapy after chemoradiation therapy. Thereafter, follow-up occurred at 3-month intervals for 1 year, then at 6-month intervals for 3 years, and yearly thereafter. The last date of patient follow-up was August 18, 2006. Follow-up consisted of physical examination, complete blood cell count, liver function testing, chest x-ray, and CT scanning as clinically indicated. Elevation in CA 19-9 level in and of itself was not to be considered a criterion for disease recurrence.

Survival for all patients and for patients with pancreatic head tumors were the primary end points. Secondary end points were disease-free survival and toxicity, which was scored per the US National Cancer Institute’s Common Toxicity Criteria version 2.0. All end points were prespecified in the original design of the trial and all analyses were conducted on an intention-to-treat basis. Failure for overall survival was defined as death due to any cause and was measured from date of randomization to date of death or last follow-up for censored patients. Failure for disease-free survival was defined as local, regional, or distant relapse, appearance of a second primary, or death due to any cause and measured from date of randomization to date of first failure or last follow-up for censored patients. Patients who did not have a failure for overall survival or disease-free survival were censored as of their last follow-up visit.

Patients were stratified by nodal status (uninvolved vs involved), tumor diameter (<3 cm vs ≥3 cm), and surgical margin status (negative vs positive vs unknown). The permuted block randomization method was used with patient factors balanced according to the permuted block randomization method.¹⁵ At an original expected accrual of 5 patients per month, 330 patients were targeted to detect a 33% reduction in the hazard rate of overall survival for the chemoradiation plus gemcitabine group compared with the chemoradiation plus fluorouracil group (increase in median survival from 18 to 27 months; hazard ratio [HR], 0.67) with 80% power and a 2-sided α level of .05, assuming an exponential distribution. In early 2001, based on unexpectedly rapid accrual (13 patients per month) and after approval by the US intergroup, the RTOG data monitoring committee, and the National Cancer Institute, the sample size was increased to find a smaller treatment effect with more power. Four hundred seventy analyzable patients would provide 85% power to detect a 28% decrease in the HR of overall survival (increase in median survival from 18 to 25 months; HR, 0.71) for all patients and 80% power in patients with pancreatic head lesions.

All analyses were performed using SAS version 9.1 statistical software (SAS Institute Inc, Cary, North Carolina). The χ² tests were used to compare differences among pretreatment characteristics between treatment groups. For CA 19-9, the variable was categorized as less than 180 U/mL vs 180 U/mL or higher to be consistent with a protocol-specified CA 19-9 analysis to be performed subsequently and based on published literature.¹⁶ The z tests were used to test for differences in binomial proportions of grade 3 toxic effects or higher (worst overall, worst nonhematologic, and worst hematologic). Overall and disease-free survival were estimated univariately using the Kaplan-Meier method¹⁷ and treatment groups were compared using the log-rank test.¹⁸ Multivariate analyses were performed with Cox proportional hazard models¹⁹ to test for treatment differences (between groups) while adjusting for the stratification variables of nodal involvement (no vs yes), tumor diameter (<3 cm vs ≥3 cm), and margin status (negative vs positive and negative vs unknown), as well as any other variables that were imbalanced between the treatment groups. A tumor diameter of 3 cm was used for stratification based on a prior large institutional study.⁹ All tests were performed at a significance level of .05. All variables are coded such that an HR of more than 1 indicates an increased risk for the second level of the variable and an HR of less than 1 indicates a benefit for the second level of the variable.

Between July 1998 and July 2002, 538 patients at 164 institutions were randomized. Based on additional information obtained per protocol after randomization, 87 patients (16%) were ineligible due to lack of submission of serum tumor marker CA 19-9 (n = 24), therapy started more than 8 weeks after resection (n = 19), staging was incomplete or ineligible (n = 17), caloric intake did not meet protocol requirements (n = 11), ineligible primary tumor site (n = 8), eligibility unable to be confirmed (n = 3), history of previous radiation therapy or cancer (n = 3), and/or patients withdrew consent (n = 2). These patients and reasons for ineligibility were evenly distributed between treatment groups. Of the remaining 451 patients, 230 were randomly assigned to chemoradiation plus fluorouracil and 221 to chemoradiation plus gemcitabine (Figure 1). Demographic factors were similar between groups with the exception of T stage (Table 1). The chemoradiation plus gemcitabine group had a higher percentage of patients with T3 or T4 disease (81%) compared with those randomized to the chemoradiation plus fluorouracil group (70%; P = .01). The majority (86%) of tumors were of the pancreatic head. Sixty-six percent of patients had nodal metastases. Thirty-four percent of patients had microscopically positive margins and 25% had pathology reports with no comment regarding the status of surgical margins (unknown). The protocol specified that the initial treatment results be reported after either there had been a total of 207 deaths in both groups or there had been a total of 114 deaths in the fluorouracil group. This analysis has been performed with 348 deaths, with 181 deaths occurring in the fluorouracil group. All surviving patients have been followed up for a minimum of 4.1 years.

Grade 3 or higher toxic effects are summarized in Table 2. Hematologic toxicity was more common in the gemcitabine group. The incidence of toxicity of grade 3 or higher was 58% in the gemcitabine group vs 9% in the fluorouracil group (P < .001). There were no differences in febrile neutropenia or infection between treatment groups. There was a single grade 5 event in the gemcitabine group due to nonneutropenic infection. There was no significant difference in nonhematologic toxic effects between treatment groups. The most common grade 3 or higher nonhematologic toxicity was diarrhea, which affected 19% of the patients in the fluorouracil group and 15% of the patients in the gemcitabine group. Other common grade 3 or higher toxic effects were mucous membrane or stomatitis (fluorouracil group, 15%; gemcitabine group, 10%) and nausea and vomiting (fluorouracil group, 11%; gemcitabine group, 10%). Other toxic effects affecting less than 10% of the population were gastrointestinal tract (nondiarrhea, nonnausea, and nonvomiting), neurological, cardiac, pain, weight loss, and metabolic in nature. Worst overall events of grade 3 or higher occurred in 79% of patients in the gemcitabine group vs 62% in the fluorouracil group (P = .001) and was due to the difference in hematologic toxicity.

Of the 230 patients assigned to the fluorouracil group, 200 (87%) completed chemotherapy as planned, and 199 (86%) completed radiation therapy as planned; while of the 221 patients assigned to the gemcitabine group, 198 (90%) completed chemotherapy as planned and 193 (88%) completed radiation therapy as planned.

The assumption of proportionality of hazards was tested and satisfied. With a median follow-up of 1.5 years for all patients and 4.7 years for surviving patients, there was no difference in overall or disease-free survival between treatment groups (Figure 2). Treatment crossover was not allowed. Among the 230 patients in the fluorouracil group, 113 (49%) received additional salvage chemotherapy at the time of tumor recurrence; with 82% (93/113) receiving gemcitabine as salvage therapy. Among the 221 patients in the gemcitabine group, 93 (42%) received additional salvage chemotherapy at the time of tumor recurrence; with 62% (58/93) receiving additional gemcitabine as salvage therapy.

Patients with pancreatic head tumors had a median survival of 20.5 months and a 3-year survival of 31% in the gemcitabine group vs 16.9 months and 22% in the fluorouracil group (HR, 0.82 [95% CI, 0.65-1.03]; P = .09; Figure 2). After adjusting for the protocol-specified stratification variables of nodal status, which strongly affected survival (P = .001), tumor diameter, and surgical margin status on multivariate analysis for patients with pancreatic head tumors, the treatment effect yielded an HR of 0.80 (95% CI, 0.63-1.00; P = .05) toward improved survival for the gemcitabine group (Table 3). Although T stage was imbalanced between the groups, it was not a significant factor on multivariate analysis. The χ² analysis showed no association between unknown margin status with volume of patients treated by an institution in this study; with institutional volume being categorized as low (1-3 patients), medium (4-5 patients), or high (≥6 patients). In addition, Kaplan-Meier analysis showed no difference in survival between patients with unknown vs negative margin status.

Pattern of tumor relapse was recorded on the site of the first relapse only, and these sites were categorized as local, regional, or distant (Table 4). Local relapse was defined as recurrence at the primary resection site; regional relapse was defined as recurrence in regional lymph nodes associated with the primary resection site, and all other relapses were defined as distant. Distribution of relapse was similar among all patients and among patients with pancreatic head tumors. Local relapse occurred in 28% of patients in the fluorouracil group and 23% in the gemcitabine group. Regional relapse was similar in both groups at 7% to 8%. Distant relapse was higher than 70% in both groups.

The present study demonstrates that patients with pancreatic head cancers treated with gemcitabine and chemoradiation had a median survival of 20.5 months and a 3-year survival rate of 31% compared with those treated with fluorouracil and chemoradiation who had a median survival of 16.9 months and a 3-year survival rate of 22% (P = .09). After multivariate analysis for prognostic variables, the treatment effect yielded an HR of 0.80 (95% CI, 0.63-1.00; P = .05). While gemcitabine was associated with increased grade 3 or 4 hematologic toxicity, patients were asymptomatic and there was not an increase in febrile neutropenia or infection. Patients receiving gemcitabine did not have an increase in nonhematologic toxicity and there were no differences in the ability to complete therapy.

The results of this trial (RTOG 97-04) are consistent with other phase 3 adjuvant trials evaluating the effect of gemcitabine, such as the recently reported CONKO-001 (Charité Onkologie) trial.²⁰ The CONKO-001 trial randomized patients with completely resected pancreatic cancer to 6 cycles of gemcitabine vs surgery alone and demonstrated a statistically significant improvement in disease-free survival without an improvement in overall survival.

In both the CONKO and RTOG trials, salvage therapy with gemcitabine administered at relapse may have reduced the effect observed for overall survival. These results cannot directly be compared, particularly given the fundamental differences in treatment approaches, design, and patient characteristics,²¹ including frequency of R0 resections (42% in RTOG vs 83% in CONKO), and the CONKO-001 requirement for CA 19-9 to be lower than 2.5 × upper limit of normal (approximately 90 U/mL). The RTOG 97-04 study did not exclude patients by CA 19-9, with 13% of patients having values of 180 U/mL or higher and 21% having values higher than 90 U/mL.

The RTOG 97-04 study used a continuous infusion of fluorouracil instead of bolus. Studies in other gastrointestinal cancers have demonstrated a benefit of the continuous infusion of fluorouracil compared with bolus in combination with radiation therapy and safety in pancreatic cancer has been demonstrated.^22-24 In comparing RTOG 97-04 with previous adjuvant trials using chemoradiation, this study is the first phase 3 trial requiring prospective quality assurance of radiation therapy. This was influenced by a prior US intergroup phase 3 trial experience with upper abdominal radiation therapy reporting a 35% risk of major or minor deviation from protocol-prescribed radiation therapy, most of which required correction prior to the start of radiation therapy due to the risk of toxic effects on critical organs or failure to treat the appropriate target volumes.²⁵ The RTOG 97-04 study compares favorably with the outcome in similar phase 3 trials using chemoradiation in patients with pancreatic head adenocarcinoma (Table 5).^6,7 In RTOG 97-04, 75% of patients had T3 or T4 disease, 66% had lymph node–positive disease, 34% had microscopically positive margins, and the associated overall local recurrence rate was 26%. The Gastrointestinal Study Group (GITSG) trial⁶ included only patients with negative surgical margins and only 30% had lymph node–positive disease; the local recurrence rate in the chemoradiation group was 47%. The European Organization for the Research and Treatment of Cancer (EORTC) trial⁷ included only patients with T1 or T2 disease, 47% had lymph node–positive disease, and 19% had microscopically positive margins; the local recurrence rate in the chemoradiation group was 51%. The survival and local relapse outcomes in RTOG 97-04 compare well despite having a much greater proportion of patients with T3 or T4 disease, lymph node–positive disease, and microscopically positive margins compared with these trials. To more easily compare future adjuvant pancreatic trials, uniform eligibility criteria are needed. The RTOG 97-04 study is the first prospective, phase 3 trial evaluating the correlation of postoperative CA 19-9 values to outcome. The results of this planned secondary analysis will help determine whether postoperative CA 19-9 values higher than 180 U/mL should be a stratification variable.

The RTOG 9704 study was powered to analyze survival for patients with pancreatic head tumors. Patients with pancreatic head tumors may present with different clinical features at diagnosis such as obstructive jaundice compared with pancreatic body or tail tumors. A different operation is typically used for pancreatic head vs pancreatic tail tumors (pancreaticoduodenectomy vs distal pancreatectomy). Patients with resected pancreatic body or tail tumors have been generally found to have a worse outcome than those with pancreatic head tumors.^26,27

The finding of 25% of patients having unknown surgical margin status warrants further comment. Although this rate of unknown margin status is higher than that reported in previous trials,^6,7,28,29 in these trials a lack of attainment and review of a patient's pathology report was categorized as unknown. In contrast, in RTOG 97-04 in which the review of pathology reports was 100%, unknown margin status was specifically defined as having no comment regarding margins in the pathology report. Interestingly, there was no difference in survival observed among patients with unknown or negative surgical margin status.

The RTOG 97-04 study included chemoradiation therapy with fluorouracil in both treatment groups. Therefore, the effect of chemoradiation therapy cannot be assessed. The European Study Group for Pancreatic Cancer 1 (ESPAC-1) phase 3 adjuvant trial questioned the role of adjuvant chemoradiation.^28,29 However, the ESPAC-1 trial has been specifically critiqued,^30-34 including with regard to its lack of specification of radiation therapy guidelines, lack of radiation therapy quality assurance, and wide range of radiation therapy doses—as high as 60 Gy.³⁵ The results of the fluorouracil plus chemoradiation therapy after surgery group of ESPAC-1 appear inferior to those seen in the EORTC and GITSG trials (13.9 months vs 17.1 and 21 months, respectively) which used similar fluorouracil plus radiation therapy. Furthermore, in ESPAC-1, 63% of patients developed a local recurrence. The European Study Group for Pancreatic Cancer 3 (ESPAC-3) is currently comparing adjuvant fluorouracil and gemcitabine in patients with resected pancreatic cancer; while an EORTC phase 2/3 trial evaluating the feasibility and potential benefit of combining concurrent gemcitabine plus radiation therapy in the adjuvant setting is ongoing.

The RTOG 97-04 study was not designed to determine the role of adjuvant radiation because chemoradiation was given in both groups. Based on CONKO-001 and ESPAC-1, a single-agent modality with gemcitabine or fluorouracil could be considered an option for patients with resected pancreatic cancer and negative margins. However, patients with resected pancreatic cancer develop both systemic and local recurrence. Unfortunately, systemic therapies with fluorouracil and gemcitabine have only modest activity. Therefore, it may not be possible to demonstrate a survival benefit from a local modality such as fluorouracil and chemoradiation therapy when fluorouracil and gemcitabine have only a modest impact on delaying systemic recurrence. The ESPAC-1 study demonstrated a local recurrence of 63%. Even though one-third of the patients in the RTOG 97-04 study had positive margins, the local recurrence rate in the gemcitabine and chemoradiation group was only 23%. Laboratory correlative studies from RTOG 97-04 are ongoing and are evaluating molecular genetic alterations that promote local and systemic relapse. Future trials should emphasize novel systemic treatments to reduce systemic metastases and modern image-guided radiation to prevent local recurrence while reducing radiation-related toxic effects.

Corresponding Author: William F. Regine, MD, University of Maryland, 22 S Greene St, Baltimore, MD 21030 (wregine@umm.edu).

Author Contributions: Dr Regine had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Regine, Abrams, Hoffman, Benson, Macdonald, Willett, Rich.

Acquisition of data: Regine, Winter, Abrams, Hoffman, Safran.

Analysis and interpretation of data: Regine, Winter, Abrams, Safran, Hoffman, Benson, Macdonald, Kudrimoti, Haddock, Willett.

Drafting of the manuscript: Regine, Winter, Abrams, Safran, Hoffman, Benson.

Critical revision of the manuscript for important intellectual content: Regine, Winter, Abrams, Safran, Hoffman, Benson, Konski, Macdonald, Kudrimoti, Fromm, Haddock, Schaefer, Willett, Rich.

Statistical analysis: Winter.

Administrative, technical, or material support: Hoffman, Konski, Benson, Kudrimoti, Fromm, Haddock, Schaefer, Willett, Rich.

Study supervision: Regine, Abrams, Safran, Hoffman, Benson, Macdonald, Willett, Rich.

Funding/Support: The National Cancer Institute supported the cooperative clinical trials groups, which participated in this study, including the Radiation Therapy Oncology Group (RTOG), the Eastern Cooperative Oncology Group, and the Southwest Oncology group via grants U10 CA21661, U10 CA37422, and U10 CA32115. Eli Lilly, the makers of gemcitabine, provided financial support to RTOG Headquarters and the statistical center.

Role of the Sponsor: The sponsors of this trial were involved in the design of the study but were not involved in the conduct of the study, collection, management, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript.