Adjuvant Multimodality with Gemcitabine versus 5-Fluorouracil in Pancreatic Cancer: Status of U.S. Intergroup Trial

Walter Sall, MD
Abramson Cancer Center of the University of Pennsylvania
Last Modified: November 12, 2003

Presenter: William F. Regine, M.D.
Affiliation: University of Maryland

Even with potentially curative resection, 5 year survival for pancreatic cancer patients is less than 20%. Approximately 25 years ago, the Gastrointestinal Study Group found a small survival advantage with the use of adjuvant 5-FU chemotherapy and radiation. Despite the use of split course radiation therapy and outdated chemotherapy dosing regimens in this study, it continues to be the primary large phase III trial on which the standard of care for this disease is based in the United States. Gemcitabine is the first new chemotherapy agent to be approved for pancreatic cancer in 35 years. Adjuvant 5-FU is being compared to gemcitabine in the RTOG 97-04 study, a phase III Intergroup study. The study accrued more than 500 patients from July 1998 - July 2002.

Study Design:
Patients with pancreatic adenocarcinoma undergoing potentially curative resection were eligible. Patients were randomized to receive either gemcitabine or 5-FU chemotherapy before and after concurrent 5-FU/radiation. Pre-radiation chemotherapy consisted of either 5-FU continuous infusion, 250mg/m2/d for 3 weeks or gemcitabine, 1000mg/m2, once weekly for three weeks. Chemoradiation for both arms consisted of 50.4 Gy in 180cGy fractions using a multi-field technique with 250mg/m2/d continuous infusion 5-FU. Post-radiation chemotherapy consisted of 3 months of continuous infusion 5-FU or 3 months of gemcitabine.

Entry Criteria:
Gross total resection was required. Exclusion criteria included prior chemotherapy or radiation, non-adenocarcinoma histology, or non-pancreatic primary tumors, including ampullary or duodenal primaries. All radiation fields underwent central review prior to initiation of radiation.

Primary endpoints included overall survival, progression-free survival and toxicity. Secondary endpoint included correlation of CA19-9 levels and outcome.

Expected Accrual:
330 patients were expected to accrue over 5 years. The study was designed to have an 80% power to detect a hazard ratio of 1.5 for overall survival.

Activated in July, 1998, 330 patients had accrued by May, 2001. 15-20% of patients had tumors of the body or tail of the pancreas. Because of concerns about the high numbers of body and tail lesions, which may have a different natural history than pancreatic head lesions, accrual goals were raised to 500 patients. Accrual was therefore closed in July 2002 with 538 patients. Crude data shows approximately 2/3 of patients are node (+) and 1/3 have positive margins. Data analysis is not complete. Results are expected in the Summer of 2004.

This study may have far reaching implications for the treatment of pancreatic cancer. It may change the 5-FU based treatment paradigm for resected pancreatic cancer for the first time in 25 years. Hopefully, given the impressive activity of gemcitabine in this disease, sequencing of this drug with radiation will lead to improved outcomes. The results of this study are eagerly anticipated.