Purpose: To determine the maximal tolerated dose of radiation delivered to the primary tumor bed, in combination with full-dose gemcitabine (1000 mg/m(2) weekly x 3), after resection of pancreatic cancer.
Methods and materials: Patients with resected pancreatic carcinoma and poor prognostic features, a positive resection margin, or involved lymph nodes were eligible. Radiotherapy (RT) was directed at the preoperative tumor volume with a conformal technique. Regional lymph node basins were not included. The initial starting radiation dose was 24 Gy in 1.6-Gy fractions. Escalation was achieved by increasing the fraction size in 0.2-Gy increments, keeping the duration of RT to 3 weeks. Gemcitabine was given i.v. for 30-40 min at a dose of 1000 mg/m(2) before RT on Days 1, 8, and 15 of a 28-day cycle. After completion of RT and chemotherapy, an additional cycle of gemcitabine was delivered.
Results: Between November 1997 and October 2001, 32 patients were entered: 30 after Whipple resection (positive margins in 2, positive nodes in 22, and both in 6), 1 after distal pancreatectomy, and 1 after incomplete resection of a tumor involving the body (both patients with positive margins and nodes). Treatment was well tolerated. Of the 32 patients, 27 completed all protocol therapy and 29 maintained their pretreatment weight within 5%. Five patients experienced dose-limiting toxicity, four with Grade 3 vomiting requiring hospitalization and one fatal toxicity secondary to pneumonia/sepsis. At the final radiation dose level (42 Gy), 2 patients experienced GI dose-limiting toxicity. At the 39-Gy-dose level, 5 of 6 patients were without dose-limiting toxicity. Isolated local or regional progression was documented in 1 patient. Distant progression was documented in 26 of 32 patients (6 with concurrent local or regional progression). The median survival was 16.5 months (95% confidence interval 12.3-19.9)
Conclusion: The results of our study indicate that the maximal tolerated radiation dose, administered using conformal techniques targeted to the tumor bed, is 39 Gy. In this high-risk population, data on locoregional control suggest that the reduction in radiation dose and field size minimizes toxicity and does not result in excess failures at these sites.