Article Abstract

The role of sequential radiation following adjuvant chemotherapy in resected pancreatic cancer

Authors: Ashwin Shinde, Vivek Verma, Richard Li, Scott Glaser, Mustafa Raoof, Susanne G. Warner, Vincent Chung, Gagandeep Singh, Yi-Jen Chen, Karyn A. Goodman, Arya Amini


Background: Optimal adjuvant therapy for resectable pancreatic cancer remains controversial. The Radiation Therapy Oncology Group (RTOG) 0848 is evaluating adjuvant chemotherapy (CT) with or without sequential chemoradiation in the absence of disease progression. In the absence of completed RTOG 0848 data, we addressed the question through a retrospective analysis.
Methods: The National Cancer Database was queried for non-metastatic, histologically-confirmed pancreatic carcinoma status post resection and adjuvant CT. Patients receiving neoadjuvant therapy were excluded. Multivariable logistic regression was performed to identify factors associated with receipt of sequential radiation therapy (RT) starting at least 4 months after CT, with generation of propensity scores. Log-rank analysis assessed univariate overall survival (OS), while multivariable Cox proportional hazards modeling examined multivariate OS. Six-month landmark analysis was performed to allow for completion of all planned adjuvant therapy.
Results: Of 15,661 patients, 14,167 (90.5%) underwent CT alone, and 1,494 (9.5%) received RT. Patients receiving RT had nodal disease and R+ resection more often (P<0.05). Median follow-up was 53.6 months. Sequential RT after adjuvant CT improved outcomes compared to adjuvant CT alone (3-year OS 44.6% vs. 35.3%; P<0.001). On multivariate analysis, sequential RT continued to improve OS (HR =0.68; 95% CI, 0.63–0.73; P<0.001). After propensity score adjustment, the benefit of sequential RT was maintained (HR =0.68, P<0.001).
Conclusions: Sequential RT after adjuvant CT for resected pancreatic cancer was associated with improved survival. The final results of RTOG 0848 are eagerly awaited to provide definitive evidence in this clinical scenario.