Dose-escalated pancreas radiotherapy for unresected pancreatic adenocarcinoma: Patterns of care and survival in the United States.

676 Background: Randomized trials have not shown an overall survival (OS) benefit to adding radiotherapy (RT) to chemotherapy for patients with locally advanced pancreatic adenocarcinoma (LAPC). However, trials such as LAP07 did not incorporate escalated-dose RT (EDR), which may confer longer survival. The adoption of EDR across the United States and associated outcomes are unknown. Methods: We queried the National Cancer Data Base for non-surgically managed patients with LAPC diagnosed between 2004 and 2019. RT with biologically effective doses (BED 10 ) > 70 and ≤ 200 Gy were labeled EDR and conventional-dose RT (CDR) for 39-70 Gy doses. Associations with receipt of EDR and with OS were identified using multivariate analysis (MVA) logistic and Cox regressions, respectively. Bonferroni corrections were applied. Cochran-Armitage and Mann-Kendall trend tests were performed to assess trends in use of RT. Results: Of 64,303 patients, the most common treatments were chemotherapy alone (CT) (71%), chemoradiation (CRT) (27%), and pancreas RT alone (2%). 18,970 patients received pancreas RT, of which 91% was CDR (median dose 50.4 Gy; median 28 fractions) and 9% was EDR (50 Gy; 5 fractions). Use of pancreas RT increased from 14% in 2004 to a peak of 18% in 2010, decreased to a nadir of 13% in 2016, and subsequently increased to 15% by 2019 ( P trend < 0.001). EDR use increased from 7% in 2004 to 22% in 2019 ( P trend < 0.0001). Median BED 10 increased from 53 to 59 Gy ( P trend < 0.001). Of patients receiving pancreas RT, use of intensity-modulated RT (IMRT) and stereotactic body RT (SBRT) respectively increased from 25% and 2% in 2004 to 60% and 27% in 2019, while use of 3-D conformal RT (3D-CRT) decreased from 64% to 9% ( P trend < 0.02 for all comparisons). On MVA logistic regression, primary tumor location in the body/tail vs. head (aOR 1.22, 95% CI 1.07-1.40; P = 0.003) associated with greater EDR receipt, whereas T3-4 vs. T2 disease (aOR 0.81, CI 0.71-0.92; P = 0.002) associated with lesser receipt. At a median follow up of 59.1 months (CI 57.5-61.0), median OS estimates for CDR and EDR were 10.2 months (CI 10.1-10.3) and 13.3 months (CI 13.2-13.5; P < 0.0001), respectively. On MVA Cox regression, N1 vs. N0 disease (HR 1.08, CI 1.06-1.10; P < 0.001) correlated with higher risk of death, whereas CRT vs. CT (HR 0.83, CI 0.81-0.85; P < 0.001) correlated with lower risk of death. Subset MVA of 14,634 CRT patients correlated higher RT dose—as both a categorical variable (EDR vs. CDR; HR 0.84, CI 0.79-0.90; P < 0.001) and continuous variable (BED 10 ; HR 0.994, CI 0.992-0.995; P < 0.001)—with lower risk of death. Conclusions: Fewer than 1 in 6 patients with unresected LAPC received pancreas RT, despite a nominal increase in utilization in recent years. Although retrospective, these NCDB data suggest longer OS with the addition of RT for unresected LAPC, suggesting continued unmet need. EDR is associated with longer survival vs. CDR.