Rethinking dogma: Association between radiation dose and pathologic outcomes or survival in patients receiving neoadjuvant radiotherapy for pancreatic cancer.

Abstract

716 Background: Controversy endures regarding the role of radiotherapy (RT) in operable pancreatic ductal adenocarcinoma (PDAC). Neoadjuvant studies have used a range of RT doses, with a recent surge in enthusiasm for dose-escalated RT. No randomized comparisons of the efficacy of different RT dosing regimens are available. We report pathologic response and survival in PDAC patients who received a range of neoadjuvant RT doses in a prospectively maintained single-institution dataset. Methods: This cohort includes 148 PDAC patients treated with preoperative RT. Conventionally fractionated RT (CFRT) was defined as daily dose of ≤3 Gy in ≥10 fractions, and stereotactic body RT (SBRT) as daily dose of ≥5 Gy in ≤5 fractions. Biologic effective dose (BED) was calculated as follows: BED = nd * (1 + d/α:β), where n = the number of doses, d = dose per fraction; α:β was set at 10. Major pathologic response (MPR) was defined as a College of American Pathologists (CAP) tumor regression grade of 0-1. P<0.05 was the threshold for statistical significance. Survival was calculated from diagnosis. Results: Median age was 66 years; 56% of patients were women. Most (96%) patients received chemotherapy; 57% received SBRT (BED 10 54.8-87.5 Gy), and 43% CFRT (BED 10 39.0-70.1 Gy). Mean BED 10 was significantly higher in the SBRT group (73.7 Gy vs 51.9 Gy, p = 0.05). No significant differences were noted between CFRT and SBRT in terms of tumor response (MPR in 13% vs 19%, respectively), nodal status (N0 in 60% vs 45%, respectively), or R0 resection rates (71% vs 58%); χ 2 p >0.05 for all comparisons. Mean BED 10 did not differ significantly among patients with or without MPR (67.0 vs 63.8 Gy, p=0.36) or with or without R0 resection (63.2 vs 67.5 Gy). Similarly, receiving BED10 ≥ median was not associated with a higher likelihood of MPR or N0 resection, but was associated with a lower likelihood of R0 resection (51 vs 72%, p=0.02). Binary logistic regression showed no relationship between increased BED10 and likelihood of MPR (R 2 =0.012, p=0.33). There was no difference in median overall survival (OS) between CFRT and SBRT (median OS 29.3 vs. 31.6 months, p-NS), nor were there any OS or local control (LC) differences in patients treated with doses above or below the median BED. Similarly, Cox regression demonstrated no relationship between increasing RT dose and OS or LC. Margin status was associated with OS: median OS was 34.7 (95% CI 25.0-44.3) vs 21.4 (14.0-28.9) months in patients with negative vs positive margins, respectively (p < 0.01). A trend towards improved OS was noted in patients with MPR (46.5 vs 29.6 months, p =0.08). Conclusions: No RT dose-response effects on MPR, LC, or OS were observed in a large cohort of PDAC patients treated with neoadjuvant RT, suggesting that alternative strategies to dose escalation should be explored in studies of neoadjuvant RT for PDAC.