Differing Radiation Doses Associated with Improved Overall Survival in Trimodality versus Bimodality Treatment for Esophageal Cancer

Abstract

The CROSS (Chemoradiotherapy for Oesophageal Cancer Followed by Surgery Study) trial published in 2012 demonstrated impressive overall survival (OS) for esophageal carcinoma patients treated with trimodality therapy involving preoperative chemoradiation followed by esophagectomy. This study notably used a lower dose of radiation (41.4 Gy) compared to previous trials (50.4 Gy). A randomized study comparing these two doses is likely not feasible; therefore, we studied trends and outcomes of 41.4 Gy and 50.4 Gy using the National Cancer Data Base (NCDB). Patients diagnosed with localized esophageal adenocarcinoma or squamous cell carcinoma from 2004 to 2015 who received definitive chemoradiation (bimodality therapy) or trimodality therapy were identified in the NCDB. Only patients who received radiation doses of 41.4 or 50.4 Gy were included. Kaplan-Meier and Cox regression analyses were performed to compare OS between patients receiving 41.4 Gy and 50.4 Gy as part of bimodality or trimodality therapy. Of the 9,671 patients who met criteria for inclusion, only 6.8% received 41.4 Gy during 2004-2011, increasing to 11.6% between 2012 and 2015 (p<0.001). Just under half of the patients receiving each dose proceeded to surgery (48% for 41.4 Gy vs. 47% for 50.4 Gy). Median follow up was 21 months. For trimodality patients, median and 2-year OS for the 41.4 (n=361) and 50.4 Gy (n=4257) groups were 39.0 vs. 37.2 months and 67% vs. 64%, respectively (p=0.03). Pathologic complete response rates were similar at 7% for 41.4 Gy vs. 8% for 50.4 Gy (p=0.20). For those receiving bimodality therapy, median and 2-year OS for the 41.4 (n=423) and 50.4 Gy (n=4630) groups were 15.4 vs. 17.2 months and 34% vs. 39%, respectively (p=0.01). After controlling for significant covariates on multivariable analysis including T-stage, N-stage, Charlson-Deyo comorbidity index, age, and insurance status, 41.4 Gy was associated with improved OS in the trimodality setting (HR 0.84, 95% CI 0.71-0.98, p=0.03) but worse OS in the bimodality setting (HR 1.14, 95% CI 1.01-1.28, p=0.03). Ninety-day post-operative mortality in trimodality patients trended towards being significantly higher for 50.4 Gy (7.1%) vs. 41.4 Gy (3.9%) (p=0.056). The use of 41.4 Gy has increased over time but still remains low. This lower dose was associated with significantly improved OS in the trimodality setting, whereas 50.4 Gy was superior in the bimodality setting. These results suggest that higher radiation doses may be harmful in the trimodality setting, and dose comparisons may warrant further investigation.