Neoadjuvant Chemoradiation Dose and Outcomes in Esophageal Cancers, a National Cancer Data Base Study

Abstract

With excellent outcomes with lower dose chemoradiation (CRT) for neoadjuvant treatment of esophageal cancer followed by surgical resection, recent interest in the optimal radiation dose has arisen, specifically whether low dose CRT is as effective as higher doses. We looked at neoadjuvant radiation dose and outcomes in the National Cancer Database (NCDB). The NCDB for esophageal cancer from years 2004-2012 was queried for patients receiving neoadjuvant CRT followed by surgical resection with curative intent. Patients were excluded if they did not have squamous cell carcinoma (SCC) or adenocarcinoma (AC), had metastatic disease, or insufficient follow up. Cox proportional hazards model was used to identify variables that predict for overall survival. Linear regression was used to determine odds ratio of receipt of treatment. We identified 7,325 patients receiving neoadjuvant CRT followed by definitive resection: 1,276 SCC and 6,049 AC. Median follow-up was 26.3 months. We examined 4 radiation dose levels: 4000-4140 cGy (n = 252), 4500 cGy (n = 2,075), 5040 cGy (n = 4,451), and 5400 cGy (n = 547). The utilization of low dose (4000-4140 cGy) radiation has increased from 5.6% in 2010 to 31.5% in 2013. There was no significant difference between groups based on age, race, gender, insurance status, and distance from facility. Age over 60 (HR = 1.09, P = 0.032), AC histology (HR = 1.19, P = 0.001), and increasing stage (HR = 1.69, P < 0.00) were significant negative predictors of survival. Patients with Medicare/Medicaid had higher risk of death than those with private insurance (HR = 1.26, P < 0.00). Treatment at an academic center had improved survival (HR = 0.86, P < 0.00) and had higher odds ratio of receiving lowest radiation dose level (OR = 1.14, P < 0.00). After adjusting for all significant predictors of survival, radiation dose level was not associated with differences in overall survival (P = 0.39) or pathologic complete response (P = 0.21). There was also no difference in 30 or 90 day survival (P = 0.75 and 0.36, respectively). In the NCDB, there was no difference in pCR or OS when using low dose neoadjuvant CRT versus higher doses. Lower dose CRT appears to be increasing in utilization, especially in academic centers.