Heart Dose and Outcomes in Radiation Treatment for Esophageal Cancer.

ABSTRACT

Purpose Studies have shown that radiation dose to the heart may be associated with worse outcomes in patients receiving chemoradiation for lung cancer. As esophageal cancer radiation treatment can result in relatively high cardiac doses, we evaluated a single-institution database of patients treated for esophageal cancer for heart dose and outcomes. Methods We retrospectively reviewed 59 patients with stage IIA-IIIB esophageal cancer treated with neoadjuvant chemoradiation to 50.4 Gy followed by esophagectomy from 2007-2015. Patient demographics and outcome data, including pathological response, local recurrence, distant metastases, and overall survival, were obtained. Mean heart dose (MHD), heart V5, V40, and V50, were calculated. Differences in patient characteristics between the three radiation therapy modalities three-dimensional (3D) conformal radiotherapy (3D-CRT), intensity modulated radiotherapy (IMRT), and proton beam radiation therapy (PBT) were tested using non-parametric Kruskal-Wallis (K-W) analysis of variance (ANOVA). Patient characteristics and heart dosimetric parameters were screened by univariate Cox regression for an association to overall survival, and univariate predictors (p < 0.05) were then selected as inputs into a multivariate Cox regression model using stepwise backward elimination. Kaplan-Meier risk-stratified survival curves were plotted for the best univariate or multivariate Cox model variables. An exploratory subgroup univariate Cox regression was conducted in each of the treatment modalities (proton, IMRT, 3D-CRT). Results The median follow-up was 20 months. The median overall survival was 73 months. Eleven patients (20%) experienced a complete pathologic response (pCR). Only two patients (4%) experienced a local recurrence. On univariate analysis, predictors of survival were age, prior radiation, pathologic response in involved lymph nodes, and tumor length post-treatment. On a multivariate analysis, only pathologic nodal response (yN) remained significant (p = 0.007). There was no relationship between any heart dosimetric variables analyzed and any clinical outcomes. Conclusions In this retrospective review, radiation dose to the heart was not associated with inferior treatment outcomes in patients receiving trimodality therapy for esophageal cancer.