Long-term cardiopulmonary mortality after radiation for locally advanced esophageal cancer.

Abstract

99 Background: Radiation (RT) is considered an integral component of trimodality therapy for locally-advanced esophageal carcinoma to improve locoregional control and potentially survival. However, the long-term risk of cardiopulmonary mortality (CPM) is not well-understood in this population. Methods: Patients age 20-85 with esophageal carcinoma with T3, T4 or node positive (N+) disease who underwent esophagectomy were identified within 17 Surveillance, Epidemiology, and End Results registries from 1988-2006. Patients with metastatic disease or <6 mo follow up were excluded. CPM was calculated for patients receiving vs not receiving RT and compared by the Kaplan-Meier method. The log-rank test was used for univariate associations and Cox proportional hazards model was used for multivariate analysis (MVA). Results: A total of 4,079 patients met the defined selection criteria of whom 2,408 were treated with RT, and 1,671 were not. Median age was 62.2 yrs (22-84) and follow-up was 22 mos (6-248). There was no significant difference in CPM in patients who received RT versus those who did not (p=0.8). At 10 yr, the majority of deaths were from esophageal cancer (73 with vs 78% without RT) compared to CPM (13.7 with vs 11.6% without RT). On univariate analysis ( table ), age <60, diagnosis era, and histology were significant independent predictors of CPM. On MVA, age <60 (HR 0.36) and diagnosis era (0.63 for 1994-2000 and 0.55 for 2000-2006) remained statistically significant for CPM. Conclusions: RT for esophageal cancer is not associated with an increased long-term risk of CPM in the overall population. Older age and earlier diagnosis era predict for CPM. Although survival in esophageal cancer is dominated by cancer deaths, advances in RT are still needed to prevent excess treatment-related mortality. [Table: see text]