Pattern of local-regional relapse and survival after bimodality therapy for esophageal cancer: Implications for the surveillance strategy.

Abstract

4083 Background: Evidence for definitive chemoradiotherapy (bimodality therapy [BMT]) has been established for patients with esophageal and gastroesophageal junction cancer (EGEJC) who do not qualify for surgery. Surveillance for these patients is often recommended but the literature lacks guidance for an evidence-based surveillance strategy after BMT. Methods: We analyzed 276 patients with EGEJC who underwent BMT and had pre- and postchemoradiation endoscopic biopsies and imaging studies available for review. Patients who underwent planned surgery or salvage surgery (SS) within 6 months from BMT were excluded. We reviewed the pattern of relapse over time. Local-regional disease (LRD) after BMT was classified as regional disease (RD) or luminal-only disease (LD). Overall survival (OS) probabilities were estimated using the Kaplan-Meier method and compared using the log-rank test. Results: For 276 patients, the median follow-up time was 53.0 months (95% confidence interval [CI], 47.3-58.7). A total of 184 (66.7%) patients had a persistent disease or relapse after BMT: 120 distant metastases (43.5% of 276) and 64 LRD (23.2% of 276). Of 64 LRD, 58 (91%) were diagnosed within 2 years of BMT and 63 (98%) were diagnosed within 3 years (see Table). Twenty-three of 64 LRD patients underwent SS. For patients with SS, the median OS time from diagnosis of LRD was 58.0 months (95% CI, not reached), and that for patients without SS was 9.0 months (95% CI, 7.3-10.7); this difference was highly significant (p < 0.001). Conclusions: Our data suggest that 91% of LRD occurred within 2 years after BMT and the OS with SS for LRD was better than that without SS. These data can contribute to the development of an evidence-based surveillance strategy. [Table: see text]