Relationship between treatment duration and outcomes in patients receiving concurrent chemoradiotherapy (CCRT) and surgery for locoregionally advanced (LRA) esophageal and gastroesophageal junction (E/GEJ) adenocarcinoma (ACA).
e14665 Background: Retrospective studies in breast, colon, and head and neck cancer have suggested that delays in administration of adjuvant chemotherapy result in inferior outcomes. We explore this question in patients with LRA E/GEJ ACA. Methods: From 11/99 to 7/06, 152 patients with cT3, N1, or M1a disease were enrolled on one of two clinical trials at the Cleveland Clinic. Two courses of CCRT consisting of 30Gy radiation (@1.5Gy BID) with continuous infusion cisplatin (20mg/m2/day x 4 days) and 5Fu (1000mg/m2/d x 4 days) were given both before and after surgical resection. In the second trial, 75 patients also received gefitinib during the 4 week induction and for a total of two years postoperatively. Using recursive partitioning analysis (RPA), we retrospectively explored the relationship between treatment duration and loco-regional control (LRC), distant metastatic control (DMC), freedom from recurrence (FFR), and overall survival (OS) in the 115 patients (76%) who completed all treatment. Outcomes were estimated using the Kaplan-Meier method and compared among groups using the log-rank test. Results: RPA analysis identified three groups of patients: 19 patients in whom resection occurred ≤45 days from the start of CCRT to resection (short treatment), 63 patients who underwent resection >45 days from the start of CCRT and required <50 days to complete treatment after surgery (intermediate treatment), and 33 patients who underwent resection >45 days from the start of CCRT but required ≥50 days to complete treatment after surgery (prolonged treatment). With a median follow-up of 90 (range 57-126) months, we found that patients with shorter treatment times had better 5 year DMC [64% (short) v 26% (intermediate) v 16% (prolonged); p=0.004], FFR [57% (short) v 23% (intermediate) v 16% (prolonged); p=0.015)], and OS [42% (short) v 30% (intermediate) v 12% (prolonged); p=0.08]. LRC was not different between the groups. Conclusions: Treatment delays in patients receiving multimodality therapy for LRA E/GEJ ACA may result in a greater risk of recurrence and decreased survival.