Multimodality treatment for distal esophageal (DE) and gastroesophageal junction (GEJ) adenocarcinoma (ACA) with celiac lymph node (CLN) involvement

Abstract

4574 Background: CLN involvement is a predictor of poor outcome in patients (pts) with DE or GEJ ACA. Pre-treatment identification of such patients depends on clinical staging including endoscopic ultrasound (EUS), computerized tomography (CT), and positron emission tomography (PET). This review of our CLN positive pts was undertaken to define the impact of clinical staging on prognosis after concurrent chemoradiotherapy (CCRT) and surgery. Methods: We retrospectively identified all pts with DE or GEJ ACA, CLN involvement by EUS, CT or PET, and no evidence of distant hematogenous metastases, who were treated with the same CCRT and surgery protocol at the Cleveland Clinic. Pts not staged with all three modalities were excluded. Induction CCRT consisted of radiation (30 Gy at 1.5 Gy bid), and chemotherapy with cisplatin (20 mg/m2/d) and fluorouracil (1000 mg/m2/d) both given as continuous intravenous infusions during the first 4 days of radiation. Surgery was performed in 4–6 weeks and identical CCRT was planned 6–10 weeks post-operatively. Outcomes examined included locoregional control (LRC), distant metastatic control (DMC), freedom from recurrence (FFR) and overall survival (OS). Results: Between 2/00 and 12/07, 54 pts with clinically staged CLN involvement were treated with this protocol. CLN involvement was found by EUS in 70%, CT in 69% and PET in 54% of pts. No single or combination of clinical staging tests proved predictive of outcome except for LRC which was worse in pts with all 3 tests positive (p = 0.008). With a median follow-up of 27 (range 8–71) months, the 2-year Kaplan-Meier projected LRC is 87%, DMC 18%, FFR 18%, and OS 28%. DE (vs. GEJ) primary site predicted for better DMC (p < 0.001), FFR (p = 0.002), and OS (p = 0.025). Negative surgical margins predicted for better DMC and FFR (p=0.005 both outcomes). Only tumor location remained significant in multivariable analysis. Conclusions: CLN involvement portends a poor but not hopeless prognosis after multimodality therapy. Despite excellent LRC, distant failure predominates. DMC, FFR, and OS were worse in pts with GEJ primaries, but were independent of how the CLN involvement was clinically identified. No significant financial relationships to disclose.