Patterns of Recurrence in Adenocarcinoma of the Esophagogastric Junction: ERECS Trial Post Hoc Analysis.

de Pascale, Stefano; Rosati, Riccardo; Bagnardi, Vincenzo; De Pasqual, Carlo Alberto; Ferrari, Giovanni; Frassoni, Samuele; Giacopuzzi, Simone; Gualtierotti, Monica et al. · Ann Surg Oncol · 2026

retrospective_cohort · Level III

Where this comes from

Abstract

Adenocarcinoma of the esophagogastric junction (AEG) is associated with high recurrence rates despite multimodal therapy. Understanding recurrence patterns and their predictive factors is critical for optimizing risk stratification and guiding therapeutic decision-making. This post hoc analysis of the multicenter retrospective ERECS trial included patients who had Siewert type 1 or 2 AEG treated with perioperative chemotherapy (BMT) or neoadjuvant chemoradiotherapy (TMT) followed by Ivor Lewis esophagectomy between January 2018 and December 2022. Uni- and multivariable Fine and Gray regression models identified predictors of recurrence. Overall survival after recurrence (OSAR) was estimated using the Kaplan-Meier method. Of 455 patients, 7.9% experienced local recurrence (LR) and 36.9% had distant recurrence (DR) during a median follow-up period of 29.8 months. Among the DR patients, 56% had single-site (sDR) and 44% multiple-site (mDR) involvement. Lymph nodes, liver, and lungs were the most frequently involved distant sites. In the multivariable analysis, TMT and mucinous histology independently predicted LR, whereas advanced pathologic stage, poor response to neoadjuvant therapy, and perineural invasion were associated with DR. For sDR, OSAR was better than mDR (12-month OSAR: 62% vs. 42%; p = 0.037). Distant recurrence remains the principal driver of treatment failure in locally advanced AEG. Advanced pathologic stage, poor treatment response, and perineural invasion identify patients at high systemic risk who may benefit from intensified systemic therapy. In this study, post-recurrence survival was significantly better for sDR (which does not fully correspond to oligometastatic disease) than for mDR. Prospective validation of a tailored surveillance strategy for high-risk patients is warranted.