Nodal station–mapped recurrence patterns after neoadjuvant immunochemotherapy and R0 esophagectomy for esophageal squamous cell carcinoma.

Abstract

e16121 Background: Neoadjuvant immunochemotherapy (nICT) followed by R0 esophagectomy is increasingly used for resectable esophageal squamous cell carcinoma (ESCC), yet nodal-station–level lymph node (LN) failure patterns are unclear and total harvested LNs (TLN) may not reflect lymphadenectomy adequacy. We applied AJCC 8th-edition station mapping and an in-field/out-of-field framework to characterize LN relapse and assess associations with tumor regression grade (TRG) and nodal burden. Methods: Multicenter retrospective cohort from a prospectively maintained database (six centers). Consecutive ESCC patients underwent nICT and R0 esophagectomy (Jan 2022–Dec 2024). Response: Mandard TRG good (1–2) vs poor (3–5). LN metrics included dissection extent (TLN; regional and station counts), metastatic burden (positive LNs [PLN]; positive regions/stations), and relapse localization (LN recurrences mapped to region/station and labeled in-field vs out-of-field). DFS was time from surgery to first recurrence or death; OS from surgery to death. Survival was assessed using Kaplan–Meier/log-rank and Cox models. Results: Among 259 patients (median follow-up 26 months), 85 (33.0%) developed recurrence; 79 were localizable. Local recurrence occurred in 45/79 (57.0%) and distant metastasis in 45/79 (57.0%), including 11/79 (13.9%) with concomitant local and distant events. Out-of-field relapse occurred in 51/79 (64.6%) and in-field relapse in 37/79 (46.8%); 9/79 (11.4%) had both. Poor responders had higher recurrence (42.7% vs 17.8%, P = 0.0002) and mortality (25.6% vs 11.0%, P = 0.0042) than good responders. TLN (continuous) was not associated with DFS or OS. Each additional PLN increased the risk of DFS events (HR 1.05, 95% CI 1.01–1.08; P = 0.0082) and death (HR 1.05, 95% CI 1.01–1.09; P = 0.0132). Exploratory station-level analyses suggested heterogeneous signals (e.g., station 16 protective; 2R/2L, 8M/8Lo, and station 20 unfavorable), likely confounded by tumor burden, location, and surgical strategy. Conclusions: After nICT and R0 esophagectomy for ESCC, relapse was predominantly out-of-field. TLN did not predict outcomes, whereas TRG and nodal burden—and hypothesis-generating station patterns—better stratified risk, supporting station-informed, response-adapted postoperative surveillance and adjuvant decision-making.