Comparative efficacy and outcomes of different treatments for pathologically high-risk superficial esophageal cancer: a systematic review and network meta-analysis.

BACKGROUND: Endoscopic resection (ER) is increasingly used for superficial esophageal cancer (SEC), but pathological examination may reveal features associated with lymph-node metastasis or recurrence. The optimal management of patients with pathologically high-risk SEC remains uncertain because high-quality randomized evidence is lacking and direct comparisons among the available treatment strategies remain limited. METHODS: We performed a systematic review and frequentist network meta-analysis according to PRISMA (PROSPERO CRD420251164650). PubMed, Embase, Web of Science, and the Cochrane Library were searched from inception to December 17, 2025. Random-effects models were used as the primary analyses for overall survival (OS), recurrence-free survival (RFS), and disease-specific survival (DSS). Study-level subgroup analysis, network meta-regression, and sensitivity analyses were used to examine the robustness and comparability of the evidence. RESULTS: Thirty-four studies were included. The OS, RFS, and DSS networks contained 2,621, 2,173, and 1,274 participants, respectively. Compared with primary esophagectomy (ESO), ER followed by esophagectomy (ER+ESO) had an OS hazard ratio (HR) of 0.562 (95% CI 0.311-1.015; P = 0.056). The corresponding HRs were 0.480 for RFS (95% CI 0.235-0.983; P = 0.045) and 0.206 for DSS (95% CI 0.069-0.619; P = 0.005). No other strategy differed significantly from ESO. Global inconsistency was not detected for OS or DSS, whereas the RFS network showed significant between-design inconsistency. Study-level subgroup and network meta-regression analyses did not identify statistically significant effect modification by pathological-stage composition, T1b proportion, or lymphovascular invasion burden. CONCLUSIONS: ER+ESO was associated with favorable RFS and DSS estimates in selected patients suitable for surgery, with a similar direction of effect for OS. These findings may support ER+ESO as an important post-ER treatment option when surgical tolerance is acceptable. ER followed by chemoradiotherapy may represent a clinically relevant organ-preserving strategy, particularly for patients in whom esophagectomy is not preferred or feasible. SYSTEMATIC REVIEW REGISTRATION: https://www.crd.york.ac.uk/PROSPERO/view/CRD420251164650, identifier CRD420251164650.