Indocyanine green-guided sentinel node navigation surgery in early gastric cancer: A systematic review and meta-analysis.

BACKGROUND: For patients with early gastric cancer (EGC) who exceed the absolute indications for endoscopic submucosal dissection (ESD) or have undergone non-curative ESD, sentinel node navigation surgery (SNNS) represents a promising individualized surgical approach. This study aimed to conduct a systematic review and meta-analysis to evaluate the feasibility and diagnostic value of indocyanine green (ICG) guided SNNS in patients with EGC. METHODS: Eligible studies were systematically searched in PubMed, Web of Science, Ovid MEDLINE, Scopus and Cochrane Library from inception to March 2026. Identification rates and patient-level sensitivity were pooled separately with random-effects Freeman-Tukey double-arcsine models in StataNow/MP 19.5. Diagnostic analyses were restricted to studies with an adequate reference standard and reconstructable 2×2 data. Cohort overlap was handled with a conservative primary set and prespecified replacement analyses. Evidence after non-curative ESD was synthesized narratively. RESULTS: Nine non-overlapping cohorts were included in the primary identification analysis, yielding a identification rate of 98.8% (95% CI: 97.6-99.5%). The random-effects pooled identification rate was 99.9% (95% CI = 98.7-100.0%; I2 = 32.0%). Eight studies contributed valid patient-level diagnostic data (84 true positives, 4 false negatives, 0 false positives, and 577 true negatives), and the pooled sensitivity was 98.5% (95% CI 86.6-100.0%; I2 = 47.6%). Specificity was 100.0% (577/577), but estimation was structurally constrained because all studies reported zero false positives. Subgroup associations were exploratory. Evidence after non-curative ESD was heterogeneous and was not pooled. CONCLUSION: ICG-guided sentinel node navigation surgery shows high technical identification and promising patient-level sensitivity in selected EGC cohorts. These findings do not establish oncological non-inferiority of limited surgery. False-negative consequences, zero-cell constraints, overlap, heterogeneous pathology, and concentration of evidence in high-volume Asian centres require cautious interpretation and prospective validation.