FAMISH score for predicting metachronous gastric lesions after endoscopic resection: validation in Japanese patients

Endoscopic submucosal dissection (ESD) is a curative treatment for early gastric cancer (EGC); however, the risk of metachronous gastric lesions (MGL) persists. The FAMISH score was recently developed in Europe to stratify MGL risk, but its external validity in Japanese patients remains unclear. We aimed to externally validate the FAMISH score in a Japanese cohort undergoing endoscopic resection.

We retrospectively analyzed patients who underwent endoscopic resection between 2005 and 2024. The FAMISH score (0-9), based on six clinical and endoscopic variables, was calculated for each patient. Predictive performance for MGL was assessed using receiver operating characteristic (ROC) analysis and cumulative incidence analyses with and without competing risks.

A total of 387 patients were included (non-MGL: 320; MGL: 67). The MGL group had significantly higher FAMISH scores, both in the proportion with scores ≥3 (82.1% vs. 65.6%, p = 0.01) and in median score (5.0 vs. 3.0, p < 0.01). ROC analysis showed modest discrimination (AUC 0.657, 95% CI 0.586-0.728). A cutoff ≥3 yielded 82.1% sensitivity and 34.4% specificity. Competing-risk analysis demonstrated significantly higher cumulative incidence in the high-score group (Gray's test, p = 0.007), and Kaplan-Meier analysis showed similar results (log-rank p = 0.010).

The FAMISH score was significantly associated with MGL development after endoscopic resection in Japanese patients. Despite its modest discriminatory performance, the FAMISH score may have potential to provide additional information for risk stratification; however, its clinical utility in guiding post-ER surveillance remains to be established.