Work

Risk factors and prediction model for submucosal invasion in early protruding-type esophagogastric junction adenocarcinoma: conventional endoscopy combines with endoscopic ultrasonography

Year: 2025

Type: article

Abstract: BACKGROUND: Precise depth evaluation of protruding-type early esophagogastric junction adenocarcinoma (EGJAC) is critical for therapeutic planning. This study evaluated the diagnostic value of conventional endoscopy (CE) and endoscopic ultrasonography (EUS) for identifying suspected submucosal-invasive carcinomas and established a predictive model for SM2 invasion. METHODS: A retrospective analysis included 111 patients with protruding-type EGJAC (2015-2024). CE macroscopic morphology, EUS diagnostic accuracy, and clinicopathological correlations were systematically assessed. multivariate analysis identified SM2 predictors. Construct a predictive nomogram for SM2-invading protruding-type early EGJAC incorporating CE macroscopic morphology, air-induced morphological changes, gastroesophageal reflux disease (GERD), EUS-verified, and submucosal cystic echoes. The predictive accuracy of the nomogram was evaluated through receiver operating characteristic (ROC) curve analysis and calibration validation. RESULTS: CE and EUS showed comparable overall diagnostic accuracy (65.7% vs. 69.3%). A numerical improvement in diagnostic yield was observed with EUS for CE-SM2 low-confidence lesions (41.7% vs. 24.0%). CE-SM2 low-confidence lesions demonstrated both significantly elevated GERD co-occurrence rates (56.0% vs. 13.8%; P = 0.038) and predominant abrupt elevation without depression morphology (AEw/oD: 64.0%), contrasting with abrupt elevation with depression morphology (AEwD: 71.4%) in high-confidence counterparts (P < 0.001). Multivariate analysis identified AEwD (OR = 8.57, 95% CI: 1.29-56.81) and absence of submucosal cystic echoes (OR = 6.97, 95% CI: 1.84-26.37) as independent SM2 predictors. The nomogram model for predicting SM2 invasion showed an area under the receiver operating characteristic (ROC) curve of 0.843 (95% CI: 0.732-0.954). CONCLUSIONS: A validated predictive model has been developed to identify SM2 invasion in protruding-type early EGJAC. This predictive tool additionally facilitates optimal treatment selection guidance for patients with protruding-type early EGJAC. In addition, accurate assessment of CE-SM2 low-confidence invasion in protruding-type early EGJAC requires integrated evaluation incorporating CE, EUS, and GERD parameters.

Source: BMC Gastroenterology

DOI: 10.1186/s12876-025-04501-x