Five-Year Unified Risk Prediction Model for Gastric and Esophageal Cancer Based on Nationwide Health Check-Up Data in...

Area of Science:

• Gastroenterology
• Oncology
• Epidemiology

Background:

• Gastric cancer (GC) and esophageal squamous cell carcinoma (ESCC) are leading causes of cancer death in Japan.
• A unified risk prediction model for both GC and ESCC could improve screening and surveillance strategies.

Purpose of the Study:

• To develop and validate a 5-year risk prediction model for GC and ESCC using data from health check-up esophagogastroduodenoscopy (EGD).
• To create a practical scoring system (J-GESS score) for risk stratification.

Main Methods:

• Retrospective analysis of 33,422 individuals undergoing screening EGD in Japan.
• Cox regression identified predictors; a scoring system was developed.
• Model performance assessed using Brier score, calibration, and c-indices, with internal validation via bootstrap resampling and decision curve analysis (DCA).

Main Results:

• The 5-year cumulative incidence of GC or ESCC was 0.84% (234 events).
• Key predictors included older age, male sex, smoking history, and gastric mucosal atrophy.
• The J-GESS model demonstrated good calibration (c-indices 0.81) and clinical utility via DCA.

Conclusions:

• A unified, practical, and validated 5-year prediction model (J-GESS score) for GC and ESCC was developed using routinely collected data.
• The J-GESS score facilitates effective risk stratification.
• This model can inform optimized endoscopic surveillance and screening intervals in Japan.