Validating the DeepComp AI Model to Predict Major Postoperative Complications After Gastric Cancer Surgery

What this Trial Is About

Gastric cancer is a leading cause of cancer-related deaths, with radical surgery as the main treatment. Postoperative complications are common and can greatly affect recovery and quality of life. Doctors currently lack precise tools to predict which patients are at high risk of severe complications before surgery. This research aims to validate an artificial intelligence model called DeepComp, which combines clinical data and advanced features from routine preoperative CT scans to analyze tumor and body composition, including muscle and fat, to assess patient physiological reserve. In this prospective, multicenter observational study, patients scheduled for radical gastrectomy open, laparoscopic, or robotic will undergo standard preoperative contrast-enhanced CT scans. The DeepComp AI model will be applied to these scans to predict the risk of moderate-to-severe postoperative complications. Researchers will then compare these predictions with actual clinical outcomes observed 30 days after surgery to evaluate the models accuracy and reliability. Participants will be followed for 30 days after surgery to monitor postoperative complications using the Clavien-Dindo grading system, focusing on grade II or higher events. Researchers will assess the AI models diagnostic performance and agreement with clinical observations. Additional measures include length of hospital stay up to 30 days after surgery. This study will help determine if DeepComp can be a useful tool for personalized surgical risk assessment in gastric cancer patients.

DeepComp for Prediction of Gastric Cancer Postoperative Complications (DeepComp-Prospective)