Abstract
<title>Abstract</title> <p> <bold>Objective:</bold> Identifying which patients with advanced esophageal squamous cell carcinoma (ESCC) benefit from immunotherapy maintenance following first-line chemoimmunotherapy and radiotherapy remains a critical challenge in precision oncology. This study leverages artificial intelligence (AI)-driven deep learning to establish a predictive model for optimizing immunotherapy maintenance strategies in advanced ESCC. <bold>Methods:</bold> This multicenter study collected CT images and clinical data from 362 patients with advanced ESCC across three centers. An AI-powered segmentation framework (VISTA3D) was employed for automated tumor delineation, integrating point-prompt enhanced segmentation with deep learning-based feature extraction. A ResNet18-based multiple instance learning (MIL) model combining CT habitat features and clinical variables was developed to predict overall survival and identify patients likely to benefit from immunotherapy maintenance. <bold>Results:</bold> The study included a total of 362 patients. When performing automatic segmentation based on the VISTA3D segmentation model, the model achieved a Dice coefficient of 0.87 on the external validation cohort, with a C-index reaching 0.86 (95% CI: 0.73–0.94). In analyses of PFS and OS, patients classified as low risk by the model demonstrated significantly better outcomes than those classified as high risk (HR: 0.61, 95% CI: 0.46–0.79, P<0.001; HR: 0.55, 95% CI: 0.39–0.77, P<0.001). Among high-risk patients, those receiving immunotherapy maintenance demonstrated significantly improved PFS and OS (HR: 0.39, 95% CI: 0.26–0.75, P<0.001; HR: 0.57, 95% CI: 0.31–0.89, P=0.023). <bold>Conclusion:</bold> This AI-driven deep learning model, based on automated segmentation, effectively identifies advanced ESCC patients who are most likely to benefit from immunotherapy maintenance, enabling precision treatment optimization. The model achieves robust predictive performance, offering a clinically translatable tool for AI-guided immunotherapy decision-making in precision oncology. </p>