Development of a Deep Learning Model to Predict Complete Response in Esophageal Squamous Cell Carcinoma After Neoadjuvant Immunochemotherapy

What this Trial Is About

Researchers are evaluating a deep learning model to predict pathological complete response pCR in patients with esophageal squamous cell carcinoma ESCC who have undergone neoadjuvant immunochemotherapy. This study collects clinical, imaging, and pathological data from past patients to develop and validate a model that may help forecast treatment outcomes and guide personalized therapy decisions. The study is retrospective and multicenter, focusing on patients treated between January 2019 and July 2025. The study involves collecting chest CT images and clinical data from ESCC patients who received neoadjuvant immunochemotherapy followed by surgery. Deep learning techniques will extract quantitative features from these CT images to build the predictive model. The models accuracy will be evaluated using metrics such as area under the curve AUC, sensitivity, specificity, positive predictive value, and negative predictive value. An additional analysis called SHapley Additive exPlanations SHAP will assess how CT imaging features contribute to the models predictions. Participants clinical and imaging data before and after neoadjuvant treatment will be analyzed, focusing on CT scans taken before treatment and after completion but before surgery. The main outcome measure is the rate of pathological complete response assessed at surgery within one month post-treatment. The models performance will be validated about a year after the study completion. The study duration extends through late 2026, and no direct treatment is given during the study since it is observational.

Deep Learning Model Predicts Pathological Complete Response of Esophageal Squamous Cell Carcinoma Following Neoadjuvant Immunochemotherapy