Fig. 2

Overview of the prediction models. A. Model utilizing only radiographic image data as input. Image features are extracted through an image feature extractor and fed into a classifier to generate predictions. B. Model integrating both radiographic images and clinical data (e.g., age and BMI). Clinical features were extracted using a clinical feature extractor and concatenated with image features before being passed to the classifier. Integrating image and clinical data leverages complementary information, potentially improving predictive performance