Breast cancer remains the leading malignancy among women worldwide, while conventional mammography screening is limited by inter-reader variability, radiologist workload, and false-negative diagnoses. This study evaluates a three-tier intelligent framework integrating AI classification, explainable reporting, and robot-assisted biopsy guidance. A hypothetical cohort of 3,542 mammographic cases was constructed from published distributions of three benchmark datasets and evaluated using a 60/20/20 train–validation–test split with stratified 5-fold cross-validation. The framework comprises a CNN–ViT hybrid combining ResNet-50 feature extraction with Vision Transformer attention, a triple-... More >