AI-powered hierarchical classification of ampullary neoplasms: a deep learning approach using white-light and narrow-band imaging.

Key Points

• To improve the diagnostic accuracy of ampullary neoplasms using AI and advanced imaging techniques.
• Developed a deep learning model using hierarchical classification
• Integrated white-light and narrow-band imaging modalities
• Employed synthetic data augmentation to enhance training data
• Significantly improved diagnostic performance for ampullary lesions
• Effectiveness shown across various imaging contexts
• Hierarchical model outperformed traditional classification methods