MedPromptX is a pioneering clinical decision support system that combines multimodal large language models with visual grounding to enhance chest X-ray diagnosis when electronic health records are incomplete.

• Integrates visual imagery with available clinical data to provide comprehensive diagnostic support
• Employs few-shot prompting techniques to maximize performance with limited training examples
• Uses visual grounding to highlight relevant areas in X-ray images for explainable AI diagnostics
• Addresses the real-world challenge of incomplete electronic health records in clinical settings

This innovation matters because it bridges the gap between medical imaging and clinical data, potentially improving diagnostic accuracy, reducing radiologist workload, and enabling better patient outcomes in cardiopulmonary care.

MedPromptX: Grounded Multimodal Prompting for Chest X-ray Diagnosis