Multimodal Digital Twin for Personalized Heart Disease Prediction

CardioTwin: A Digital Twin Framework for Predicting Heart Disease Progression is a multimodal healthcare AI system designed to model patient-specific cardiac function and forecast disease progression using Electronic Health Records (EHR), ECG time-series signals, and Cardiac MRI data. The framework integrates Cardiac-FM, a multimodal foundation model, to construct a personalized digital twin of the heart by learning shared representations across clinical, electrical, and structural modalities. The system combines machine learning, deep learning, and rule-guided scoring mechanisms to generate interpretable outputs, including cardiac risk prediction, disease progression estimation, modality-wise contribution analysis, and disease stage assessment. Built using Python, PyTorch, FastAPI, and React, CardioTwin enables doctor-facing visualization, what-if simulation, and personalized clinical decision support through a digital twin dashboard for cardiovascular risk monitoring.