In progress
Cardio Twin
An edge-based digital twin for real-time cardiovascular monitoring
A digital twin is a computer model of one person's heart that is kept up to date with their own data. Cardio Twin is designed to process physiological signals at the edge, on or near the patient's devices, with privacy and security in mind, and to use the twin to simulate disease progression, virtual stress tests and treatment responses before they reach the patient.
Data in
- Real-time ECG, PPG, blood pressure and heart-rate variability
- Echocardiography, MRI and CT
- Clinical records and genomics
- Activity, sleep, stress and environment
What it is meant to deliver
- Continuous risk monitoring with anomaly alerts
- Disease progression and treatment response simulation
- Patient-specific risk profiles
- Remote care and patient engagement
Overview
The full picture.
From wearable signals processed at the edge to a living model of one patient's heart. Planned framework; no results are reported here.
01Cardiovascular data
- Physiological signals, at the edgeECG, PPG, blood pressure, HRV from wearables and IoMT
- Imagingechocardiography, MRI, CT, angiography
- Clinical data (EHR)history, notes, medications, labs
- Genomics & multi-omicsgenetic risk and pathways
- Lifestyle & environmentactivity, sleep, diet, stress, social factors
02Edge processing
- Standardization
- De-identification
- Noise reduction
- Feature extraction
- Real-time streaming & learning
03Patient-specific twin
What the twin does
- Cardiovascular knowledge graph
- Personalized anatomy & physiology
- Multi-scale modeling
- Disease progression simulation
- Treatment response prediction
- What-if analysis
- Virtual stress testing
- Risk stratification
- Digital biomarkers
Agents
- Monitoring
- Prediction
- Simulation
- Treatment
- Care coordination
04Clinical applications
- Real-time monitoring & early warningcontinuous risk, anomaly alerts
- Personalized treatment planningdrug therapy, lifestyle, adaptive strategies
- Precision risk stratificationpatient-specific risk profiles
- Clinical decision supportevidence-based, interpretable, clinician in the loop
- Patient engagement & remote carefeedback, coaching, self-management
✓Trustworthy, secure & responsible AI
- Privacy & security (federated learning)
- Fairness & bias mitigation
- Explainability
- Robustness & reliability
- Human in the loop
- Regulatory compliance & ethics
- Earlier detection
- Personalized care
- Improved outcomes
- Scalable solutions
- Healthier individuals & communities
Illustrations generated with ChatGPT for the AI4Health Lab. They depict research themes, not study data.