Toward safer Artificial Heart Implants: R&D on sensor-enhanced blood interfaces by AI-assisted Artificial Patient model - safeHEARTassist

Project summary

Rationale / Needs to be addressed: Short-term ventricular assist devices (“artificial hearts”) and ventricular plugs to close access holes through the heart muscle after patient’s recovery lack localized, real-time monitoring of the blood-device interface. Thrombus, hemolysis, infection and failed endothelialization often originate at these surfaces. Objectives: (i) Develop flexible screen-printed piezoelectric sensor arrays (PyzoFlex® based) for spatiotemporal pressure mapping. (ii) Integrate biomimetic riblets and antifouling biocoatings. (iii) Train explainable AI models to detect early clotting, pseudointima formation and flow anomalies using Artificial Patient models (TRL 4). Potential applications: Real-time telemetry for pulsatile VADs, monitored ventricular plugs. Impact and potential benefits: Enables early complication detection, fewer revision surgeries, shorter hospital stays, improved explantation outcomes and accelerated recovery by safer, monitored implants.