AI-powered vest supports early detection of heart disease

Cardiovascular diseases remain a leading cause of death worldwide. Timely detection significantly increases treatment success, but geographical distance, limited specialist availability, and frequent in-clinic check-ups often delay diagnosis. A new AI-assisted textile vest is intended to support earlier risk detection and help medical teams make data-driven decisions.

As part of the maia project (Medical Artificial Intelligence Applications – Center for Applied AI in Medicine), Fraunhofer IZM, Charité University Hospital and the Technical University of Berlin collaborated to develop a wearable system that gathers cardiovascular data non-invasively.
The vest integrates more than 110 physiological parameters, collected 1,000 times per second, making it the first system globally to monitor this range of diagnostic values continuously.

A built-in chatbot supplements sensor data by capturing patient-reported symptoms, offering clinicians an expanded overview for assessment.

Signal processing occurs via a mobile edge computing module, which analyses data locally before secure wireless transmission. According to Fraunhofer IZM, the newly developed hybrid analytical framework combines conventional methods for processing sensor signals with cutting-edge machine learning to level up the quality of the resulting data.

Using sensor information, patient input and medical assessment, the AI system generates:

• a likely diagnostic impression for clinical staff
• a personalised cardiovascular risk evaluation for patients

This approach supports telemedical care, particularly for individuals living in rural regions or people with mobility limitations.

The vest uses multiple measurement techniques, including:

• Bioimpedance spectroscopy (BIS)
• Electrocardiography (ECG)
• Seismocardiography (SCG)
• Phonocardiography (PCG)
• Photoplethysmography (PPG)

Future extensions may add echocardiography and additional diagnostic modules. The system can calculate medical indicators such as blood pressure or the presence of edema.

The smart textile electrodes are semidry, biocompatible and gel-free, making them gentle on the skin while ensuring stable multichannel recording. The modular vest structure helps maintain correct electrode positioning across different body shapes.

Looking ahead, researchers aim to integrate the system into thin polyurethane stick-on patches for even greater comfort and long-term wearability.

MEDICA-tradefair.com; Source: Fraunhofer IZM