Our solution is DeepPHDet, an innovative non-invasive technology for heart sound analysis. DeepPHDet is an AI-powered algorithm that analysis heart sounds captured with digital stethoscopes. The innovative method is based on the S2 heart sounds and significantly improves the detection accuracy and explainability of the diagnosis. The collected audio data is transformed into 2D feature maps, enabling real-time or offline predictive analysis directly from heart sound recordings. DeepPHDet seamlessly integrates with digital stethoscopes and healthcare platforms, providing a non-invasive and remote solution for the early detection of Pulmonary Hypertension.
This technology represents a significant advancement in automated heart sound analysis, broadening access to data-driven diagnosis and effective care.
Pulmonary hypertension (PH) is a chronic disease affecting both the lungs and heart and with a 1% global prevalence. PH is often difficult to detect early, as it is frequently missed during routine physical exams. Current diagnostic methods, including Right Heart Catheterization and Transthoracic Echocardiography, are either invasive or require specialized teams. New monitoring technologies like CardioMEMS are costly and limited in application. The delayed diagnosis leads to severe complications and increased healthcare costs for patients, caregivers, and health systems.
These challenges in diagnosing PH, along with the high costs associated with invasive procedures, make them inaccessible in many settings and limit screening programs. There is an opportunity to develop new, non-invasive diagnostic tools that can be integrated into echocardiograms and regular check-ups. This would address the gaps in early detection and make screening more accessible and cost-effective.
- Non-invasive: no needles or surgeries, uses heart sounds collect through a digital stethoscope;
- High predictive performance: reliable tool for data-driven diagnosis in less than 5 minutes;
- Versatile: easy to integrate with digital stethoscopes with AI-supported analysis;
- Universal: suitable for any patient, including patients at home or via telemedicine.
Clinical diagnosis; regular monitoring of patients; individual data monitoring.
