Revolutionizing ECG interpretation with AI-powered universal smartphone technology

Overview

This Powerful Medical abstract describes PMcardio, a smartphone application that uses image recognition and artificial intelligence (AI) to standardize and digitize photographs of 12-lead ECGs. Once an ECG is digitized, the app interprets 38 different cardiovascular diseases and provides guideline-compliant recommendations for patient referral and further diagnostic and treatment options. The authors present it as a response to a common gap: millions of ECGs are recorded every year, often without an expert cardiologist available to interpret them, and simple computerized interpretation algorithms built into ECG devices often give inadequate interpretations. The abstract is a technology description and does not report accuracy or outcome data.

Published in: ISCE 2023 (International Society for Computerized Electrocardiology); abstract published in the Journal of Electrocardiology, Volume 78
Published on: 30 March 2023

Background

Worldwide, millions of ECGs are recorded every year, yet in many cases an expert cardiologist with sufficient training to evaluate the recording competently is not readily available, causing frequent misdiagnoses and discrepancies in patient management. Computerized interpretation of electrocardiograms was introduced to address this, but despite improvements over recent years, the simple algorithms built into ECG devices often provide inadequate interpretations. The main obstacles to better diagnostic performance are error-proneness with frequently encountered noisy waveform data, limited access to physician-validated 12-lead ECG data in digital format, the absence of standardized diagnostic criteria, and inadequate incorporation of clinical context.

Methods

The authors present PMcardio, a versatile smartphone application that uses image recognition and AI to standardize and digitize images of 12-lead ECGs. After digitization, the application interprets 38 different cardiovascular diseases and provides guideline-compliant recommendations for patient referral and for further diagnostic and treatment options.

Results

This abstract is a technology description and does not report a study cohort, diagnostic accuracy or outcome data. Its figure illustrates the app workflow in four steps: ECG digitization from a smartphone photograph, interpretation and diagnostics, referral recommendations, and sharing and collaboration between clinicians.

Conclusion

PMcardio combines smartphone-based digitization of 12-lead ECGs with AI interpretation of 38 cardiovascular diseases and guideline-compliant referral and treatment recommendations, aiming to address the misdiagnoses and inconsistent patient management that arise when an expert cardiologist is not readily available.