MENU

AI Auscultation: Juntendo University’s Heart Function Detection

Jiji Press Japan
Overview
Researchers from Juntendo University and AMI Inc. developed an AI model that detects left ventricular dysfunction using digital heart sounds and ECGs with an AUC of 0.904. The system outperformed the diagnostic accuracy of cardiology specialists in external validation, offering a scalable method for early heart failure screening.
In Depth

Key Findings

An AI-driven auscultation method has demonstrated high precision in detecting left ventricular dysfunction, defined as a left ventricular ejection fraction (LVEF) of 40% or less. In an external validation study involving 964 patients across six facilities, the AI achieved an Area Under the Curve (AUC) of 0.904, a sensitivity of 86.3%, and a negative predictive value of 98.5%. When compared directly against seven cardiology specialists, the AI’s accuracy was 82.0%, significantly higher than the 62.0% accuracy achieved by the specialists’ majority vote.

Technical Details

The system utilizes the “Cardio-EGG” digital stethoscope from AMI Inc. to simultaneously record single-lead ECGs and heart sounds from four standard chest locations for eight seconds each. The AI architecture integrates a Transformer model, designed to capture temporal sequences in audio and electrical data, with a Convolutional Neural Network (CNN) for extracting local waveform features. The model was trained on a dataset of 2,359 individuals (average age 69.7) before being validated on the separate cohort of 964 individuals (average age 68.6). The AI outperformed models using only heart sounds (AUC 0.832), only ECGs (AUC 0.804), or statistical models combining clinical data and BNP blood tests (AUC 0.719–0.848).

Background and Context

Left ventricular dysfunction often lacks obvious symptoms, leading to delayed diagnosis of heart failure. While echocardiography is the gold standard for evaluation, it requires specialized equipment and technical expertise. Traditional auscultation is simple but subjective, with results varying based on the physician’s experience. This research sought to provide an objective, AI-enhanced alternative to identify patients who require urgent echocardiography.

Significance and Outlook

Because digital auscultation is fast and requires no heavy infrastructure, it can be deployed in medical settings where immediate echocardiography is unavailable. While the current study focused on patients already deemed likely to need echocardiography, future research will evaluate the system’s utility in general health screenings and its impact on patient prognosis and clinical efficiency.

Source: https://www.jiji.com/jc/article?k=000000924.000021495&g=prt

Get our weekly technology intelligence — free

Receive an infographic that lets you judge at a glance whether each field’s analysis report is worth reading.

Subscribe Free — Weekly Tech Intelligence

By subscribing, you’ll receive Troy-Technical’s weekly technology intelligence newsletter.

  • Your email and selected fields are used only to deliver the newsletter.
  • We never share your information with third parties.
  • You can unsubscribe anytime via the link in each email.

See our Privacy Policy for details.

Takes about a minute · Unsubscribe anytime

Published by Troy-Technical, an independent site run by one engineer with a career in materials development.
About the author / Contact info@troy-technical.jp
Let's share this post !

Author of this article

TOC