TLDR: Narayana Health, in collaboration with its Medha AI division, has developed an artificial intelligence model capable of predicting left ventricular ejection fraction (LVEF) from standard ECG images in near real-time. This innovation aims to enable early detection of heart failure and expand diagnostic access, particularly in rural and resource-limited areas of India, where access to traditional echocardiography is often limited. The model, trained on over 100,000 ECG images, demonstrated high accuracy in external validation, identifying 97% of severe heart function reductions an average of 58 days before conventional diagnosis.
Bengaluru, India – Narayana Health’s Clinical Research team, in a significant stride for cardiac diagnostics, has announced the development of a pioneering artificial intelligence (AI) model. Developed in collaboration with its Advanced Analytics and AI division, Medha AI, this breakthrough technology is designed to predict left ventricular ejection fraction (LVEF)—a critical indicator of heart function—directly from standard electrocardiogram (ECG) images in near real-time.
The new AI model addresses a crucial diagnostic gap in India, where heart failure affects nearly 10 million people and accounts for approximately 1.8 million hospital admissions annually. Despite the prevalence, access to echocardiography (ECHO), the current gold standard for assessing heart function, remains uneven, especially in rural and underserved regions. Narayana Health’s AI solution offers a scalable and accessible alternative by leveraging widely available ECG machines and cloud-based analytics.
The model was rigorously trained on an extensive dataset of over 100,000 ECG images, each paired with corresponding ECHO reports. Its efficacy was further validated externally across 14 tertiary centers, encompassing more than 57,000 patients. The results were highly promising: the AI successfully identified 97% of individuals with severely reduced LVEF (≤35%). Crucially, it flagged these patients an average of 58 days before an echocardiogram was eventually ordered, providing a significant lead time for clinical intervention and improved patient outcomes.
Dr. Devi Shetty, Founder and Chairman of Narayana Health, emphasized the broader impact of this innovation, stating, “This is more than a tech breakthrough—it’s a milestone in making heart care universally accessible. Using AI to predict heart failure from a simple ECG empowers clinicians with faster decisions and broader reach. This is India’s moment to lead global healthcare innovation.” He further added, “This breakthrough represents much more than an AI model — it is a reflection of what Narayana Health stands for: combining clinical expertise, deep research, and cutting-edge technology to solve real problems at scale.”
The AI tool has been seamlessly integrated into Narayana Health’s in-house Electronic Medical Record (EMR) system, Athma, ensuring it can be incorporated into existing clinical workflows without disruption. It is currently undergoing real-world validation at key facilities, including the Narayana Institute of Cardiac Sciences, Mazumdar Shaw Medical Center, select clinics in Bengaluru, and the Rabindranath Tagore International Institute of Cardiac Sciences in Kolkata.
Also Read:
- Crafting Digital Heartbeats: How ECGTwin Personalizes Cardiac Data
- ekincare Unveils elva.AI: A Generative AI Assistant Revolutionizing Employee Healthcare Support
Designed for rapid triage and care planning, the model delivers results in near real-time. Future developments are already underway to expand its capabilities to detect other echocardiographic abnormalities and early signs of coronary artery disease. Narayana Health also plans to deploy the model via its mobile medical units, specifically targeting underserved regions as part of its broader digital public health strategy. Regulatory submissions and efforts to enable open-access distribution are also in progress, aiming to make this life-saving technology widely available.
