It may be surprising to those unaware of the history of medicine that technology in cardiac care dates to the last decade of the 19th century when the advent of chest X-rays and the Early Electrocardiogram (ECG) gave doctors new insights into the structure and function of the human heart. The 20th century that followed saw many new discoveries and inventions leading to the modern 12-lead ECG that even today is among the first procedures to be carried out when a patient presents with a heart complaint. While doctors still review the ECG patterns visually and interpret them, attempts to automate the process began as early as the late 1950s. Today, advances in software using machine learning and Artificial Intelligence (AI) allow automated ECG interpretation, even detecting patterns that may not be visible even to the trained human eye. AliveCor’s Kardia 12L is the world’s first AI-powered handheld 12-lead ECG system with its latest generation cleared by the US FDA for determining 39 cardiac problems.
What is happening to the ECG is symbolic of the changes that AI is bringing to cardiac care. Wearables are changing the game. In May 2025, the American College of Cardiology (ACC) issued guidelines to help doctors and patients use health data collected while wearing Apple Watch to track and manage cardiovascular health. Smart wearables today come with wellness features that not only help monitor a healthy lifestyle with activity tracking, mindfulness experiences, sleep tracking and cardiorespiratory fitness, but also record the heartbeat and rhythm using an electrical heart sensor, analyse pulse rate data, and identify and notify irregular heart rhythms. These wearables are empowering patients to lead healthy, active lives, of course, in collaboration with their doctors.
These consumer-facing tools have made the shift from the lab to everyday practise based on a growing body of clinical evidence. In 2019, the Apple Heart Study demonstrated that a smartwatch could flag irregular pulses that were later confirmed as atrial fibrillation, a leading and frequently silent cause of stroke. In 2021, the EAGLE trial at the Mayo Clinic found that an AI-enabled ECG helped primary care doctors detect low ejection fraction, a weak heart pump that often goes unnoticed, in patients who would otherwise have been missed. What made these findings striking was that the AI could read signs of disease that experienced cardiologists missed. This ability is what gives the technology the potential to turn a routine recording into a rich source of diagnostic information.
Regulators have been catching up. While radiology leads in the over 1500 AI algorithms cleared by the US FDA, cardiology is second with 225 AI algorithms cleared including 146 under cardiology and 69 for cardiovascular imaging. These cardiac AI models are used for purposes such as improving angiography image quality and guidance in the cath lab, guidance during electrophysiology ablation procedures, automated image assessment, echocardiography assessments, and more.
Globally, AI in cardiac care has huge potential to transform clinical outcomes. More than 80% of deaths due to cardiovascular diseases occur in low and middle-income countries with up to 63% of congenital and chronic heart cases remaining undiagnosed until irreversible complications occur. The survival gap is visible in emergencies where every minute counts. Around 950 of every 1,000 patients in the US who reach an emergency department with a heart attack survive, against about 596 in underserved regions. And it is here that AI can help with swift and accurate diagnosis. A Taiwanese study showed how AI along with ECG testing, slashed the time to diagnose heart attack patients by over ten minutes, reducing infarct size, preserving heart tissue, and reducing the risk of post-myocardial infarction (MI) complications such as heart failure.
AI is also transforming intensive care medicine by enabling advanced analysis of the complex clinical data that is generated in intensive care units (ICUs). Smart ICUs that combine AI with continuous patient monitoring transform intensive care from reactive management to proactive intervention. They analyse vital signs and lab reports, detecting subtle warning signs, to predict complications like sepsis before visible symptoms occur. AI systems help doctors to personalise care by recommending medication dosages and ventilator adjustments for optimal fluid and respiratory balance. They support clinical decision-making and optimise resource utilisation.
The future will bring AI models that fuse ECG, imaging, and electronic-record data for broader screening. But the real success of AI in cardiac care will depend on whether these systems reach patients who need it the most, with the least access to a specialist. While the rapid pace of technological advances makes it well-nigh impossible to predict what the status will be, even a decade from now, we must remember that AI in cardiology is a tool in the hands of clinicians, not a substitute for them.
(The views expressed are personal)
This article is authored by Dr. Ashok Punjabi is a consultant cardiologist with 40 years of non-invase and invasive cardiology practice.
