AI may expand diagnostic capabilities in cardiology
With the European Society of Cardiology (ESC) 2026 congress behind us, it’s clear that AI in cardiology has reached an inflection point. Across three pillars, electrocardiography (ECG), cardiac imaging, and clinical risk prediction, AI tools are moving from experimental use into everyday clinical workflows.
AI-ECG leads the way
The 12-lead ECG has been a clinical staple for over eighty years but has traditionally been confined for use as a front-line diagnostic tool. This is because it lacks the sensitivity to detect underlying structural changes in cardiac tissue. AI-supported ECG (AI-ECG) is changing that, using deep learning to detect signal patterns invisible to the human eye, including systolic and diastolic dysfunction, hypertrophic cardiomyopathy (HCM), valve abnormalities, coronary disease, and pulmonary hypertension.
Viz.ai’s Viz HCM module illustrates this transition. It received De Novo authorization from the FDA in August 2023 for HCM detection, backed by a multi-year agreement with Bristol Myers Squibb. Since then, its access has continued to expand, including being assigned a CPT (current procedural terminology) code alongside a set national payment rate for Medicare reimbursement. In the long-term, the field is moving towards fewer-lead inputs and expansion into wearables and at-home monitoring.
Cardiac imaging follows closely behind
AI-enhanced imaging, such as cardiac MRI, is a leading use of AI in clinical practice. In September 2025, the FDA cleared an update for Tempus Pixel, enabling AI-automated T1 and T2 in-line map generation from raw MRI data, saving clinicians time by removing the need for manual processing and quickly providing quantitative data for detecting cardiac tissue abnormalities. In March 2026, Philips’ SmartHeart received FDA clearance, which can speed up MRI scans while reducing the number of required patient breath-holds, improving both image quality and patient comfort. With both products commercially launched, the next focus is clinical adoption. In line with this, Philips recently sponsored educational sessions at ESC 2026 on how AI can streamline and accelerate MRI workflows.
Clinical risk prediction has long-term potential
AI-based risk stratification is expected to transform clinical care by identifying cardiovascular (CV) disease earlier in patients, resulting in more effective therapeutic interventions. At the ESC 2026 Paul Hugenholtz Lecture in Innovation, Dr. Charalambos Antoniades (University of Oxford, UK) highlighted the potential of AI in clinical decision making. Dr. Antoniades discussed the development of an AI-Risk algorithm that can predict ten-year CV mortality and major adverse cardiac events (MACE) from routine coronary CT angiographies (CCTA). This first-in-class technology, called CaRi-Heart (Caristo Diagnostics) recently received FDA De Novo approval, has been assigned a CPT code, and is expected to launch in Q4 2026 with the potential to transform clinical CVD risk assessment.
Real-world evidence from the UK’s NHS demonstrates this impact, with clinicians changing management decisions in about half of all patients when presented with AI-Risk results. Sharing this data with patients also made them more likely to adhere to lifestyle and medication changes, impacting patient care beyond the clinic. These initial results were reported in 2024, and so while the clinical impact is already valid, it is also likely to grow even further as these tools improve and adoption matures.
It is clearer than ever that the intersection of AI and cardiology is transitioning from promise to practice. Collectively, these tools are well-positioned to help drive the field towards more precise, personalized cardiovascular care and we’re excited to follow where the field goes next.
