HeartLung.AI unveils 10-study ESC program on AI CT screening

HeartLung.AI said it completed a company-record scientific program at ESC Congress 2026 in Munich, presenting 10 studies on how AI can extract cardiovascular, structural heart and lung cancer risk signals from routine CT scans. The work centers on Agatston-2.0, a next-generation calcium scoring approach that identified hidden risk in people with zero conventional CAC scores.

Why it matters: - HeartLung.AI is arguing that one routine CT scan can do more than answer a single clinical question. - The company’s ESC Congress 2026 data suggest AI can turn existing imaging into a broader preventive screen for coronary disease, valvular disease, cardiac remodeling, body composition and lung cancer risk. - The findings could reduce the need for separate scans, while giving clinicians more quantitative information from exams patients are already getting.

What happened: - HeartLung.AI presented 10 studies at ESC Congress 2026 in Munich. - The studies covered cardiovascular prevention, coronary calcium, cardiac chamber analysis, structural heart disease, cardiometabolic phenotyping and opportunistic lung cancer risk assessment from routine CT imaging. - The company said the program builds on its May announcement that 10 abstracts had been accepted to ESC.26. - Morteza Naghavi, HeartLung.AI founder and CEO, said the presentations showed the scientific depth behind AI-CVD and the potential for routine CT scans to become more powerful tools for prevention.

The details: - Agatston-2.0 is HeartLung.AI’s next-generation AI-based coronary calcium quantification framework. - The approach uses deep-learning coronary segmentation, image calibration, noise normalization, voxel-level density weighting and spatial filtering on the same non-contrast CT scan. - In a study of 3,965 MESA and Framingham Heart Study participants with conventional CAC scores of zero, Agatston-2.0 detected AI-CAC above zero in 862 people, or 21.7%. - After adjustment for traditional cardiovascular risk factors, detectable AI-CAC was associated with 73% higher risk of myocardial infarction, 85% higher risk of hard coronary heart disease and 71% higher risk of all coronary heart disease. - Ten-year progression to conventional CAC above zero was 66.3% among people with AI-detected calcium, versus 43.4% among those without it. - When Agatston-2.0 was combined with other AI-CVD biomarkers from the same CT scan, predicted 10-year cardiovascular risk within the conventional CAC-zero group ranged from about 0.3% to 15.7%. - In a second study of 1,542 MESA and Framingham participants with CAC scores from 1 to 99 and a median 12.5 years of follow-up, 33% were up-classified and accounted for 52% of 10-year CHD events. - In that same study, 41% were down-classified to less than 5% 10-year risk, and the approach produced a +34.8% net reclassification. - In the Miami Heart Study, AI-derived epicardial adipose tissue measurements in 1,259 participants found non-calcified plaque burden rose about 93% per standard-deviation increase in epicardial fat among people with BMI below 29. - Low-CAC participants in the highest epicardial-fat quartile had about three times the median non-calcified plaque burden of those in the lowest quartile. - Another Miami Heart Study analysis of 1,259 asymptomatic participants found an AI-CVD model achieved an AUC of 0.957 for obstructive stenosis, compared with 0.881 for Agatston CAC alone. - The same model posted an AUC of 0.789 versus 0.626 for high-risk plaque and 0.771 versus 0.679 for non-calcified plaque. - A MESA study of 1,675 participants compared AI-derived cardiac chamber measurements from non-gated, non-contrast chest CT with echocardiography. - A separate MESA analysis of 2,053 participants found strong agreement between non-gated chest CT and ECG-gated cardiac CT, with ICCs of 0.94 for left atrial volume, 0.95 for left ventricular volume and 0.96 for LV mass. - Those non-gated CT measurements were non-inferior to gated CT measurements for predicting future heart failure and atrial fibrillation. - A MESA study of 5,520 participants followed for about two decades examined 25 AI-CVD-derived imaging phenotypes for future aortic stenosis risk. - Aortic valve calcification was the strongest individual predictor, and the broader AI-CVD model reached a 10-year AUROC of 0.973 versus 0.809 for the clinical model. - Another study used MESA and Framingham data with follow-up up to 17 years to evaluate left atrial volume index and LA/RA and LA/LV ratios from CAC scans for future atrial fibrillation and stroke. - A lung cancer study of 5,726 MESA participants found Sybil AI could identify imaging signals associated with future lung cancer risk using only the baseline CAC scan, without smoking history or other clinical factors. - Predictive performance for lung cancer risk stayed near 70% AUC through much of follow-up and was about 68% at 15 years. - In a head-to-head comparison of 4,486 cardiac CAC and chest CT scans from Framingham and MESA, Sybil kept an AUC above 0.80 for lung cancer risk prediction through six years despite the narrower field of view on cardiac CT. - HeartLung.AI said its FDA-cleared AI-CVD platform can extract quantitative measures including coronary artery calcium, thoracic aortic and valvular calcium, cardiac chamber volumes, left ventricular mass, aortic and pulmonary artery dimensions, epicardial fat, liver attenuation, muscle and visceral fat composition, lung density and bone mineral density. - The company also said the platform requires no new hardware or local software installation and can connect to PACS or accept manual uploads to generate DICOM and PDF reports. - HeartLung.AI included a long list of physicians, scientists, engineers and academic collaborators who contributed to the ESC program. - The company provided more information about the ESC Congress 2026 research and presentations.

Between the lines: - The scientific message is broader than coronary calcium scoring. - HeartLung.AI is positioning AI-CVD as a platform that can surface multiple disease signals from one exam instead of treating each measurement as a separate workflow. - The strongest implication is not that CT replaces other testing, but that routine imaging may carry more preventive value than clinicians typically extract today. - The Agatston-2.0 data also narrow the debate around zero-calcium scans by separating truly very-low-risk patients from those with AI-detected sub-threshold disease.

What's next: - HeartLung.AI said it will keep advancing AI-CVD, Agatston-2.0 and broader CT-based prevention research through multi-institutional collaborations, clinical validation and deployment initiatives. - The company’s stated goal is to make analysis of existing imaging more comprehensive, scalable and clinically useful. - HeartLung.AI said it wants to help healthcare organizations move from late-stage disease response toward earlier detection and prevention.

The bottom line: - HeartLung.AI used ESC Congress 2026 to show that routine CT scans may be a much richer preventive dataset than conventional interpretation suggests.