HeartLung.AI to Present 10 Studies at ESC Congress 2026

HeartLung.AI will present 10 scientific studies at ESC Congress 2026 in Munich from Aug. 28-31, marking the company’s largest showing at a major cardiology meeting. The research highlights how AI may extract more preventive value from routine CT scans, including risk prediction for heart disease, stroke, aortic stenosis and lung cancer.

Why it matters: - HeartLung.AI is using routine CT scans to look for disease signals beyond the original reason for the scan. - The company’s ESC Congress 2026 program shows how AI could expand preventive screening without requiring a new imaging test. - The research spans cardiovascular disease, valvular disease, stroke risk and lung cancer risk.

What happened: - HeartLung.AI said it will present 10 scientific studies at ESC Congress 2026 in Munich, Germany, from Aug. 28-31. - The company called the program its largest scientific showing at a single major cardiology congress. - The 10 accepted studies cover coronary artery calcium scoring, cardiac chamber volumetry, heart failure, atrial fibrillation, stroke, aortic stenosis, epicardial adipose tissue, cardiometabolic phenotyping, coronary plaque characterization and lung cancer risk prediction. - The company said it expects to be the only company at ESC Congress 2026 presenting this many accepted scientific studies.

The details: - One MESA study of 1,675 participants compared AI-derived cardiac chamber measurements from non-gated, non-contrast chest CT with echocardiography for heart failure and atrial fibrillation assessment. - A second MESA study followed 5,520 participants for about two decades and tested 25 AI-CVD imaging phenotypes for future aortic stenosis risk. - In that study, aortic valve calcification was the strongest individual predictor, and the AI-CVD model reached a 10-year AUROC of 0.973 versus 0.809 for the clinical model. - A third MESA study of 2,053 participants found high agreement between non-gated chest CT chamber measurements and ECG-gated cardiac CT, including ICCs of 0.94 for left atrial volume, 0.95 for left ventricular volume and 0.96 for LV mass. - The same study found non-gated CT measurements were non-inferior to gated CT measurements for predicting future heart failure and atrial fibrillation. - In the Miami Heart Study, researchers analyzed 1,259 participants with paired CAC and CCTA scans to test epicardial adipose tissue analysis. - Among participants with BMI below 29, non-calcified plaque burden rose about 93% per standard-deviation increase in epicardial fat. - In low-CAC participants, those in the highest epicardial-fat quartile had about three times the median non-calcified plaque burden of those in the lowest quartile. - Another study examined 1,542 MESA and Framingham Heart Study participants with a median 12.5 years of follow-up and introduced Agatston-2.0. - Within the CAC 1–99 range, 33% of participants were up-classified and accounted for 52% of 10-year CHD events, while 41% were down-classified to less than 5% 10-year risk. - The Agatston-2.0 approach produced a +34.8% net reclassification using information from the same CAC scan. - A Miami Heart Study analysis of 1,259 asymptomatic participants found AI-CVD phenotyping outperformed Agatston CAC alone for detecting obstructive stenosis, high-risk plaque and non-calcified plaque on coronary CT angiography. - The AI-CVD model achieved an AUC of 0.957 for obstructive stenosis, compared with 0.881 for Agatston CAC alone. - For high-risk plaque, the model posted 0.789 versus 0.626, and for non-calcified plaque 0.771 versus 0.679. - A study of 3,965 MESA and Framingham participants with conventional CAC scores of zero found AI-derived coronary calcium in 21.7% of them. - Participants with AI-CAC above zero had higher long-term CHD risk and more progression to conventional CAC positivity. - Another study used baseline CAC scans from 5,726 MESA participants and found Sybil AI could identify imaging signals linked to future lung cancer risk without smoking history or other clinical risk factors. - Predictive performance stayed near 70% AUC through much of follow-up and was about 68% at 15 years. - A separate analysis used MESA and Framingham data with up to 17 years of follow-up to evaluate left atrial volume index and LA/RA and LA/LV ratios from routine CAC scans for future atrial fibrillation and stroke. - Investigators also compared Sybil AI on 4,486 cardiac CAC and chest CT scans from the Framingham Heart Study and MESA. - Sybil maintained an AUC above 0.80 for lung cancer risk prediction through six years despite the narrower lung field of view on cardiac CT. - HeartLung.AI said the program reflects collaboration across cardiovascular medicine, radiology, preventive cardiology, imaging science and artificial intelligence. - The company’s collaborators include researchers from multiple institutions and specialties listed in the announcement. - Morteza Naghavi, HeartLung.AI founder and president, said the breadth of the science shows how a single CT scan can potentially provide insights beyond its original indication. - Naghavi also said the company aims to use information already contained in medical images to identify disease earlier. - HeartLung.AI said full presentations are available on the ESC 2026 section of its Slide Presentations page: the ESC 2026 slide presentations page - HeartLung.AI’s FDA-cleared AI-CVD platform transforms eligible CT scans into preventive health assessments by quantifying coronary artery calcium, aortic and valvular calcification, cardiac chamber size, aorta and pulmonary artery size, epicardial and visceral fat, liver density, lung density, bone mineral density and muscle-fat composition. - The company said its platform requires no new hardware or local software installation. - Hospitals and imaging centers can connect PACS to HeartLung.AI’s cloud or manually upload scans to receive AI-generated DICOM and PDF reports.

Between the lines: - The announcement frames CT scans as a broader prevention tool rather than a single-purpose test. - The research emphasis suggests HeartLung.AI is trying to turn existing imaging into a multi-disease risk engine. - The repeated use of MESA and Framingham data points to a strategy focused on large, well-known cohorts that can support long-term risk prediction claims.

What's next: - HeartLung.AI’s findings will be presented at ESC Congress 2026 in Munich from Aug. 28-31. - The company is positioning the conference as a key venue to show whether AI-based scan analysis can support earlier preventive care at scale. - Further validation and clinical adoption will depend on how these models perform outside the study cohorts and in routine care.

The bottom line: - HeartLung.AI is betting that the same CT scan can reveal far more than one diagnosis, and ESC Congress 2026 is its biggest stage yet for that pitch.