top of page
32200c5c-7c4c-438f-a9df-3119a2398bf2.png
Autchmaber1.png
ai-cvd.png
AI-CVD Summary
Coronary Calcification
816_aicvd (8)_4_edited.jpg
‌ Aortic Calcification
ٍEpicardial Fat _edited.jpg
Epicardial Fat
816_aicvd (8)_5_edited.jpg
Cardiac Chambers
816_aicvd (8)_7_edited.jpg
Aorta & PA Size
816_aicvd (8)_10_edited.jpg
Lung LAI
iScreen Shoter - 20261028172009135.jpg
Lung HAI
816_aicvd (8)_11_edited.jpg
Liver Fat
816_aicvd (8)_12_edited.jpg
Muscle & Visceral Fat
816_aicvd (8)_13_edited.jpg
Bone Density
Frame 1000008569.png
Frame 1000008568.png
4acc1a9d-ced8-4f58-a79e-cd2ac6fe752e.png
Frame 1410148109.png
Frame 100000825.png
Frame i1000008267.png
Frame 1410148108.png
Frame 10000082699.png
Fra 1000008266.png
Frame 1000i008268.png
Frame 1410148110.png
Frame 100070.png
Frame 1410148111.png
valve.png

AHA Stroke Alert Podcast Features HeartLung.AI Research on AI-Predicted Atrial Fibrillation and Stroke

Jun 19
3 min read

AI-derived left atrial volume and chamber ratios from CAC scans predicted long-term AF and stroke in MESA and Framingham.


HOUSTON, June 18, 2026 - HeartLung.AI today announced that its research article, "AI-Derived LA Volume Index, LA/RA and LA/LV Volume Ratios From Coronary Artery Calcium Scans Predict Long-Term Atrial Fibrillation and Stroke," was featured on the June 2026 episode of the Stroke Alert Podcast, the podcast of Stroke, a journal of the American Heart Association/American Stroke Association.



The study demonstrates that artificial intelligence applied to routine, noncontrast coronary artery calcium (CAC) scans can extract cardiac chamber measurements that help predict future atrial fibrillation (AF) and stroke. The findings support a broader role for CAC scans beyond the traditional Agatston calcium score, suggesting that the same low-cost CT examination may provide additional prevention-relevant information without contrast or a separate imaging test.


Using pooled participant-level data from the Multi-Ethnic Study of Atherosclerosis (MESA) and the Framingham Heart Study Offspring cohort, investigators analyzed 6,812 participants and followed them for a median of approximately 17 years. During follow-up, 1,302 participants developed AF and 365 experienced stroke events. Participants in the highest 5 percent of AI-derived chamber metrics had significantly higher long-term risk of both outcomes.


Among the key findings, left atrial volume index was the strongest AI-derived predictor of future AF, while the left atrial-to-right atrial ratio emerged as the strongest AI-derived predictor of stroke. Adding these imaging biomarkers improved risk reclassification beyond established clinical risk models, including CHARGE-AF for atrial fibrillation and the Framingham Stroke Risk Profile for stroke.


"Being featured by the Stroke Alert Podcast is an important recognition of this work and of the broader opportunity to use AI to unlock hidden clinical value from scans already being performed," said Morteza Naghavi, MD, founder of HeartLung.AI and senior author of the study. "A CAC scan should not be viewed only as a calcium score. With AI, it can also reveal cardiac chamber remodeling associated with future atrial fibrillation and stroke - two major causes of morbidity, mortality, and health care cost."


The authors note that the findings are intended to support improved risk prediction and hypothesis generation for future clinical studies. Further research is needed to determine how AI-derived chamber measurements from CAC scans should be incorporated into clinical workflows, preventive care pathways, rhythm monitoring strategies, and stroke prevention programs.


The Stroke Alert Podcast segment summarized the article and its key findings for the stroke and vascular neurology community. The podcast feature highlights the growing interest in AI-enabled opportunistic imaging as a strategy to identify patients at elevated risk before clinical events occur.


Key study facts

Article

AI-Derived LA Volume Index, LA/RA and LA/LV Volume Ratios From Coronary Artery Calcium Scans Predict Long-Term Atrial Fibrillation and Stroke

Journal

Stroke

Cohorts

Multi-Ethnic Study of Atherosclerosis and Framingham Heart Study Offspring cohort

Participants

6,812 total: 5,670 from MESA and 1,142 from FHS

Follow-up

Median approximately 17 years

Outcomes

1,302 incident atrial fibrillation events and 365 stroke events

Main finding

AI-derived left atrial volumetric and ratio-based metrics from CAC scans improved long-term AF and stroke risk prediction


HeartLung.AI develops artificial intelligence methods to extract actionable cardiometabolic and thoracic health information from routine CT imaging. Its research initiative focuses on expanding the clinical value of coronary artery calcium and chest CT scans beyond traditional single-purpose reporting.


Useful links


Stroke Alert Podcast episode: June 2026 episode



Comments


Ready to Unlock the Value in Your CT Scans?

Join leading healthcare providers using AI-CVD to save lives and increase revenue, and lead the future of preventive cardiology.

  • Twitter
  • LinkedIn
  • Facebook
  • Instagram
  • Youtube

2450 Holcombe Blvd
TMC Innovations
Houston, TX 77021

aicvd patent_edited_edited.jpg
ISO LOGO_edited.png
autoliver patent.png
iScreen Shoter - 20261104171348520.jpg

©2026 HeartLung Corporation. All Rights Reserved. U.S. Patent Nos. 12,620,494; 12,599,350; 11,610,686. Licensed U.S. Patent Nos. 9,119,590 and 10,695,022. U.S. Patent Applications Pending: 19/390,544; 19/403,750; 19/643,672; 19/728,738; 19/741,766; 18/652,123; 18/771,542; 19/390,629; 19/417,345; 19/467,802; 19/565,385; 19/682,416; 19/690,066. International PCT Applications: PCT/US2025/058510; PCT/US2025/061047; PCT/US2025/058505; PCT/US2026/013566; PCT/US2026/024304. AI-CVD, AI-CAC, AutoBMD, AutoCAC, AutoLiver, AutoChamber, AutoLung and other product names and trademarks displayed on this website are protected by applicable intellectual property laws. All rights reserved.

bottom of page