Truace
Subscribe

Heart failure and atrial fibrillation prediction from non-gated chest CT scans using AI-based cardiac chamber volumetry: an AI-CVD study within the multi-ethnic study of atherosclerosis (MESA)

Background: We previously demonstrated that AI-derived cardiac chamber volumetry in ECG-gated CT scans predicts atrial fibrillation (AF) and heart failure (HF). Since most chest CTs are non-gated, and the American Heart Association recently highlighted their value for opportunistic cardiovascular risk assessment using AI-based coronary artery calcium scoring, we extended this concept to evaluate whether AI-derived chamber volumetry from non-gated CTs can reliably predict HF and AF.s. Methods: We compared AI-deri…

By the Truvace desk1 min read1 sourcePeer-reviewedTRV-2026-1834· cite & verify
Heart failure and atrial fibrillation prediction from non-gated chest CT scans using AI-based cardiac chamber volumetry: an AI-CVD study within the multi-ethnic study of atherosclerosis (MESA)
Synopsis of the diseases of the larynx, lungs and heart [electronic resource] : comprising Dr. Edwards' tables on the examination of the chest with alterations and additions by Hall, F. de Havilland (Francis de Havilland), 1847-1929. Public domain

In brief

Researchers tested whether an AI tool that measures heart chamber sizes could work on ordinary chest CT scans that are not synchronized to the heartbeat. They compared these measurements with those from ECG-gated scans in 2053 participants of the Multi-Ethnic Study of Atherosclerosis.

Because most chest CTs are non-gated, the finding suggests that existing scans could be reused to flag people at higher risk of heart failure or atrial fibrillation. The authors frame this cautiously, stating the approach may provide reliable measurement, and the report does not show clinical use or patient outcomes from screening itself.

Main points

  1. Agreement for detecting chamber enlargement was 98.8-99.2% (κ=0.76-0.84) between non-gated and gated scans.
  2. Paired ECG-gated scans taken 2 minutes apart showed 98.6-98.9% agreement (κ=0.72-0.77) as a variability benchmark.
  3. Age- and sex-adjusted C-index for left atrial volume was 0.75 versus 0.74 for non-gated versus gated.
  4. Ventricular volume C-indices were 0.77 for both scan types, and the ventricular ratio was 0.74 for both.

The gain

AI chamber volumetry on non-gated chest CT agreed closely with ECG-gated scans and predicted AF and HF about as well, supporting opportunistic risk screening.

The rundown

The analysis used 2053 MESA Exam 5 participants to compare AI-derived chamber volumetry between non-gated and ECG-gated chest CTs, using absolute differences, intraclass correlation coefficients and Cohen's κ.

To set a benchmark for normal scan-to-scan variation, the team analyzed paired ECG-gated scans acquired 2 minutes apart in 5749 MESA Exam 1 participants.

Predictive performance for incident heart failure and atrial fibrillation was compared through cumulative incidence curves, adjusted C-indices and non-inferiority testing with a margin of 0.05. Quartile-stratified curves showed substantial overlap, and non-inferiority p-values were below 0.001.

Sources

  1. Peer-reviewedAmerican Journal of Preventive Cardiology2026-07-05

The debate