Truace
Subscribe

An Automated Machine Learning Based Framework for Scan Level Coronary Artery Calcium Risk Stratification

Coronary artery calcium (CAC) scoring plays a crucial role in the early detection and risk stratification of coronary artery disease (CAD). Current techniques for CAC analysis rely on manual annotations by clinicians and large curated training datasets. This study introduces a fully automatic framework for CAC risk stratification that eliminates the need for manual clinical annotations. We introduce a novel pixel-attenuation-based feature extraction pipeline for automatic CAC assessment that leverages pseudo-lab…

By the Truvace desk1 min read1 sourcePeer-reviewedTRV-2026-1847· cite & verify
An Automated Machine Learning Based Framework for Scan Level Coronary Artery Calcium Risk Stratification
"Cardiac CT and Coronary Angiography Imaging in Navi Mumbai" by Goleisureintl is licensed under CC BY 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/.

In brief

Researchers built a fully automatic framework to sort patients into coronary artery calcium risk groups using non-contrast images from coronary CT angiography scans. It extracts pixel-attenuation features and uses pseudo-labelling to find regions of interest, so clinicians do not have to annotate scans by hand.

Early stratification of coronary artery disease matters because current calcium analysis depends on manual annotation and large curated datasets. The authors say the approach is cost-effective and time-efficient and could reduce additional diagnostic procedures, but the evidence comes from a small in-house dataset and one public dataset, so clinical deployment is not yet demonstrated.

Main points

  1. The pixel-attenuation pipeline reached accuracies of 92%, 93%, and 87% across two-, three-, and five-class risk tasks.
  2. F1-scores were 92%, 90%, and 78% for the same three classification tasks.
  3. It was compared against a radiomics-based pipeline and a foundation-model-based pipeline.
  4. Training and testing used an in-house ECG-gated CCTA dataset of 182 patients.

The gain

A pixel-attenuation machine learning pipeline can classify coronary artery calcium risk groups from non-contrast CT scans without manual clinical annotations, which could ease early stratification of coronary artery disease.

The rundown

The study addresses a reliance on manual clinician annotations and large curated training datasets in current CAC analysis. Its pipeline uses pseudo-labelling to identify regions of interest, reducing dependence on expert labels.

The team tested training only on non-contrast data against a combined set of contrast-enhanced and non-contrast scans, while evaluating solely on non-contrast images. Generalizability was checked on the public COCA-Coronary Calcium and Chest CTs dataset, including lower-quality non-gated scans, where performance was described as robust.

Sources

  1. Peer-reviewedJournal of Imaging Informatics in Medicine2026-10-05

The debate