Enhancing Coronary Artery Calcium Scoring via Multi-Organ Segmentation on Non-Contrast Cardiac Computed Tomography

TL;DR

This paper introduces a novel anatomically-informed algorithm for coronary artery calcium scoring on non-contrast cardiac CT scans, achieving higher accuracy and interpretability by leveraging multi-organ segmentation and anatomical insights.

Contribution

The paper presents a new method that improves calcium scoring accuracy and interpretability by integrating multi-organ segmentation and anatomical understanding, surpassing existing state-of-the-art techniques.

Findings

Achieves inter-observer level accuracy on multi-vendor datasets.

Enhances interpretability through anatomical insights.

Effectively labels coronary and aortic calcifications, reducing false positives.

Abstract

Despite coronary artery calcium scoring being considered a largely solved problem within the realm of medical artificial intelligence, this paper argues that significant improvements can still be made. By shifting the focus from pathology detection to a deeper understanding of anatomy, the novel algorithm proposed in the paper both achieves high accuracy in coronary artery calcium scoring and offers enhanced interpretability of the results. This approach not only aids in the precise quantification of calcifications in coronary arteries, but also provides valuable insights into the underlying anatomical structures. Through this anatomically-informed methodology, the paper shows how a nuanced understanding of the heart's anatomy can lead to more accurate and interpretable results in the field of cardiovascular health. We demonstrate the superior accuracy of the proposed method by evaluating it on an open-source multi-vendor dataset, where we obtain results at the inter-observer level, surpassing the current state of the art. Finally, the qualitative analyses show the practical value of the algorithm in such tasks as labeling coronary artery calcifications, identifying aortic calcifications, and filtering out false positive detections due to noise.