A Framework for Cross-Domain Generalization in Coronary Artery Calcium Scoring Across Gated and Non-Gated Computed Tomography

2026-02-25

AI Summary

This work proposes CARD-ViT, a novel framework that enables coronary artery calcium (CAC) detection and Agatston scoring directly on routine non-gated chest CT scans—despite being trained exclusively on electrocardiogram-gated CT data. Leveraging a Vision Transformer architecture with DINO self-supervised pretraining, the model achieves cross-domain generalization without requiring annotated non-gated images. Evaluated on an unseen Stanford non-gated dataset, CARD-ViT attains an accuracy of 0.707 (Cohen’s κ = 0.528), approaching the performance of models specifically designed for non-gated scans. On gated test data, it achieves 0.910 accuracy (κ > 0.87), demonstrating robustness and effectively overcoming the domain shift between gated and non-gated imaging protocols. This represents a significant advance in deploying CAC scoring in broader clinical settings where only non-gated CT is available.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

Coronary artery calcium (CAC) scoring is a key predictor of cardiovascular risk, but it relies on ECG-gated CT scans, restricting its use to specialized cardiac imaging settings. We introduce an automated framework for CAC detection and lesion-specific Agatston scoring that operates across both gated and non-gated CT scans. At its core is CARD-ViT, a self-supervised Vision Transformer trained exclusively on gated CT data using DINO. Without any non-gated training data, our framework achieves 0.707 accuracy and a Cohen's kappa of 0.528 on the Stanford non-gated dataset, matching models trained directly on non-gated scans. On gated test sets, the framework achieves 0.910 accuracy with Cohen's kappa scores of 0.871 and 0.874 across independent datasets, demonstrating robust risk stratification. These results demonstrate the feasibility of cross-domain CAC scoring from gated to non-gated domains, supporting scalable cardiovascular screening in routine chest imaging without additional scans or annotations.

Problem

• Coronary Artery Calcium Scoring
• Cross-Domain Generalization
• Gated CT
• Non-Gated CT
• Cardiovascular Risk Prediction

Innovation

• cross-domain generalization
• self-supervised learning
• Vision Transformer
• coronary artery calcium scoring
• non-gated CT