ABSTRACT
Background: Coronary computed tomography angiography (CCTA) has emerged as a reliable noninvasive modality to assess coronary artery stenosis and high-risk plaque (HRP). However, CCTA assessment of stenosis and HRP is time-consuming and requires specialized training, limiting its clinical translation. Objectives: The aim of this study is to develop and validate a fully automated deep learning system capable of characterizing stenosis severity and HRP on CCTA. Methods: A deep learning system was trained to assess stenosis and HRP on CCTA scans from 570 patients in multiple centers. Stenosis severity was categorized as >0%, 1 to 49%, ≥50%, and ≥70%. HRP was defined as low attenuation plaque (≤30 HU), positive remodeling (≥10% diameter), and spotty calcification (<3 mm). The model was then tested on 769 patients (3,012 vessels) for stenosis severity and 45 patients (325 vessels) for HRP. Results: Our deep learning system achieved 93.5% per-vessel agreement within 1 Coronary Artery Disease-Reporting and Data System (CAD-RADS) category for stenosis. Diagnostic performance for per-vessel stenosis was very good for sensitivity, specificity, positive predictive value, negative predictive value, and area under the curve with: >0% stenosis: 90.6%, 88.8%, 83.4%, 93.9%, 89.7%, respectively; ≥50% stenosis: 87.1%, 92.3%, 60.9%, 98.1%, 89.7%, respectively. Similarly, the per-vessel HRP feature achieved very good diagnostic performance with an area under the curve of 0.80, 0.79, and 0.77 for low attenuation plaque, spotty calcification, and positive remodeling, respectively. Conclusions: A fully automated unsupervised deep learning system can rapidly evaluate stenosis severity and characterize HRP with very good diagnostic performance on CCTA.
ABSTRACT
OBJECTIVES: Coronary artery calcium (CAC) scores derived from computed tomography (CT) scans are used for cardiovascular risk stratification. Artificial intelligence (AI) can assist in CAC quantification and potentially reduce the time required for human analysis. This study aimed to develop and evaluate a fully automated model that identifies and quantifies CAC. METHODS: Fully convolutional neural networks for automated CAC scoring were developed and trained on 2439 cardiac CT scans and validated using 771 scans. The model was tested on an independent set of 1849 cardiac CT scans. Agatston CAC scores were further categorised into five risk categories (0, 1-10, 11-100, 101-400, and > 400). Automated scores were compared to the manual reference standard (level 3 expert readers). RESULTS: Of 1849 scans used for model testing (mean age 55.7 ± 10.5 years, 49% males), the automated model detected the presence of CAC in 867 (47%) scans compared with 815 (44%) by human readers (p = 0.09). CAC scores from the model correlated very strongly with the manual score (Spearman's r = 0.90, 95% confidence interval [CI] 0.89-0.91, p < 0.001 and intraclass correlation coefficient = 0.98, 95% CI 0.98-0.99, p < 0.001). The model classified 1646 (89%) into the same risk category as human observers. The Bland-Altman analysis demonstrated little difference (1.69, 95% limits of agreement: -41.22, 44.60) and there was almost excellent agreement (Cohen's κ = 0.90, 95% CI 0.88-0.91, p < 0.001). Model analysis time was 13.1 ± 3.2 s/scan. CONCLUSIONS: This artificial intelligence-based fully automated CAC scoring model shows high accuracy and low analysis times. Its potential to optimise clinical workflow efficiency and patient outcomes requires evaluation. KEY POINTS: ⢠Coronary artery calcium (CAC) scores are traditionally assessed using cardiac computed tomography and require manual input by human operators to identify calcified lesions. ⢠A novel artificial intelligence (AI)-based model for fully automated CAC scoring was developed and tested on an independent dataset of computed tomography scans, showing very high levels of correlation and agreement with manual measurements as a reference standard. ⢠AI has the potential to assist in the identification and quantification of CAC, thereby reducing the time required for human analysis.