Prediction of high coronary artery calcium (CAC) scores from aortic arch calcification: An efficient tool for selection of non-optimal candidates for coronary CTA?

Purpose: Overutilization of healthcare resources is causing a high socioeconomic burden. Patients with high coronary artery calcium (CAC) scores > 1000AU are not optimal candidates for coronary CTA and better suited for other diagnostic strategies. Therefore, our objective was to evaluate whether a 4-scale aortic arch calcification severity (AoArCa) score from CT and X-Ray predicts high-CAC scores.
Methods: Patients referred to coronary/aortic CT-Angiography were enrolled. The severity of aortic arch calcification (AoArCa) was scored as grade: 0 = absent, 1 = minimal (<25 % of circumference), 2 = mild (25-50 %), 3 = moderate (50-75 %) and 4 = severe (75-100 %) on both thoracic CT and X-ray.
Results: In 130 patients, the absence of AoArCa by CT was highly accurate to rule out CAC > 1000AU (sens. 100 %). No or minimal AoArCa had a high NPV of 95.6 % to rule out CAC > 1000 and grade 0,1 + 2 a NPV of 86.96 %. The AUC of AoArCa by CT for predicting high CAC > 1000 was c = 0.84 (p < 0.001; 95 %CI: 0.771--0.91). For moderate-to-severe AoArCa, accuracy was c = 0.792 (p < 0.001). The intermodality agreement between CT and X-Ray based AoArCa Scores was good (r = 0.824, p < 0.001); ICC = 0.902. For X-ray, AUC was c = 0.715 to predict CAC > 1000 (p < 0.001). In regression models, only moderate-or-severe AoArCa, but not the other CVRF predicted CAC > 1000 (p < 0.001), and there was an association of the number of CVRF.
Conclusions: Patients with moderate-to-severe aortic arch calcification have a high probability of CAC > 1000AU, but not those with no, minimal and mild. The absence of AoArCa rules out CAC > 1000AU. AoArCa severity may serve as valuable tool for selecting the diagnostic strategy.
Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
(Copyright © 2023 The Author(s). Published by Elsevier B.V. All rights reserved.)