https://doi.org/10.1007/s00392-025-02625-4
1Medical Imaging Centre of Semmewleis University Budapest, Ungarn; 2Universitätsklinikum Heidelberg Klinik für Innere Med. III, Kardiologie, Angiologie u. Pneumologie Heidelberg, Deutschland; 3Das Radiologische Zentrum Sinsheim, Deutschland; 4MVZ-DRZ GmbH Kardiodiagnostik Heidelberg, Deutschland; 5Universitätsmedizin Mannheim der Universität Heidelberg Institut für Herz-Kreislaufforschung Mannheim, Deutschland; 6Das Radiologische Zentrum Heidelberg, Deutschland
Background and aims
Functional assessment is recommended in case of coronary CT angiography (CCTA) examinations with results of uncertain hemodynamical significance. The aim of this study was to find coronary plaque features predictive of the presence of ischemia on cardiac MRI (CMR) examinations in obstructive vessel territories on coronary CT angiography.
Methods
A database search was carried out retrospectively to include patients with both a CCTA and a CMR examination within a six-month period. Automated coronary plaque quantification was performed by a non-commercially available AI based prototype.
The ischemic myocardial segments and the corresponding vessel territories were paired by experienced physicians based on coronary anatomy observed on CCTA images.
Quantitative plaque features of individual plaques per vessel territory were added together. Uni and multivariate logistic regression were carried out to determine the plaque features predictive of ischemia.
Results
After exclusion, 400 patients with 748 obstructive vessels were included in the final analysis. Higher calcified plaque component volume, minimal luminal area, minimal luminal diameter, percentage area and percentage diameter stenosis was found in ischemic vessels.
Plaque volume, the volume of non-calcified plaque component, plaque length, and vessel volume at the site of the plaque was also found to be significantly different in the univariate logistic regression. A significant difference was found in the presence of plaques with high-risk plaque (HRP) features (Chi2=6,847 P = 0,0089). When creating a multivariate logistic regression analysis in the best model independent predictors of ischemia on CMR were minimal luminal diameter, plaque volume and the presence HRPs.
Conclusion
The most significant predictors of vessel territory-based ischemia turned out to be
minimal luminal diameter, plaque volume and the presence of HRPs. As of conclusion, plaque features have additional value over luminal stenosis severity in ischemia detected by stress CMR.