Identification of non-calcified coronary plaque characteristics using machine learning radiomic analysis of non-contrast high-resolution CT. (3rd October 2022)

• Record Type: Journal Article
• Authors: Kruk, M; Wardziak, L; Kolossvary, M; Maurovich-Horvat, P; Demkow, M; Kepka, C
• Is Part Of: European heart journal. Volume 43(2022)Supplement 2
• Publication Date: 2022-10-03
• Subjects: Cardiology -- Periodicals; Heart -- Diseases -- Periodicals; 616.12005
• DOI: 10.1093/eurheartj/ehac544.170
• Languages: English
• ISSNs: 0195-668X

Abstract

Abstract: Objective: To explore whether machine learning (ML) radiomic analysis of low-dose, high-resolution, non-contrast, ECG gated cardiac CT scan allows identification of non-calcified coronary plaque characteristics. Background: Novel imaging and analysis techniques may provide the ability to detect non-calcified or high risk coronary plaques on a non-contrast CT scan, advancing cardiovascular diagnostics. Methods: We prospectively enrolled 125 patients with a non-calcified plaque and an adverse plaque characteristic (APC), and 25 controls without visible atherosclerosis on coronary CT angiography (CCTA). All patients underwent the non-contrast CT exam prior to CCTA. 419 radiomic features were calculated to identify: presence of any CAD, obstructive CAD (stenosis>50%), plaque with ≥2 APC, degree of calcification and specific APCs. ML models were trained on a training set (917 segmentations) and tested on a separate (validation) set (292 segmentations). Results: Among the radiomic features 88.3% was associated with any plaque, 0.9% with obstructive CAD and 76.4% with presence of at least two APCs. Overall, 80.2%, 88.5% and 36.5%, of features were associated with calcified, partially calcified, and noncalcified plaques, respectively. Regarding APCs, 61.1%, 61.8%, 84.2%, and 61.3%, of features were associated with low attenuation (LAP), napkin-ring sign (NRS), spotty calcification (SC), and positive remodeling (PR), respectively. ML models outperformed conventional methods for the presence of plaque, obstructive stenosis, presence of 2 APC, as well as for noncalcified plaque and partially calcified plaque, but not for calcified plaque. ML models also significantly outperformed identification of LAP and PR, but neither NRS nor SC. Conclusions: Radiomic analysis of non-contrast CT heart exams may allow identification of specific non-calcified coronary plaque characteristics which could aid cardiovascular risk stratification or pre-screening of individuals prior to contrast enhanced CCTA exam. Funding Acknowledgement: Type of funding sources: Public grant(s) – National budget only. Main funding source(s): National Science Center