Improved long-term prognostic value of coronary CT angiography-derived plaque measures and clinical parameters on adverse cardiac outcome using machine learning

• Published in: European radiology 2021, Vol.31 (1), p.486-493
• Main Authors: Tesche, Christian, Bauer, Maximilian J., Baquet, Moritz, Hedels, Benedikt, Straube, Florian, Hartl, Stefan, Gray, Hunter N., Jochheim, David, Aschauer, Theresia, Rogowski, Sebastian, Schoepf, U. Joseph, Massberg, Steffen, Hoffmann, Ellen, Ebersberger, Ullrich
• Format: Article
• Language: English
• Subjects: Algorithms; Angiography; Calcium; Cardiac; Cardiovascular disease; Cardiovascular diseases; Computed Tomography Angiography; Coronary Angiography; Coronary artery; Coronary artery disease; Coronary Artery Disease - diagnostic imaging; Coronary Stenosis - diagnostic imaging; Decision trees; Diagnostic Radiology; Diagnostic systems; Female; Health risks; Heart; Heart diseases; Humans; Imaging; Integration; Internal Medicine; Interventional Radiology; Learning algorithms; Machine Learning; Male; Medical imaging; Medicine; Medicine & Public Health; Neuroradiology; Parameters; Plaque, Atherosclerotic - diagnostic imaging; Predictions; Predictive Value of Tests; Prognosis; Radiology; Regression analysis; Regression models; Retrospective Studies; Risk analysis; Risk Assessment; Risk factors; Segments; Stenosis; Tomography, X-Ray Computed; Ultrasound
• ISSN: 0938-7994; 1432-1084
• DOI: 10.1007/s00330-020-07083-2

Objectives To evaluate the long-term prognostic value of coronary CT angiography (cCTA)-derived plaque measures and clinical parameters on major adverse cardiac events (MACE) using machine learning (ML). Methods Datasets of 361 patients (61.9 ± 10.3 years, 65% male) with suspected coronary artery disease (CAD) who underwent cCTA were retrospectively analyzed. MACE was recorded. cCTA-derived adverse plaque features and conventional CT risk scores together with cardiovascular risk factors were provided to a ML model to predict MACE. A boosted ensemble algorithm (RUSBoost) utilizing decision trees as weak learners with repeated nested cross-validation to train and validate the model was used. Performance of the ML model was calculated using the area under the curve (AUC). Results MACE was observed in 31 patients (8.6%) after a median follow-up of 5.4 years. Discriminatory power was significantly higher for the ML model (AUC 0.96 [95%CI 0.93–0.98]) compared with conventional CT risk scores including Agatston calcium score (AUC 0.84 [95%CI 0.80–0.87]), segment involvement score (AUC 0.88 [95%CI 0.84–0.91]), and segment stenosis score (AUC 0.89 [95%CI 0.86–0.92], all p