Integrated prediction of lesion-specific ischaemia from quantitative coronary CT angiography using machine learning: a multicentre study

• Published in: European radiology Vol. 28; no. 6; pp. 2655 - 2664
• Main Authors: Dey, Damini, Gaur, Sara, Ovrehus, Kristian A., Slomka, Piotr J., Betancur, Julian, Goeller, Markus, Hell, Michaela M., Gransar, Heidi, Berman, Daniel S., Achenbach, Stephan, Botker, Hans Erik, Jensen, Jesper Moller, Lassen, Jens Flensted, Norgaard, Bjarne Linde
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01-06-2018; Springer Nature B.V
• Subjects: Adult; Aged; Aged, 80 and over; Angiography; Artificial intelligence; Atherosclerosis; Automation; Blood vessels; Calcification; Cardiac; Cardiovascular disease; Computed tomography; Computed Tomography Angiography - methods; Coronary Angiography - methods; Coronary artery; Coronary artery disease; Coronary Artery Disease - diagnostic imaging; Coronary Artery Disease - physiopathology; Coronary Stenosis - diagnostic imaging; Coronary Stenosis - physiopathology; Data processing; Density; Diagnostic Radiology; Female; Fractional Flow Reserve, Myocardial - physiology; Heart diseases; Hemodynamics; Humans; Imaging; Internal Medicine; Interventional Radiology; Ischemia; Learning algorithms; Machine Learning; Male; Medical imaging; Medicine; Medicine & Public Health; Middle Aged; Myocardial Ischemia - diagnostic imaging; Myocardial Ischemia - physiopathology; Neuroradiology; Patients; Plaque, Atherosclerotic - diagnostic imaging; Plaque, Atherosclerotic - physiopathology; Predictions; Radiology; Risk; Severity of Illness Index; Stenosis; Ultrasound; Vascular Calcification - diagnostic imaging; Vascular Calcification - physiopathology; Computed tomography angiography; Ischaemia; Atherosclerotic plaque; Coronary stenosis; Machine learning
• ISSN: 0938-7994; 1432-1084
• DOI: 10.1007/s00330-017-5223-z

Objectives
We aimed to investigate if lesion-specific ischaemia by invasive fractional flow reserve (FFR) can be predicted by an integrated machine learning (ML) ischaemia risk score from quantitative plaque measures from coronary computed tomography angiography (CTA).
Methods
In a multicentre trial of 254 patients, CTA and invasive coronary angiography were performed, with FFR in 484 vessels. CTA data sets were analysed by semi-automated software to quantify stenosis and non-calcified (NCP), low-density NCP (LD-NCP, < 30 HU), calcified and total plaque volumes, contrast density difference (CDD, maximum difference in luminal attenuation per unit area) and plaque length. ML integration included automated feature selection and model building from quantitative CTA with a boosted ensemble algorithm, and tenfold stratified cross-validation.
Results
Eighty patients had ischaemia by FFR (FFR ≤ 0.80) in 100 vessels. Information gain for predicting ischaemia was highest for CDD (0.172), followed by LD-NCP (0.125), NCP (0.097), and total plaque volumes (0.092). ML exhibited higher area-under-the-curve (0.84) than individual CTA measures, including stenosis (0.76), LD-NCP volume (0.77), total plaque volume (0.74) and pre-test likelihood of coronary artery disease (CAD) (0.63);
p
< 0.006.
Conclusions
Integrated ML ischaemia risk score improved the prediction of lesion-specific ischaemia by invasive FFR, over stenosis, plaque measures and pre-test likelihood of CAD.
Key Points
•
Integrated ischaemia risk score improved prediction of ischaemia over quantitative plaque measures
•
Integrated ischaemia risk score showed higher prediction of ischaemia than standard approach
•
Contrast density difference had the highest information gain to identify lesion-specific ischaemia