Table 1. List of all relevant studies for coronary anatomy imaging included in this systematic review.
| First author | Year | Model output | Sample size | Imaging modality | Model | Index test | Reference test | External validation |
|---|---|---|---|---|---|---|---|---|
| Rodrigues et al. (3) | 2016 | Pericardial and mediastinal fat classification | 20 | CCTA | RF | Manual feature extraction algorithms | Expert reader | No |
| Kang et al. (4) | 2015 | Coronary stenosis classification | 42 | CCTA | SVM | Expert reader | Invasive coronary angiography | No |
| Araki et al. (5) | 2016 | Coronary plaque calcification | 15 | IVUS | SVM | NA | cIMT | No |
| Itu et al. (6) | 2016 | FFR prediction | 87 | CCTA | MLP | Computational fluid dynamics CT-FFR | Invasive FFR | Yes |
| Wolterink et al. (7) | 2016 | CAC quantification | 250 | CCTA | CNN | NA | Expert reader | No |
| Su et al. (8) | 2017 | Media adventitia border detection | 4 | IVUS | MLP | NA | Expert reader | No |
| Yong et al. (9) | 2017 | Coronary lumen segmentation | 64 | OCT | CNN | NA | Expert reader | No |
| Xu et al. (10) | 2017 | Coronary plaque classification | 18 | OCT | CNN and SVM | NA | Expert reader | No |
| Zreik et al. (11) | 2019 | Coronary plaque classification | 163 | CCTA | RNN | NA | Expert reader | No |
| Zreik et al. (12) | 2018 | LV segmentation for coronary stenosis significance classification | 156 | CCTA | CNN + SVM | NA | Invasive FFR | No |
| Kolluru et al. (13) | 2018 | Coronary plaque classification | 48 | OCT | CNN | NA | Expert reader | No |
| Zhang et al. (14) | 2018 | Coronary plaque classification | 61 | IVUS | SVM | NA | Expert reader | No |
| Oh et al. (15) | 2018 | Lipid core plaque detection | 116 | IVUS | CNN | NA | Expert reader | No |
| van Rosendael et al. (16) | 2018 | Clinical outcome prediction | 8,844 | CCTA | Boosted ensemble algorithm | Conventional clinical risk scores | Clinical outcomes | No |
| Stuckey et al. (17) | 2018 | CAD detection | 606 | cPSTA | Elastic net | NA | Invasive coronary angiography | No |
| Lessmann et al. (18) | 2018 | CAC detection | 1,744 | CCTA | CNN | NA | Expert reader | No |
| Šprem et al. (19) | 2018 | Motion artefact detection in CACS | 585 | CCTA | CNN | NA | Conventional CACS | No |
| Hae et al. (20) | 2018 | Prediction of myocardium subtended by coronary stenosis | 932 | CCTA | SVM | NA | Invasive coronary angiography | Yes |
| Dey et al. (21) | 2018 | FFR prediction | 254 | CCTA | Boosted ensemble algorithm | Conventional CCTA | Invasive FFR | No |
| van Hamersvelt et al. (22) | 2019 | LV segmentation for coronary stenosis significance classification | 126 | CCTA | SVM | NA | Expert reader | No |
| Cho et al. (23) | 2019 | FFR classification | 1,501 | Invasive coronary angiography | XGBoost | NA | Invasive FFR | Yes |
| Liu et al. (24) | 2019 | Vulnerable plaque detection | 2,300 (images) | OCT | CNN | NA | Expert reader | No |
| Gessert et al. (25) | 2019 | Coronary plaque segmentation | 49 | OCT | CNN | NA | Expert reader | No |
| Abdolmanafi et al. (26) | 2019 | Coronary artery wall pathology detection | 45 | OCT | CNN | NA | Expert reader | No |
| Liu et al. (27) | 2019 | Bifurcation lesion detection | 308 | Invasive coronary angiography | CNN | NA | Expert reader | No |
| Gharaibeh et al. (28) | 2019 | CAC quantification | 34 | IVUS | CNN | NA | Expert reader | No |
| Jun et al. (29) | 2019 | Thin cap fibroatheroma classification | 100 | IVUS | CNN | NA | OCT | No |
| Lee et al. (30) | 2019 | Coronary artery segmentation | 4,980 | Invasive coronary angiography | CNN | NA | Expert reader | No |
| Yang et al. (31) | 2019 | Coronary artery segmentation | 2,042 | Invasive coronary angiography | CNN | NA | Expert reader | Yes |
| Wang et al. (32) | 2019 | Media adventitia border detection | 22 | IVUS | MLP | P6 and P8 detectors | Expert reader | No |
| Johnson et al. (33) | 2019 | Clinical outcome prediction | 6,892 | CCTA | KNN | Conventional CT and clinical risk scores | Clinical outcomes | No |
| Kolossváry et al. (34) | 2019 | Coronary plaque classification | 21 | CCTA | Least angle regression + radiomics | Histogram assessment by expert reader | Histology (ex vivo) | No |
| Wang et al. (35) | 2019 | FFR prediction | 63 | CCTA | RNN | Conventional CCTA | Invasive FFR | No |
| Datong et al. (36) | 2019 | CAC detection | 820 (images) | CCTA | CNN | NA | Expert reader | No |
| Oikonomou et al. (37) | 2019 | Clinical outcome prediction | 5,487 | CCTA | RF + radiomics | Conventional clinical risk scores | Clinical outcomes | Yes |
| Masuda et al. (38) | 2019 | Coronary plaque classification | 78 | CCTA | Extreme gradient boosting | Conventional CCTA | IVUS | No |
| Kigka et al. (39) | 2019 | Coronary plaque progression prediction | 40 | CCTA | RF | NA | Clinical outcomes | No |
| Zhang et al. (40) | 2019 | Coronary risk prediction | 4,415 | CCTA | Boosted ensemble algorithm | Conventional clinical risk scores | Clinical outcomes | No |
| Commandeur et al. (41) | 2019 | Epicardial adipose tissue quantification | 850 | CCTA | CNN | NA | Expert reader | No |
| Hong et al. (42) | 2019 | Coronary artery segmentation | 156 | CCTA | CNN | NA | Expert reader | No |
| Huo et al. (43) | 2019 | CAC detection | 2,332 | CCTA | CNN | NA | Expert reader | No |
| Wang et al. (44) | 2020 | MPVI prediction | 9 | IVUS | SVM and RF | GLMM | Follow-up MPVI | No |
| Lee et al. (45) | 2020 | FFR prediction | 1,328 | IVUS | AdaBoost | NA | Invasive FFR | No |
| Wu et al. (46) | 2020 | Coronary stenosis detection | 63 | Invasive coronary angiography | CNN | NA | Expert reader | No |
| Sampedro-Gómez et al. (47) | 2020 | Stent restenosis prediction | 263 | Invasive coronary angiography | ERT | Conventional clinical risk scores | Clinical outcomes | No |
| Miyoshi et al. (48) | 2020 | Coronary neointimal coverage classification, yellow colour classification, red thrombus detection | 107 | Invasive coronary angioscopy | GAN | SVM | Expert reader | Yes |
| Zhang et al. (49) | 2020 | Coronary stenosis classification | 228 | Invasive coronary angiography | HEAL | NA | Expert reader | No |
| Du et al. (50) | 2021 | Coronary artery segmentation, stenosis classification, total occlusion detection, calcification detection, thrombus detection, dissection detection | 10,073 | Invasive coronary angiography | CNN and GAN | NA | Expert reader | No |
| He et al. (51) | 2020 | Coronary plaque segmentation | 24 | OCT | CNN | NA | Expert reader | No |
| Yabushita et al. (52) | 2021 | Coronary artery segmentation | 146 | Invasive coronary angiography | CNN | NA | Expert reader | No |
| Hamaya et al. (53) | 2020 | Clustering epicardial functional stenosis with low CFR | 364 | Invasive coronary angiography | Unsupervised hierarchical clustering | K-mean clustering | Clinical outcomes | No |
| Lee et al. (54) | 2019 | Coronary plaque segmentation | 55 | OCT | CNN | A-line CNN detector | Expert reader | No |
| Min et al. (55) | 2020 | Thin cap fibroatheroma classification | 602 | OCT | CNN | NA | Expert reader | No |
| Commandeur et al. (56) | 2020 | Clinical outcome prediction | 1,912 | CCTA | Extreme gradient boosting | Conventional CT and clinical risk scores | Clinical outcomes | No |
| Muscogiuri et al. (57) | 2020 | CAD classification | 288 | CCTA | CNN | NA | Expert reader | No |
| Benz et al. (58) | 2020 | Coronary artery image reconstruction | 43 | CCTA | CNN | Adaptive statistical iterative reconstruction | Invasive coronary angiography | No |
| Wang et al. (59) | 2020 | CAC quantification | 530 | CCTA | CNN | NA | Expert reader | No |
| Al’Aref et al. (60) | 2020 | Coronary stenosis prediction from CACS | 13,054 | CCT | Boosted ensemble algorithm | NA | CCTA | No |
| Kawasaki et al. (61) | 2020 | FFR prediction | 47 | CCTA | RF | NA | Invasive FFR | No |
| Fischer et al. (62) | 2020 | CAC quantification | 200 | CCTA | RNN | NA | Expert reader | No |
| van Velzen et al. (63) | 2020 | CAC quantification | 7,240 | CCTA | CNN | NA | Expert reader | No |
| Zreik et al. (64) | 2020 | Coronary stenosis classification | 187 | CCTA | CNN + SVM | NA | Invasive FFR | No |
| Kumamaru et al. (65) | 2020 | FFR prediction | 1,052 | CCTA | CNN + GAN | Conventional CCTA | Invasive FFR | No |
| Candemir et al. (66) | 2020 | Coronary stenosis classification | 493 | CCTA | CNN | NA | Expert reader | Yes |
| Shu et al. (67) | 2022 | Clinical outcome prediction | 154 | CCTA | SVM + radiomics | NA | Expert reader | Yes |
| van den Oever et al. (68) | 2020 | CAC rule out | 100 | CCTA | CNN | NA | Expert reader | Yes |
| Han et al. (69) | 2020 | Coronary stenosis classification | 150 | CCTA | CNN | Expert reader | Invasive coronary angiography | No |
| Han et al. (70) | 2020 | Rapid plaque progression prediction | 1,083 | CCTA | Boosted ensemble algorithm | Conventional clinical risk scores | Clinical outcomes | No |
| Lin et al. (71) | 2020 | Pericoronary adipose tissue prognosis prediction | 177 | CCTA | Boosted ensemble algorithm + radiomics | Conventional CT and clinical risk scores | Clinical outcomes | No |
| Chen et al. (72) | 2020 | Coronary artery segmentation | 124 | CCTA | CNN | Expert reader | Invasive coronary angiography | No |
| Tesche et al. (73) | 2021 | Clinical outcome prediction | 361 | CCTA | Boosted ensemble algorithm | Conventional CT and clinical risk scores | Clinical outcomes | No |
| Al’Aref et al. (74) | 2020 | CL precursors detection | 46 | CCTA | Boosted ensemble algorithm | Traditional CCTA CL precursors | Invasive coronary angiography | Yes |
| Hong et al. (75) | 2020 | CCTA image noise reduction | 82 | CCTA | CNN | NA | Invasive coronary angiography | No |
| Podgorsak et al. (76) | 2020 | Coronary segmentation and FFR prediction | 64 | CCTA | CNN | Expert reader | Invasive FFR | No |
| Eberhard et al. (77) | 2020 | FFR prediction | 56 | CCTA | CNN | Invasive FFR | Clinical outcomes | No |
| Son et al. (78) | 2020 | CAC prediction | 20,130 | Retinal fundus imaging | CNN | NA | CCTA | No |
| Carson et al. (79) | 2020 | FFR prediction | 25 | CCTA | MLP and RNN | MPR | Invasive FFR | Yes |
| Gangl et al. (80) | 2019 | Coronary plaque segmentation | 104 (images) | OCT | CNN | NA | Expert reader | No |
| Głowacki et al. (81) | 2020 | Coronary stenosis prediction from CACS | 435 | CCT | Extreme gradient boosting | NA | CCTA | No |
| Hoshino et al. (82) | 2020 | FAI clusters | 220 | CCTA | Unsupervised hierarchical clustering | Invasive FFR | Clinical outcomes | No |
| Kawaguchi et al. (83) | 2018 | FFR prediction | 934 | CCTA | CNN | NA | Invasive FFR | No |
CCTA, coronary computed tomographic angiography; RF, random forest; SVM, support vector machine; IVUS, intra-vascular ultrasound; NA, not available; cIMT, carotid intima-media thickness; FFR, fractional flow reserve; MLP, multi-layer perceptron; CT, computed tomography; CAC, coronary artery calcification; CNN, convolutional neural network; OCT, optical coherence tomography; RNN, recurrent neural network; LV, left ventricle; CAD, coronary artery disease; cPSTA, cardiac phase space tomography analysis; CACS, coronary artery calcium score; KNN, k-nearest neighbours; MPVI, morphological plaque vulnerability index; GLMM, generalised linear mixed model; ERT, extremely randomised tree; GAN, generative adversarial network; HEAL, hierarchical attentive multi-view; CFR, coronary flow reserve; CL, culprit lesion; MPR, multi-variant polynomial regression; FAI, fat attenuation index.