Table 5. Confusion matrix for classification of CAD, the model (single lead) vs. cCTA.
| Lead | Test data set (n = 1054) No. (%) | ||||
| True positive | False positive | True negative | False negative | Accuracy | |
| I | 247 (23.4) | 52 (4.9) | 385 (36.5) | 370 (35.2) | 632 (60.0) |
| II | 281 (26.6) | 220 (20.9) | 412 (39.1) | 141 (13.4) | 693 (65.7) |
| V1 | 157 (14.9) | 88 (8.3) | 544 (51.6) | 265 (25.2) | 701 (66.5) |
| V2 | 251 (23.8) | 191 (18.1) | 441 (41.8) | 171 (16.3) | 692 (65.7) |
| V3 | 177 (16.8) | 98 (9.3) | 534 (50.7) | 245 (23.2) | 711 (67.5) |
| V4 | 178 (16.9) | 101 (9.6) | 531 (50.4) | 244 (23.1) | 709 (67.3) |
| V5 | 94 (8.9) | 55 (5.2) | 577 (54.7) | 328 (31.2) | 671 (63.7) |
| V6 | 286 (27.1) | 272 (25.8) | 360 (34.2) | 136 (12.9) | 646 (61.3) |
The model performance in predicting CAD was assessed in the test group. Abbreviations: CAD: coronary artery disease; cCTA: coronary computed tomography angiography.