Table 3 Performances of the deep learning models for patient-level diagnosis.
From: A novel deep learning model for a computed tomography diagnosis of coronary plaque erosion
AUC (95% CI) | Sensitivity (%) (95% CI) | Specificity (%) (95% CI) | PPV (95% CI) | NPV (95% CI) | FPR (%) (95% CI) | FNR (%) (95% CI) | |
|---|---|---|---|---|---|---|---|
Five-fold cross-validation | |||||||
MD-CTA model | 0.901 (0.873–0.930) | 81.2 (72.8–88.0) | 86.6 (82.4–90.2) | 68.4 (61.7–74.4) | 92.8 (89.8–95.0) | 13.4 (9.8–17.6) | 18.8 (12.0–27.2) |
CNN model | 0.621 (0.567–0.675) | 59.8 (50.1–69.0) | 60.2 (54.5–65.7) | 34.9 (30.4–39.7) | 80.8 (76.7–84.3) | 39.8 (34.3–45.5) | 40.2 (31.0–49.9) |
Test set validation | |||||||
MD-CTA model | 0.899 (0.841–0.957) | 87.1 (70.2–96.4) | 85.3 (75.3–92.4) | 71.1 (58.3–81.2) | 94.1 (86.5–97.6) | 14.7 (7.6–24.7) | 12.9 (3.6–29.8) |
CNN model | 0.724 (0.622–0.826) | 71.0 (52.0–85.8) | 68.0 (56.2–78.3) | 47.8 (38.1–57.8) | 85.0 (76.2–90.9) | 32.0 (21.7–43.8) | 29.0 (14.2–48.0) |
