Table 4 Classification test on Manual, AI-only, and AI-assisted methods using cutoff points at the Standard (0.5), Optimum (maximum sensitivity and specificity), and maximum-sensitivity (Max-Sens)
Method | Test | CTR Cutoff | Sensitivity (%) | Specificity (%) | Accuracy (%) | F1 | AUC |
|---|---|---|---|---|---|---|---|
Manual | Standard | 0.5 | 100 | 83.6 | 89.3 | 0.866 | 0.978 |
Optimum | 0.518 | 95.8 | 91.8 | 93.2 | 0.907 | ||
Max-Sens | 0.505 | 100 | 86.0 | 90.9 | 0.884 | ||
AI-only | Standard | 0.5 | 97.5 | 82.8 | 87.9 | 0.849 | 0.962 |
Optimum | 0.521 | 90.3 | 91.0 | 90.8 | 0.872 | ||
Max-Sens | 0.219 | 100 | 0.10 | 34.8 | 0.516 | ||
AI-assisted | Standard | 0.5 | 100 | 84.3 | 89.8 | 0.873 | 0.977 |
Optimum | 0.516 | 96.0 | 91.3 | 92.95 | 0.904 | ||
Max-Sens | 0.501 | 100 | 85.0 | 90.2 | 0.876 |