Table 3.
ROC analysis of the prediction model for the training and validation sets.
| Training set | Validation set | |||
|---|---|---|---|---|
| Method 1 | Method 2 | Method 1 | Method 2 | |
| AUC | 0.838 | 0.928 | 0.808 | 0.903 |
| 95% CI | 0.764-0.912 | 0.864-0.992 | 0.669-0.947 | 0.807-0.999 |
| Sensitivity | 0.871 | 0.903 | 0.956 | 0.870 |
| Specificity | 0.805 | 0.866 | 0.588 | 0.823 |
| Accuracy | 0.823 | 0.876 | 0.800 | 0.850 |
| PLR | 4.464 | 6.733 | 2.323 | 4.927 |
| NLR | 0.160 | 0.112 | 0.074 | 0.158 |
| PPV | 0.628 | 0.718 | 0.759 | 0.870 |
| NPV | 0.943 | 0.960 | 0.909 | 0.823 |
| P ∗ | 0.036 | 0.035 | ||
PLR: positive likelihood ratio; NLR: negative likelihood ratio; NPV: negative predictive value; PPV: positive predictive value. ∗Compared by DeLong test.