Table 3.
Comparison of the performance and discriminative ability between the current and other models
| Cohort | Models | 1-year risk of EGVB | |
|---|---|---|---|
| AUROC (95% CI) | C-index (95% CI) | ||
| Training | ANN | 0.959 (0.945–0.973) | 0.956 (0.728–0.972) |
| NIEC | 0.669 (0.605–0.731) | 0.717 (0.646–0.735) | |
| Rev-NIEC | 0.725 (0.669–0.780) | 0.681 (0.636–0.726) | |
| Validation | ANN | 0.945 (0.877–0.987) | 0.936 (0.753–0.965) |
| NIEC | 0.743 (0.600–0.887) | 0.707 (0.643–0.772) | |
| Rev-NIEC | 0.797 (0.667–0.927) | 0.701 (0.631–0.771) | |
Abbreviations: ANN Artificial neural networks, AUROC Area under the receiver operating characteristic curves, CI Confidence interval, C-index Concordance index, EGVB Esophagogastric variceal bleeding, NIEC North Italian Endoscopic Club, Rev-NIEC Revised North Italian Endoscopic Club