Table 3. Performance of multiparametric radiomics nomogram and liver radiomics in diagnosing EV.
Variable | Multiparametric radiomics nomogram | LL Rad-sore | RL Rad-sore | |||||
---|---|---|---|---|---|---|---|---|
Train (n=95) | Validation (n=41) | Train (n=95) | Validation (n=41) | Train (n=95) | Validation (n=41) | |||
AUC | 0.95 (0.911–0.99) | 0.92 (0.78–1) | 0.88 (0.80–0.96) | 0.87 (0.67–1) | 0.86 (0.77–0.94) | 0.83 (0.65–1) | ||
Mean cross-validated AUC | 0.87 | 0.70 | 0.76 | |||||
ACC | 0.91 | 0.93 | 0.85 | 0.81 | 0.83 | 0.81 | ||
Sensitivity | 0.86 | 0.89 | 0.71 | 0.78 | 0.70 | 0.78 | ||
Specificity | 0.92 | 0.94 | 0.89 | 0.81 | 0.87 | 0.81 | ||
Cut off value | 0.67 | 0.74 | 0.67 |
Severity in cirrhosis in the training and validation cohorts. AUC, the area under the curve; ACC, accuracy; EV, esophageal varices.