Table 4. Comparison of the diagnostic performance of the BETNET model with three machine-learning algorithms in the validation dataset.
| Parameters | BETNET | SE_Net | SE_inception_v4 | Xception |
|---|---|---|---|---|
| AUC, 95% CI | 0.983 (0.973–0.990) | 0.963 (0.949–0.974) | 0.971 (0.959–0.980) | 0.964 (0.951–0.975) |
| Sensitivity (%) | 99.19 | 94.20 | 96.20 | 97.80 |
| Specificity (%) | 97.45 | 98.40 | 98.00 | 95.00 |
| Accuracy (%) | 98.3 | 96.3 | 97.1 | 96.4 |
| Youden index | 0.9663 | 0.9276 | 0.9420 | 0.9287 |
| P | – | 0.0004* | 0.0337* | 0.0027* |
AUCs of the BETNET model and the other three models were calculated by DeLong et al.’s. method. P: The difference of AUCs between the BETNET model and other three models was compared by Z-test; *, P<0.05. AUC, area under the ROC curve; CI, confidence interval.