Skip to main content
. 2023 Jun 13;18:426. doi: 10.1186/s13018-023-03909-z

Fig. 3.

Fig. 3

Diagnostic performance of the models and the reading clinicians. A The diagnostic performance of 8 clinicians was dotted in the ROC curves of the 2D and 3D CNN models according to the sensitivity and specificity for the surgery test set. B The diagnostic performance of 8 clinicians was dotted in the Precision and Recall (PR) curves of the 2D and 3D CNN models according to the precision and sensitivity for the surgery test set. C The diagnostic performance of 8 clinicians was dotted in the ROC curves of the 2D and 3D CNN models according to the sensitivity and specificity for the internal test set. D The diagnostic performance of 8 clinicians was dotted in the PR curves of the 2D and 3D CNN models according to the precision and sensitivity for the surgery test set on the internal test set