A Simple Method to Train the AI Diagnosis Model of Pulmonary Nodules

Table 3

AUC and ROC curve best cutoff of the trained model and the doctors.

	AUC	Optimal cutoff	Sensitivity (%)	Specificity (%)	(compared to the trained model)
Trained model (data augmentation)	0.896	0.791	88.2	90.9
Doctor A	0.759	0.519	88.2	63.6	0.1212
Doctor B	0.775	0.549	88.2	66.7	0.1673
Doctor C	0.745	0.490	82.4	66.7	0.0963