Table 2 Diagnostic performance of different models for predicting PNETs in training and test groups.

Model	Cohort	AUC (95% CI)	Accuracy	Sensitivity	Specificity	PPV	NPV
Deep learning model^a	Training	0.948 (0.9108–0.9854)	0.898	0.865	0.920	0.877	0.912
Deep learning model^a	Test	0.795 (0.6929–0.8968)	0.775	0.805	0.744	0.767	0.784
Clinical model^a	Training	0.823 (0.7513–0.8942)	0.812	0.730	0.866	0.783	0.829
Clinical model^a	Test	0.847 (0.7639–0.9309)	0.775	0.683	0.872	0.848	0.723
Nomogram	Training	0.962 (0.9392–0.9843)	0.892	0.919	0.875	0.829	0.942
Nomogram	Test	0.871 (0.7958–0.9465)	0.787	0.732	0.846	0.833	0.750

Quick links

Search