Transfer learning–based PET/CT three-dimensional convolutional neural network fusion of image and clinical information for prediction of EGFR mutation in lung adenocarcinoma

Table 3 Comparison of the prediction performance of deep learning models and radiomics models in the test set

Model	AUC (95%CI)	Accuracy	Sensitivity	Specificity	PPV	NPV
CT_RS	0.639 (0.529–0.749)	0.652	0.887	0.244	0.670	0.556
CT_TL	0.701 (0.595–0.808)	0.688	0.746	0.585	0.757	0.571
PET_RS	0.661 (0.552–0.769)	0.643	0.676	0.585	0.738	0.511
PET_TL	0.645 (0.534–0.756)	0.589	0.549	0.659	0.736	0.458
DS_RS	0.620 (0.509–0.730)	0.670	0.831	0.390	0.702	0.571
DS_TL	0.722 (0.622–0.822)	0.661	0.676	0.634	0.762	0.531
TS_RS	0.711 (0.613–0.809)	0.616	0.577	0.683	0.759	0.483
TS_TL	0.730 (0.629–0.830)	0.670	0.676	0.659	0.774	0.540

Bold numbers indicate the best results for each evaluation metric
AUC Area under the receiver operating characteristic curve, PPV Positive predictive value, NPV Negative predictive value, CT_RS CT radiomics, CT_TL CT transfer learning, PET_RS PET radiomics, PET_TL PET transfer learning, DS_RS PET/CT radiomics, DS_TL dual-stream transfer learning, TS_RS PET/CT radiomics combined with clinical features, TS_TL three-stream transfer learning

ISSN: 1471-2342