Fig. 2: Kidney cancer diagnosis performance of the model and radiologists.
From: Deep learning for end-to-end kidney cancer diagnosis on multi-phase abdominal computed tomography
a The ROC curves of the model and the performance of the six radiologists are plotted for each tumor subtype. For the three RCCs, the results of testing the model on the TCIA dataset are also plotted. b Confusion matrices for the model and individual radiologists. Comparison of the sensitivities (c) and specificities (d). P values were calculated using the two-sided permutation test.
