Table 2 Area under the ROC curve after dropping one (or several) particular sequences (i.e.: setting the value to 0) in test time for the Gleason Grade Group \(\ge\)2 significance criterion.

From: Deep learning for fully automatic detection, segmentation, and Gleason grade estimation of prostate cancer in multiparametric magnetic resonance images

MRI sequence dropped

ProstateX

IVO

Lesion

Patient

Lesion

Patient

None (Baseline)

0.959

0.865

0.945

0.910

b400/500

0.944

0.861

0.940

0.868

b800/1000/1400

0.946

0.873

0.895

0.783

All b-numbers

0.951

0.844

0.845

0.720

ADC

0.905

0.870

0.940

0.836

\(K^{trans}\)

0.894

0.865

–

–

All DCE

–

–

0.895

0.820

All but T2

0.804

0.808

0.782

0.545

Back to article page