Skip to main content

Table 2 The performance of radiomics signatures in the testing set for benign and malignant prostate lesion discrimination

From: MRI-based radiomics models to assess prostate cancer, extracapsular extension and positive surgical margins

 

AUC

p-value

95 % CI

ACC

SEN

SPE

Cut-off

\({S}_{T2WI}\)

0.775

0.008

0.696, 0.842

0.699

0.654

0.782

0.622

\({M}_{T2WI}\)

0.851

0.780, 0.907

0.794

0.840

0.727

0.463

\({S}_{ADC}\)

0.863

0.024

0.793, 0.916

0.809

0.827

0.781

0.620

\({M}_{ADC}\)

0.912

0.852, 0.954

0.868

0.877

0.873

0.552

\({S}_{MP-MRI}\)

0.855

0.010

0.784, 0.909

0.801

0.815

0.782

0.574

\({M}_{MP-MRI}\)

0.905

0.843, 0.949

0.846

0.790

0.927

0.678

  1. The p-values were derived from DeLong tests. The first row of p-values compares the AUCs of the radiomics signatures with random guesses by chance (AUC of 0.5), and the second row compares the AUCs of the radiomics signatures with those of the corresponding integrated models
Back to article page

Cancer Imaging

ISSN: 1470-7330

Contact us