Skip to main content

Table 3 Predictive performance of different models in training and validation groups

From: Biparametric magnetic resonance imaging-based radiomics features for prediction of lymphovascular invasion in rectal cancer

  

Training group

  

Validation group

  

Models

AUC (95%CI)

Accuracy

Sensitivity

Specificity

AUC (95%CI)

Accuracy

Sensitivity

Specificity

cutoff

T2WI model

0.873 (0.804–0.941)

0.760

0.889

0.714

0.869 (0.728–1.00)

0.721

0.909

0.656

-1.12

DWI model

0.943 (0.883–1.000)

0.917

0.895

0.922

0.922 (0.836–1.000)

0.850

0.875

0.844

-0.749

Combined model

0.973 (0.945–1.000)

0.906

0.947

0.896

0.949 (0.880–1.000)

0.875

1.000

0.781

-1.26

  1. AUC Area under the receiver operating characteristics curve, CI Confidence interval, T2WI T2-weighted imaging, DWI Diffusion-weighted imaging
Back to article page

BMC Cancer

ISSN: 1471-2407

Contact us