Skip to main content

Table 3 ROC curve analysis and comparison of predictive values

From: Assessing synchronous ovarian metastasis in gastric cancer patients using a clinical-radiomics nomogram based on baseline abdominal contrast-enhanced CT: a two-center study

 

Training set (n = 101)

External validation set (n = 46)

Nomogram

Radscore

Subjective evaluation

Nomogram

Radscore

Subjective evaluation

AUC

0.910

0.827

0.773

0.850

0.790

0.675

95% CI

0.845–0.959

0.740–0.901

0.690–0.856

0.738–0.963

0.652–0.927

0.539–0.810

Specificity

0.794

0.714

0.730

0.821

0.821

0.571

Sensitivity

0.921

0.816

0.816

0.833

0.778

0.778

Accuracy

0.842

0.752

0.762

0.826

0.804

0.652

PLR

4.464

2.855

3.023

4.667

4.356

1.815

NLR

0.099

0.258

0.252

0.203

0.270

0.389

PPV

0.729

0.633

0.646

0.750

0.737

0.538

NPV

0.943

0.865

0.868

0.885

0.852

0.800

Delong test (P value) *

/

0.007

0.026

/

0.186

0.047

NRI a

/

-0.185

-0.169

/

-0.056

-0.306

  1. AUC area under the curve, PLR positive likelihood ratio, NLR negative likelihood ratio, NPV negative predictive value, PPV positive predictive value, NRI net reclassification index
  2. *Compared with nomogram model
Back to article page

Cancer Imaging

ISSN: 1470-7330

Contact us