Table 5 Classification metrics of the three models in the training and test datasets.

From: Combination of clinical information and radiomics models for the differentiation of acute simple appendicitis and non simple appendicitis on CT images

 

CT_train

CT_test

Radiomics_train

Radiomics_test

Combined_train

Combined_test

AUC (95% CI)

0.604 (0.494, 0.713)

0.669 (0.528, 0.809)

0.766 (0.674, 0.858)

0.804 (0.730, 0.878)

0.884 (0.820, 0.949)

0.817 (0.728, 0.905)

Cutoff

0.569

0.515

0.504

0.483

0.446

0.452

Accuracy (95% CI)

0.642 (0.637, 0.646)

0.702 (0.700, 0.704)

0.726 (0.723, 0.730)

0.728 (0.726, 0.730)

0.858 (0.856, 0.861)

0.754 (0.753, 0.756)

Sensitivity (95% CI)

0.566 (0.433, 0.699)

0.693 (0.630, 0.756)

0.660 (0.533, 0.788)

0.707 (0.645, 0.770)

0.887 (0.801, 0.972)

0.746 (0.687, 0.806)

Specificity (95% CI)

0.717 (0.596, 0.838)

0.783 (0.614, 0.951)

0.792 (0.683, 0.902)

0.913 (0.798, 1.000)

0.830 (0.729, 0.931)

0.826 (0.671, 0.981)

Positive predict value (95% CI)

0.667 (0.529, 0.804)

0.966 (0.937, 0.995)

0.761 (0.638, 0.884)

0.986 (0.968, 1.005)

0.839 (0.743, 0.935)

0.975 (0.950, 0.999)

Negative predict value (95% CI)

0.623 (0.501, 0.745)

0.222 (0.132, 0.313)

0.700(0.584, 0.816)

0.259 (0.164, 0.355)

0.880 (0.790, 0.970)

0.268 (0.165, 0.371)

Positive likelihood ratio (95% CI)

2.000 (1.226, 3.262)

3.186 (1.460, 6.956)

3.182 (1.817, 5.572)

8.134 (2.157, 30.670)

5.222 (2.857, 9.544)

4.291 (1.755, 10.495)

Negative likelihood ratio (95% CI)

0.605 (0.426, 0.86)

0.393 (0.292, 0.529)

0.429 (0.287, 0.639)

0.321 (0.250, 0.411)

0.136 (0.064, 0.293)

0.307 (0.227, 0.415)

  1. *CT_train: CT model evaluated in the training cohort. CT_test: CT model evaluated in the test cohort. Radiomics_train: radiomics model evaluated in the training cohort. Radiomics_test: radiomics model evaluated in the test cohort. Combined_train: the combined model evaluated in the training cohort. Combined_test: combined model evaluated in the test cohort.
Back to article page