Table 4 Performance of various machine learning algorithms in the training and test groups based on mpMRI sequences.

From: Preoperative MRI-based radiomics analysis of intra- and peritumoral regions for predicting CD3 expression in early cervical cancer

Model

Group

AUC(95%CI)

Accuracy (95% CI)

Sensitivity (95% CI)

Specificity (95% CI)

PPV (95% CI)

NPV (95% CI)

DecisionTree

Training group

0.861 (0.798–0.898)

0.766 (0.695–0.837)

0.968 (0.917–1.000)

0.608 (0.500–0.713)

0.659 (0.560–0.756)

0.96 (0.900–1.000)

Test group

0.567 (0.516–0.697)

0.459 (0.344–0.574)

0.667 (0.458–0.864)

0.35 (0.212–0.500)

0.35 (0.214–0.500)

0.667 (0.454–0.857)

AdaBoost

Training group

0.947 (0.880–0.979)

0.879 (0.823–0.929)

0.887 (0.800–0.962)

0.873 (0.795–0.939)

0.846 (0.746–0.929)

0.908 (0.842–0.969)

Test group

0.705 (0.567–0.835)

0.639 (0.508–0.754)

0.714 (0.500–0.895)

0.600 (0.444–0.750)

0.484 (0.313–0.655)

0.800 (0.643–0.933)

RandomForest

Training group

0.998 (0.968–1)

0.979 (0.950–1.000)

0.968 (0.912–1.000)

0.987 (0.96–1.000)

0.984 (0.948–1.000)

0.975 (0.931–1.000)

Test group

0.627 (0.491–0.773)

0.623 (0.492–0.738)

0.429 (0.211–0.650)

0.725 (0.583–0.861)

0.450 (0.231–0.684)

0.707 (0.548–0.846)

SVM

Training group

0.926 (0.875–0.978)

0.886 (0.829–0.936)

0.929 (0.86–0.984)

0.857 (0.782–0.929)

0.813 (0.708–0.909)

0.947 (0.892–0.988)

Test group

0.807 (0.696–0.919)

0.774 (0.661–0.871)

0.63 (0.455–0.808)

0.886 (0.763–0.974)

0.810 (0.632–0.957)

0.756 (0.614–0.884)

LogisticRegression

Training group

0.834 (0.783–0.879)

0.780 (0.709–0.851)

0.807 (0.702–0.902)

0.76 (0.663–0.852)

0.725 (0.616–0.822)

0.833 (0.743–0.913)

Test group

0.676 (0.626–0.706)

0.672 (0.557–0.787)

0.667 (0.476–0.864)

0.675 (0.524–0.829)

0.519 (0.357–0.71)

0.794 (0.657–0.926)

Back to article page