Table 3 The performance of the 15 selected models from Integration Set in validation cohort.

From: Differentiating solitary brain metastases from glioblastoma by radiomics features derived from MRI and 18F-FDG-PET and the combined application of multiple models

Group

Model

AUC

ACC

Sensitivity

Specificity

A

LASSO-SVM

0.93

0.83

0.76

0.92

LDA-SVM

0.92

0.86

0.84

0.91

LASSO-LR

0.91

0.89

0.88

0.9

LDA-LR

0.90

0.87

0.84

0.91

LDA-KNN

0.90

0.85

0.80

0.91

B

PLS-LR

0.86

0.78

0.76

0.82

NCA-KNN

0.84

0.83

0.86

0.81

PLS-RF

0.83

0.80

0.80

0.80

PLS-SVM

0.83

0.78

0.82

0.74

PLS-Adaboost

0.83

0.82

0.82

0.82

C

PCA-RF

0.80

0.72

0.74

0.70

NCA-Adaboost

0.79

0.78

0.82

0.74

PCA-LR

0.78

0.77

0.66

0.88

PCA-Adaboost

0.78

0.77

0.70

0.84

LASSO-Adaboost

0.68

0.67

0.68

0.65

  1. AUC area under curve, ACC accuracy, LASSO least absolute shrinkage and selection operator, LDA linear discriminant analysis, PLS partial least squares regression, NCA near-collar component analysis, PCA principal component analysis, SVM support vector machine, LR the logistic regression, KNN K nearest neighbors, RF random forest, Adaboost Adaptive Boosting.
Back to article page