Table 2 Predictive performance of various machine learning-based radiomics models for predicting HER2 status of EC in the training, internal validation, and external validation cohorts.

From: Preoperative prediction of the HER2 status and prognosis of patients with endometrial cancer using multiparametric MRI-based radiomics: a multicenter study

Model

Training cohort

Internal validation cohort

External validation cohort 1

External validation cohort 2

Validation cohorts

AUC (95% CI)

ACC (%)

AUC (95% CI)

ACC (%)

AUC (95% CI)

ACC (%)

AUC (95% CI)

ACC (%)

Average AUC

Average ACC (%)

KNN

0.782 (0.717–0.846)

73.95

0.693 (0.587–0.799)

67.39

0.649 (0.509–0.789)

65.63

0.714 (0.622–0.806)

68.60

0.685

67.21

SVM

0.893 (0.850—0.936)

82.54

0.822 (0.733–0.911)

80.44

0.786 (0.658–0.913)

78.13

0.834 (0.760–0.908)

80.17

0.814

79.58

LR

0.843 (0.789–0.898)

76.19

0.727 (0.619–0.835)

76.09

0.691 (0.554–0.829)

67.19

0.772 (0.688–0.855)

74.38

0.730

72.55

RF

0.817 (0.759–0.874)

80.95

0.736 (0.636–0.836)

69.57

0.584 (0.437–0.730)

62.50

0.671 (0.573–0.769)

68.60

0.664

66.89

NB

0.906 (0.861–0.950)

84.92

0.628 (0.505–0.750)

68.48

0.714 (0.576–0.851)

76.56

0.652 (0.552–0.751)

67.77

0.665

70.94

XGBoost

0.865 (0.815–0.915)

84.13

0.769 (0.673–0.865)

72.83

0.681 (0.539–0.824)

68.75

0.713 (0.621–0.806)

69.42

0.721

70.33

  1. Abbreviations: AUC, area under the curve; ACC, accuracy; CI, confidence interval; EC, endometrial cancer; HER2, human epidermal growth factor receptor2; KNN, k-nearest neighbors; SVM, support vector machine; LR, logistic regression; RF, random forest; NB, naive Bayes; XGBoost, eXtreme Gradient Boosting.
Back to article page