Table 3 Predictive performance of the clinical model, SVM-based radiomics model, and fusion model for predicting HER2 status 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

Cohort

Model

AUC (95% CI)

SEN (%)

SPE (%)

ACC (%)

Training

Clinical

0.716 (0.651–0.781)

65.17

73.81

70.23

SVM-based radiomics

0.893 (0.850–0.936)

87.64

82.54

84.65

Fusion

0.914 (0.877–0.952)

86.52

85.71

86.05

Internal validation

Clinical

0.673 (0.570–0.777)

63.16

70.37

67.39

SVM-based radiomics

0.822 (0.733–0.911)

84.21

77.78

80.44

Fusion

0.846 (0.764–0.929)

86.84

79.63

82.61

External validation 1

Clinical

0.635 (0.532–0.763)

47.83

80.49

68.75

SVM-based radiomics

0.786 (0.658–0.913)

69.57

82.93

78.13

Fusion

0.809 (0.700–0.918)

73.91

82.93

79.69

External validation 2

Clinical

0.641 (0.545–0.737)

46.30

88.06

69.42

SVM-based radiomics

0.834 (0.760–0.908)

81.48

79.10

80.17

Fusion

0.865 (0.796–0.934)

85.19

82.09

83.47

  1. Abbreviations: AUC, area under the curve; ACC, accuracy; CI, confidence interval; HER2, human epidermal growth factor receptor2; SEN, sensitivity; SPE, specificity; SVM, support vector machine.
Back to article page