Skip to main content

Table 3 Performance on the test dataset: NODDI and routine MRI radiomic texture model

From: Radiomic texture analysis based on neurite orientation dispersion and density imaging to differentiate glioblastoma from solitary brain metastasis

Model

Accuracy (%)

AUC

(95% CI)

Sensitivity

(%)

Specificity

(%)

PPV

(%)

NPV

(%)

NODDI-Necrosis

72.1

0.713 (0.574–0.853)

77.2

58.8

82.9

50.0

NODDI-Solid tumor

85.2

0.821 (0.695–0.946)

93.1

64.7

87.2

78.5

NODDI-Peritumoral edema

78.6

0.762 (0.610–0.914)

81.8

70.6

87.8

60.0

NODDI-TBV

83.6

0.882 (0.789–0.975)

84.1

82.3

92.5

66.7

NODDI-ABV

72.1

0.779 (0.667–0.897)

65.9

88.2

93.5

50.0

Routine MRI-Necrosis

60.7

0.651 (0.483–0.819)

54.5

76.4

85.7

39.4

Routine MRI-Solid tumor

77.1

0.786 (0.649–0.923)

77.3

76.5

89.5

56.5

Routine MRI-Peritumoral edema

72.1

0.806 (0.692–0.919)

68.2

82.4

90.9

50.0

Routine MRI-TBV

78.7

0.723 (0.561–0.885)

86.3

58.8

84.4

62.5

Routine MRI-ABV

83.8

0.818 (0.684–0.952)

90.9

76.4

90.9

76.5

  1. 95% CI, 95% confidence interval; ABV, abnormal bulk volume; AUC, area under the ROC curve; NODDI, neurite orientation dispersion and density imaging; TBV, tumor bulk volume; MRI, magnetic resonance imaging; PPV, positive predictive value; NPV, negative predictive value