Table 4 Eleven texture features with the best recognition performance in all selected brain regions of all sequences and their combined recognition performance.

From: Abnormal neonatal brain microstructure in gestational diabetes mellitus revealed by MRI texture analysis

Sequence

Encephalic region

Texture feature

AUC

95% CI

Sensitivity (%)

Specificity (%)

T2WI

RA

S(4,0)DifEntrp

0.723

0.602–0.843

79.4

59.5

FLAIR

GCC

S(5, -5)AngScMom

0.862

0.777–0.947

70.3

88.2

S(5, -5)Entropy

0.859

0.772–0.946

70.3

88.2

RH

WavEnLH_s-1

0.781

0.672–0.889

78.4

67.6

LCN

S(5,0)Contrast

0.727

0.610–0.845

89.2

47.1

DWI

GCC

S(2,2)SumAverg

0.719

0.599–0.839

51.4

94.1

LH

S(4,4)Entropy

0.709

0.585–0.833

81.1

55.9

RH

S(2, -2)Entropy

0.742

0.627–0.858

85.3

59.5

S(3,3)SumEntrp

0.723

0.602–0.844

61.8

78.4

S(3, -3)Entropy

0.734

0.617–0.851

52.9

83.8

S(4, -4)DifEntrp

0.763

0.652–0.873

64.7

78.4

Eleven-feature combination

0.980

0.957–1.000

97.1

89.2