Table 2 Second-order features measuring fineness, coarseness and texture homogeneity; their trend shows that the CST is the one reporting less fine texture and more homogeneity in terms of texture, despite its wider histogram distribution.

From: Assessing robustness of quantitative susceptibility-based MRI radiomic features in patients with multiple sclerosis

Texture category

Feature

Trend

GLCM

Difference Entropy

ā†“ CST

Maximal Correlation Coefficient

ā†“ CST

Joint Entropy

ā†“ CST

Joint Energy

ā†‘ CST

GLRLM

Short Run Emphasis

ā†“ CST

Long Run Emphasis (*)

ā†‘ CST

Run Percentage (**)

ā†‘ CST

Gray Level Variance (***)

ā†“ CST

Run Entropy

ā†“ CST, OR

Run Length Non-Uniformity Normalised (****)

ā†“ CST

GLSZM

Long Area Emphasis (*)

ā†‘ CST

Zone Percentage (**)

ā†‘ CST

Gray Level Variance (***)

ā†“ CST, IFOF, OR

Zone Entropy

ā†“ CST, OR

NGTDM

Complexity

ā†“ CST, IFOF

GLDM

Small Dependence Emphasis

ā†“ CST

Large Dependence Emphasis (*)

ā†‘ CST

Dependence Non-Uniformity Normalised (****)

ā†“ CST

Gray Level Variance (***)

ā†“ CST

  1. CST =Ā cortico-spinal tract, IFOFĀ = inferior fronto-occipital fasciculus, OR =Ā optic radiation, GL =Ā grey levels, GLCMĀ = GL co-occurrence matrix, GLRLMĀ = GL run length matrix, GLSZM =Ā GL size zone matrix, NGDTM =Ā neighbouring grey difference tone matrix, GLDM =Ā GL dependence matrix.
  2. *Marks corresponding features in different categories.
Back to article page