Skip to main content

Table 4 Top 10 radiomic features with highest gain value in 2D-analysis patch-based model

From: Detection of pancreatic cancer with two- and three-dimensional radiomic analysis in a nationwide population-based real-world dataset

Feature

Gain value

NGTDM: busyness (original)

9735.6

First order: median (original)

8999.0

GLSZM: large area emphasis (wavelet-HL)

4841.2

GLDM: dependence entropy (wavelet-LL)

4013.8

First order: mean absolute deviation (wavelet-HL)

3577.7

NGTDM: coarseness (wavelet-LH)

2750.1

GLDM: dependence non-uniformity (wavelet-HL)

2456.0

GLRLM: long run emphasis (wavelet-HL)

2421.2

GLCM: correlation (wavelet-LL)

2407.8

First order: skewness (original)

2298.6

  1. NGTDM Neighboring Gray Tone Difference Matrix, GLSZM Gray Level Size Zone Matrix, GLDM Gray Level Dependence Matrix, GLRLM Gray Level Run Length Matrix, GLCM Gray Level Co-occurrence Matrix
Back to article page