Table 2 Association between genes modules and 18F-FDG PET features.

From: Transcriptomics in cancer revealed by Positron Emission Tomography radiomics

Module #

Regulator 1

Regulator 2

Regulator 3

Groups size

1

Irregularity

29.5%

Angular second moment

6.0%

Inverse difference moment

5.9%

[14|10][05|16]

2

Irregularity

27.8%

Angular second moment

5.8%

High-intensity large area emphasis

22.8%

[11|11][07|16]

3

Irregularity

27.0%

Maximum distance to background

22.5%

Maximum distance to background

5.7%

[15|16][09|14]

4

Irregularity

29.5%

Angular second moment

6.5%

MATV

4.9%

[14|11][05|15]

5

Intensity variability

25.9%

Large area emphasis

15.4%

[16|23][06]

6

SUVCOV

79.2%

Maximum distance to background

5.7%

[15|24][06]

7

Angular second moment

18.3%

Compactness v2

25.2%

High-intensity large area emphasis

26.5%

[10|24][07|04]

8

Zone percentage

77.3%

Inertia

14.0%

[08|31][29]

9

ratio 3ds vol

25.5%

Maximum distance to background

22.5%

Large area emphasis

15,4%

[11|07][09|18]

10

Size-zone variability

6.6%

Angular second moment

19.7%

High-intensity emphasis

25.7%

[07|04][06|28]

11

Irregularity

27.8%

Large area emphasis

26.9%

Dissimilarity

71.8%

[12|10][17|06]

12

Irregularity

29.5%

Angular second moment

5.9%

Inverse difference moment

5.9%

[11|14][05|15]

13

Intensity variability

25.9%

Maximum distance to background

5.7%

[15|24][06]

14

Compactness v2

86.5%

Large area emphasis

21.7%

[20|19}[06]

15

Homogeneity

18.6%

SUVmax

5.5%

Inverse difference moment

20.0%

[08|08][05|24]

16

High-intensity large area emphasis

19.3%

Large area emphasis

20.3%

Intensity variability

6.1%

[07|05][13|20]

17

Irregularity

27.8%

Large area emphasis

28.0%

Maximum distance to background

18.1%

[12|10][11|12]

18

Intensity variability

25.9%

Large area emphasis

15.4%

[16|23][06]

19

Inertia

20.6%

Large area emphasis

27.7%

[16] [25|04]

20

Inertia

20.6%

SUVCOV

34.4%

Zone percentage

73.1%

[07|09][16|13]

  1. The cut-off from regulator 1 splits patients into two groups, that are then split using the regulator 2 and 3. Regulator 2 and 3 are applied to the groups of patients with a value < and > = the threshold on the regulator 1, respectively. The values are expressed as percentage based on the distribution within the cohort. Heterogeneity feature, shape feature.
  2. Values for each regulator correspond to the cut-off between the two groups.