Multifeature Quantification of Nuclear Properties from Images of H&E-Stained Biopsy Material for Investigating Changes in Nuclear Structure with Advancing CIN Grade
Table 3
Morphological and textural features with statistically significant differences between low and high CIN classes, ranked alphabetically.
Feature name
corrected
r (correlation)
Low CIN
High CIN
Mean
SE
Mean
SE
1
dwt2D Mean Absolute Deviation
4.48E − 02
1.34E − 01
−0.34
6.23
0.13
5.50
0.28
2
dwt2H Mean Absolute Deviation
2.02E − 03
1.82E − 02
−0.39
10.44
0.24
8.67
0.59
3
dwt2H Mean Value
6.25E − 03
3.03E − 02
0.27
−0.84
0.029
−0.68
0.080
4
dwt2H Median Absolute Deviation
7.20E − 03
3.24E − 02
−0.33
1.17
0.065
0.95
0.072
5
dwt2H Median Value
5.03E − 05
3.17E − 03
0.18
−0.014
0.0024
−0.0079
0.0048
6
dwt2H Standard Deviation
2.50E − 02
7.87E − 02
−0.31
23.33
0.39
20.80
1.10
7
Gray-level nonuniformity
2.02E − 03
2.13E − 02
0.52
85.21
1.98
116.83
7.71
8
Kurtosis
2.98E − 03
2.35E − 02
0.37
2.59
0.02
2.71
0.04
9
Local binary pattern mean value
1.02E − 02
3.77E − 02
0.40
16.61
0.23
18.03
0.45
10
Local binary pattern standard deviation
1.37E − 04
4.30E − 03
−0.47
0.12
0.00082
0.12
0.0014
11
Nucleus area
7.20E − 03
3.02E − 02
0.46
1141
50
1640
137
12
Nucleus convex area
8.27E − 03
3.26E − 02
0.45
1253
53
1772
147
13
Nucleus equivalent diameter
5.03E − 03
2.64E − 02
0.47
36.53
0.78
43.59
1.85
14
Nucleus extent
5.00E − 02
1.43E − 01
0.22
0.64
0.0045
0.66
0.0079
15
Nucleus major axis length
1.16E − 02
4.07E − 02
0.37
52.6
1.7
60.1
2.3
16
Nucleus minor axis length
3.75E − 03
2.36E − 02
0.47
28.0
0.5
34.1
1.7
17
Nucleus perimeter
1.24E − 02
4.12E − 02
0.40
135.9
3.2
156.6
6.4
18
Nucleus solidity
1.37E − 04
2.87E − 03
0.52
0.91
0.0027
0.93
0.0037
19
Tamura coarseness 1
1.15E − 03
1.81E − 02
0.45
7.12
0.12
8.01
0.24
20
Tamura coarseness 2
4.04E − 03
2.31E − 02
0.48
23.88
0.57
33.99
2.77
21
Tamura coarseness 3
2.98E − 03
2.09E − 02
0.45
9.11
0.37
14.27
1.51
22
Tamura coarseness 4
1.73E − 03
2.18E − 02
0.49
18.21
0.70
23.94
1.40
Benjamini and Hochberg FDR correction; MATLAB function: mafdr(); SE = standard error.
