Table a - List of the 107 Image Features Used for Texture Analysis(Supplementary)

Table A - List of the 107 image features used for texture analysis(Supplementary).

1st order statistics *
Energy
Entropy
Kurtosis
Maximum
Mean
Mean absolute deviation
Median
Minimum
Range
Root mean square
Skewness
Standard deviation
Uniformity
Variance
2nd order statistics
i) Grey level co-occurrence matrix
Autocorrelation
Cluster prominence
Cluster shade
Cluster tendency
Contrast
Correlation
Difference entropy
Dissimilarity
Energy
Entropy
Homogeneity 1
Homogeneity 2
Maximum probability
Sum average
Sum entropy
Sum variance
Variance
Information measure of correlation 1
Information measure of correlation 2
Inverse difference normalised
Inverse difference moment normalised
ii) Grey level run length matrix
Short run emphasis
Long run emphasis
Grey level non uniformity
Run length non uniformity
Run percentage
Low grey level run emphasis
High grey level run emphasis
Short run low grey level emphasis
Short run high grey level emphasis
Long run low grey level emphasis
Long run high grey level emphasis
3D Shape and size
Compactness 1
Compactness 2
Sphericity
Surface-to-volume ratio
Volume

Note. – * without and with Laplacian of Gaussian filtering (σ =1;1.5;2;2.5)

Table B - Multivariate analysis and relative risk for death according to preoperative imaging-based T and N staging and histology (Supplementary).

Total population (n = 56)
threshold / Log RR / SE / p
T stage / T1-3 vs. T4 / 1.84 / 0.80 / 0.022
N stage / N0 vs. N+ / 20.27 / 9.42 x 103 / 0.998
Histology / ADK vs. SRC / 0.62 / 0.67 / 0.349

Note: ADK = adenocarcinoma; SRC= signet-ring cell

Table C - Multivariate analysis and relative risk for death when adding imaging-based T and N staging and histology in our model (Supplementary).

Total population (n = 56)
Threshold / Log RR / SE / p
Energy / 2.2 x 108 / 3.62 / 3.04 / 0.233
Entropy / 6.81 / 6.22 / 1.97 / 0.001
Filter 1
Root mean square / 9.31 / -2.13 / 2.19 / 0.331
Filter 1.5
Entropy / 5.51 / 5.08 / 1.96 / 0.009
Maximum Hounsfield unit value / 57.95 / 4.95 / 1.82 / 0.006
Filter 2
Mean absolute deviation / 2.66 / -5.07 / 1.99 / 0.011
Skewness / 2.91 / 5.68 / 3.79 / 0.134
T stage / T1-3 vs. T4 / -0.40 / 2.04 / 0.843
N stage / N0 vs. N+ / 0.21 / 1.59 x 104 / 0.998
Histology / ADK vs. SRC / 2.58 / 1.74 / 0.137

Note: Log RR = logarithm of relative risk; SE = standard error; p = p value; Max. HU = maximum Hounsfield unit; ADK = adenocarcinoma; SRC= signet-ring cell

Table D - Multivariate analysis and relative risk for death when adding imaging-based T and N staging and histology in our model, after retaining relevant features through a stepwise procedure (Supplementary).

Threshold / Log RR / SE / p
Energy / 2.2 x 108 / 3.41 / 1.74 / 0.049
Entropy / 6.81 / 6.09 / 2.07 / 0.003
Entropy (filter 1.5) / 5.51 / 4.80 / 1.84 / 0.009
Maximum HU value (filter 1.5) / 57.95 / 4.87 / 1.81 / 0.007
Mean absolute deviation (filter 2) / 2.66 / -5.27 / 1.67 / 0.001
Skewness (filter 2) / 2.91 / 4.39 / 1.96 / 0.025
N stage at imaging / N0 vs. N+ / 23.52 / 1.64 x 104 / 0.998
Histology / ADK vs. SRC / 2.09 / 1.31 / 0.109

Note: Log RR = logarithm of relative risk; SE = standard error; p = p value; HU = Hounsfield unit; ADK = adenocarcinoma; SRC= signet-ring cell