Table A3.

Explanatory table of contributive radiomics features.

Radiomic Feature	Formula	Explanation
GLDM: Grey level Dependance Matrix: Dependencies between a central voxel and voxels at a given length. Dependency is defined by a maximal spread of density between those voxels. Ng number of grey levels Nd number of discreet lengths in the image Nz number of dependencies zones in the image P(i,j) is the dependency matrix
gldmDNUN	$\frac{{\displaystyle {\sum }_{i=1}^{N_g}{\left({\displaystyle {\sum }_{j=1}^{N_d}\mathsf{P}\left(i,j\right)}\right)}^2}}{N_z}$	Measures the grey level similarity in the ROI
gldmDLGLE	$\frac{{\displaystyle {\sum }_{i=1}^{N_g}{\displaystyle {\sum }_{j=1}^{N_d}\frac{\mathsf{P}\left(i,j\right)j^2}{i^2}}}}{N_z}$	Distribution of large matrix of dependencies of low grey levels
GLSZM: Grey Level Size Zone Matrix: a grey level zone is a zone of contiguous voxels with a similar grey level. Ng number of grey levels Ns number of possible sizes of grey levels in the ROI Np number of voxel in the ROI Nz numbe of zones in the image P(i,j) size of the zone (matrix)
glszmGLNUN	$\frac{{\displaystyle {\sum }_{i=1}^{N_g}{\left({\displaystyle {\sum }_{j=1}^{N_s}\mathsf{P}\left(i,j\right)}\right)}^2}}{N_z^2}$	Variability of the grey levels of zones. A small value is in favour of image homogeneity
glszmSAE	$\frac{{\displaystyle {\sum }_{i=1}^{N_g}{\displaystyle {\sum }_{j=1}^{N_s}\frac{\mathsf{P}\left(i,j\right)}{j^2}}}}{N_z}$	Number of small size zones
Shape features
ShapeVV	$V_{voxel}={\displaystyle \sum _{k=1}^{N_v}{V}_k}$	Sum of the volume of voxels of ROI
shapeSVR	$\frac{A}{V}$	ROI’s surface on its volume. A small ratio is linked to a simple shape such as a sphere

gldmDNUN: gldm Dependence Non Uniformity Normalized; gldmDLGLE: gldm Large Dependence Low Grey Level Emphasis; glszmGLNUN: glszm Grey Level Non Uniformity Normalized; glszmSAE: glszm Small Area Emphasis; shapeSVR: shape Surface Volume Ratio; ROI: Region of Interest. Bold: highlight labels of lines.