Table 1.

List of the features selected and included in the radiomic signature. Mean values of the coefficients of the Elastic Net logistic regression (averaged over 500 model instances).

Feature group	Feature name	Description	Coefficient
Shape	Maximum 2D diameter (column)	The largest pairwise Euclidean distance between tumor surface mesh vertices in the coronal plane	–0.1815
	Least Axis Length	The smallest axis length of the ROI-enclosing ellipsoid	–0.241
	Elongation	Describes the relationship between the two largest principal components in the ROI shape	0.379
GLCM (gray-level co-occurrence matrix)	Inverse Difference Moment Normalized (IDMN), also denoted as homogeneity	Is a measure of the local homogeneity of an image (IDMN weights are the inverse of the contrast weights (decreasing exponentially from the diagonal i=j in the GLCM). It measures the smoothness (homogeneity) of the gray-level distribution of the image; it is (approximately) inversely correlated with contrast—if contrast is small, usually homogeneity or IDMN is large (39)	–0.7857
	Difference Entropy	A measure of the randomness/variability in neighborhood intensity value differences; measures the degree of disorder related to the gray-level difference distribution of the image. Entropy is (approximately) inversely correlated with uniformity; images with a larger number of gray levels have larger entropy (39)	–0.4252
Volume			–0.0458

The response variables were codified as 1 = CRS3, 0 = CRS1–2.