Table 1.

Overview of univariate and multivariate analysis shown for all region of interests (ROIs) considered.

		Univariate		Multivariate
ROI	Features	AUC	p-value	Coefficient	p-value
GTV	GLRLM_run entropy	0.64	0.034
	GLCM_inverse variance*	0.65	0.035	−10.83	0.035
lung_exterior	GLSZM_zone percentage	0.63	0.046
	GLCM_contrast	0.65	0.043
	GLCM_homogeneity normalized*	0.72	0.004	52.51	0.004
	NGLDM_low dependence emphasis	0.66	0.025
iso_exterior	Intensity_median	0.65	0.028
	GLCM_correlation	0.61	0.043
	GLSZM_zone size non-uniformity normalized*	0.68	0.015	−12.84	0.112
	Intensity_percentile_90*	0.68	0.025	−0.01	0.072
gradient	GLSZM_zone size non-uniformity normalized*	0.68	0.046	−20.26	0.046
GTV+Rim	Intensity_median*	0.63	0.026	0.002	0.144
	GLSZM_zone size non-uniformity normalized*	0.69	0.010	−16.75	0.026

Only features are listed which had a significant performance in the univariate analysis per principal component group. Features with an asterisk were retained in the final models after backward selection and their coefficients and p-values in the multivariate analysis are listed.