Skip to main content
. 2021 Oct 29;10(21):5064. doi: 10.3390/jcm10215064

Table 4.

Univariate analysis for all PET and RFs and each scanner considered alone or in combination.

Mean AUC
Scanner 1 Scanner 2 Scanner 1 + 2
“Conventional” PET features
SUV-related
SUVmax 0.855 0.714 0.760
SUVmean 0.840 0.714 0.752
SUVlbm 0.851 0.730 0.767
SUVbsa 0.859 0.724 0.769
L-L SUV R 0.847 0.727 0.766
L-BP SUV R 0.860 0.727 0.776
Metabolic volumes
MTV 0.594 0.601 0.562
TLG 0.771 0.669 0.633
First-order statistics
Histogram-related
Histo skewness 0.682 0.585 0.629
Histo kurtosis 0.584 0.591 0.560
Histo excess kurtosis 0.584 0.592 0.560
Histo entropy_log10 0.869 0.718 0.763
Histo entropy_log2 0.869 0.719 0.762
Histo energy 0.852 0.704 0.743
Shape-related
Shape volume_mL 0.656 0.626 0.542
Shape volume_vx 0.656 0.626 0.542
Shape sphericity 0.615 0.641 0.580
Shape Compacity 0.637 0.634 0.530
Second-order statistics
GLCM-related
GLCM homogeneity 0.846 0.692 0.733
GLCM energy 0.846 0.702 0.743
GLCM contrast 0.849 0.705 0.740
GLCM correlation 0.572 0.677 0.556
GLCM entropy_log10 0.853 0.732 0.747
GLCM entropy_log2 0.850 0.734 0.747
GLCM dissimilarity 0.845 0.702 0.739
GLRLM-related
GLRLM SRE 0.843 0.692 0.731
GLRLM LRE 0.841 0.676 0.723
GLRLM LGRE 0.813 0.688 0.733
GLRLM HGRE 0.819 0.695 0.725
GLRLM SRLGE 0.799 0.687 0.729
GLRLM SRHGE 0.828 0.698 0.730
GLRLM LRLGE 0.831 0.692 0.741
GLRLM LRHGE 0.762 0.678 0.686
GLRLM GLNU 0.752 0.752 0.663
GLRLM RLNU 0.719 0.705 0.642
GLRLM RP 0.483 0.680 0.722
NGLDM-related
NGLDM coarseness 0.643 0.783 0.590
NGLDM contrast 0.808 0.808 0.721
NGLDM busyness 0.783 0.643 0.700
GLZLM-related
GLZLM SZE 0.777 0.677 0.703
GLZLM LZE 0.821 0.821 0.704
GLZLM LGZE 0.785 0.785 0.722
GLZLM HGZE 0.793 0.689 0.722
GLZLM SZLGE 0.614 0.614 0.574
GLZLM SZHGE 0.794 0.689 0.718
GLZLM LZLGE 0.837 0.697 0.725
GLZLM LZHGE 0.755 0.671 0.664
GLZLM GLNU 0.810 0.810 0.715
GLZLM ZLNU 0.808 0.716 0.726
GLZLM ZP 0.834 0.834 0.714