Table A4.
Top 10 selected features according to rank products for each dataset.
| Dataset | cMRI | PWI | DKI | cPD |
|---|---|---|---|---|
| CER&NER | T1pc-average-CER | CBV-average-CER | ak-IMC2-CER | T1pc-average-CER |
| T1pc-90th_percentile-CER | MTT-homogeneity-CER | ak-IMC1-CER | T1pc-90th_percentile-CER | |
| T1pc-10th_percentile-CER | CBV-90th_percentile-CER | fa-autocorrelation-CER | CBV-90th_percentile-CER | |
| FLAIR-cluster shade-CER | MTT-IDN-CER | fa-sum variance-CER | CBV-average-CER | |
| FLAIR-skewness-CER | MTT-maximum probability-CER | fa-sum of squares-CER | CBF-90th_percentile-CER | |
| FLAIR-sum of squares:variance-CER | CBF-90th_percentile-CER | fa-average-CER | ak-IMC2-CER | |
| T2-cluster shade-CER | MTT-dissimilarity-CER | fa-sum average-CER | MTT-homogeneity-CER | |
| FLAIR-autocorrelation-CER | Rsquare-homogeneity-CER | ad-cluster prominence-NER | CBF-average-CER | |
| FLAIR-sum variance-CER | MTT-difference entropy-CER | mk-IMC2-CER | MTT-IDN-CER | |
| FLAIR-sum average-CER | CBF-average-CER | mk-IMC1-CER | ak-IMC1-CER | |
| Total ROI | T1pc-90th_percentile | CBV-90th_percentile | ad-cluster prominence | ad-cluster prominence |
| T1pc-coefficient of variation | Rsquare-IMC2 | md-cluster prominence | md-cluster prominence | |
| T1pc-cluster shade | Rsquare-IMC1 | ad-cluster shade | T1pc-90th_percentile | |
| T1pc-sum average | CBF-90th_percentile | rd-sum entropy | T1pc-coefficient of variation | |
| T1pc-skewness | CBV-IMC2 | md-cluster shade | CBV-90th_percentile | |
| T1pc-sum variance | MTT-correlation | md-sum entropy | ad-cluster shade | |
| T1pc-autocorrelation | CBF-IMC2 | rd-cluster prominence | T1pc-skewness | |
| T1pc-sum of squares:variance | Rsquare-contrast | rd-cluster shade | rd-sum entropy | |
| T1pc-maximum probability | Rsquare-dissimilarity | ak-cluster prominence | T1pc-sum average | |
| T1pc-average | Rsquare-IDMN | rk-sum entropy | T1pc-cluster shade |