Skip to main content
. 2021 Oct 24;19:443. doi: 10.1186/s12967-021-03117-5

Table 2.

Diagnostic performance of ME_ADC0–1000, BE_IVIM_D, ME_ADCall b, DKI-D and DKI-K by using RF, L1R-LR, PCA-LR, and SVM, respectively

Maps AUC (95% CI)
RF L1R-LR PCA-LR SVM mAUC
ME-ADC0–1000 0.83 (0.80–0.87) 0.83 (0.79–0.87) 0.76 (0.70–0.81) 0.81 (0.76–0.85) 0.81
BE-IVIM-D 0.85 (0.81–0.89) 0.83 (0.78–0.87) 0.75 (0.70–0.82) 0.80 (0.75–0.85) 0.81
ME-ADCall b 0.84 (0.80–0.87) 0.82 (0.79–0.87) 0.77 (0.70–0.83) 0.79 (0.74–0.85) 0.81
DKI-D 0.83 (0.80–0.86) 0.83 (0.78–0.86) 0.75 (0.74–0.82) 0.80 (0.77–0.84) 0.80
DKI-K 0.84 (0.81–0.89) 0.83 (0.78–0.87) 0.74 (0.70–0.80) 0.79 (0.75–0.85) 0.80

RF: random forest; SVM: support vector machine; PCA: principal component analysis; L1R: L1 regularization; LR: linear regression; mAUC: mean values of AUCs of RF, L1R-LR, PCA-LR and SVM