Table 2 Shows the overall performance metrics of the EMLM when trained using the Weka selected DTI features of the tumorous region, NAWM region, and tumorous + NAWM region separately for the standard DTI model.

From: Accurate low and high grade glioma classification using free water eliminated diffusion tensor metrics and ensemble machine learning

Region of interest

Weka attribute selection

Class

Sensitivity

Specificity

AUC-ROC score

Tumorous region

FA of the core and edema, AD of the enhanced

LGG

0.63

0.61

0.60

HGG

0.66

0.61

0.65

NAWM

FA of the NAWM

LGG

0.70

0.61

0.66

HGG

0.68

0.61

0.65

Tumorous + NAWM region

FA of the core and edema, AD of the enhanced

FA of the NAWM

LGG

0.67

0.68

0.68

HGG

0.64

0.64

0.64

Back to article page