Fig. 4.
Assessment of the objective combination of PET and MRI variables for grade prediction. The least absolute shrinkage and selection operator (LASSO) was applied to extract a composite vector for tumour grade prediction from 24 (dynamic included) or 20 (static only) parameters determined by analysis of simultaneously acquired PET and MRI data. Five-fold cross-validation was performed to select lambda minimum to give the minimum cross-validated error for classifying LGG (grade II) versus HGG (grades III and IV) when (A) the PET variables were restricted to static data or (B) included dynamic PET data. The feature coefficients (b value) are indicated. (C) Boxplots showing ve parameter in grades II, III, IV and CWM