Classification of features radiomics

Dear PyRadiomics / 3DSlicer community,

I have been extracting radiomic features with PyRadiomics.
But, when I try to carry out a classification using machine learning techniques, usually the different classifiers classify one class very well and the other does not. I have already tried some pre-processing techniques, such as normalization, scaling, dimensionality reduction, class balancing, feature selection, but without better results.
I am using 3 different datasets, and in both results it has the same characteristic.
The radiomics features have any particular characteristics that need to be addressed?
Thank you for your help.
Tony