Feature extraction from MRI slices

Hello Pyradiomics community,
I am a little bit confused because I am new in radiomics.
I have 3 MRI slices with 3 corresponding masks, and I want to know if I should extract radiomic features from each slice and calculate the mean ? or it is okay to have features for each slice?

I’m eagerly awaiting your response. Many thanks.

Normally you use a 2D region in 2D images and 3D region in 3D images. Are the 3 slices coming from a single volume? If yes, then it would be probably more appropriate to compute features in a 3D region.

Thank you very much @lassoan for your reply !