antspynet.deep_atropos() usage for subjects containing tumors

[muqeemmm]

I am interested in segmenting white matter (WM) tissue from a T1 brain sequence image containing tumor. I am sharing a code snippet as follows:
`# 0) Load native-space T1
t1_native = ants.image_read(str(t1_pth))
tumor_mask_native = ants.image_read(str(tumor_pth))
out_dir = output_dir

# 1) Run Deep Atropos WITH built-in preprocessing (works in template/MNI space)
da          = antspynet.deep_atropos(t1_native, 
                                     do_preprocessing    = True, 
                                     verbose             = True, 
                                     use_spatial_priors  = 0)
seg_mni     = da["segmentation_image"]                                            # label image (template/MNI space)

seg_probs   = da["probability_images"]                                            # list of 6 tissue probability images (template/MNI space)
wm_prob     = seg_probs[2]                                                        # white matter probability map`

My main concern is with presence of tumor itself. Is the current implementation robust enough to handle such cases?

[ntustison]

You'd have to try it and see.