Warping lesion ROIs and HCPex Atlas ROIs onto patient specific diffusion space

As described in this paper, I’d like to warp MRtrix generated lesion ROIs from one patient with low quality DTI to another patient with high quality DTI (human connectome project data).

What is the best way to accurately warp one patient’s lesion ROIs into another patient specific diffusion space?

I would also really appreciate any guidance on how to use flirt, fnirt, or mrregister to warp ROIs from atlases (like the HCPex Atlas which is in ICBM152 2009c Asymmetric brain template space) into patient specific space.