Don't worry. I think it's important to realize the only place AFNI uses the oblique information is in 3dWarp. In all other places, the cardinal information is used. Datasets are displayed on their cardinal equivalents always. The display is based on the origin and voxel size. If the dataset is created with to3d, the origin and voxel size are computed independently of the oblique matrix from DICOM data, relying on the manufacturer to be consistent, though there is an option to override this. Matching obliqueness between the two datasets makes the cardinal matrices of each match too. You can read a little more about obliquity on this page:
[
afni.nimh.nih.gov]
[
afni.nimh.nih.gov]