Slicer can read patient name with various encodings:
There is a good chance that there is something wrong with the DICOM file encoding, but of course it is possible that the DICOM file is valid and DCMTK toolkit does not interpret it correctly. We need an example file to verify. You can create an example data set by creating a patient in the dental imaging software (with non-Latin characters in the patient name) and acquire an image of anything.