Person names in many languages may be written in a local (non-Latin) script, as well as in a transliteration to a Latin script (Romanization). Healthcare information systems in those environments may support one or both name formats. Local scripts may be encoded using Unicode in UTF-8.
For the purpose of exchange in DICOM, there are three typical uses of name component groups using Unicode in UTF-8:
Names in a Latin script may be encoded in the first (alphabetic) component group, and names in a local script (alphabet, abugida, or syllabary) in the third (phonetic) component group (see Table 6.2-1). The second (ideographic) component group is null. This is the preferred use for cross-enterprise or international communication.
Where the local script historically has a single byte character set defined for Specific Character Set (0008,0005), i.e., Cyrillic, Arabic, Greek, Hebrew, Thai, and the various versions of Latin, only the first name component group might be used. Encoding may be in Unicode in UTF-8, as described in this Annex, as an equivalent for use of that defined single byte character set in the first name component group. (See note 1.)
Names in the local script may be encoded in the first component group, and names in a Latin script in the third component group, both encoded in Unicode in UTF-8.
Notes: 1. A previous edition of DICOM required the first name component group to use a single byte character set (see PS3.5-2008). Unicode in UTF-8 may now be used in that component group simply as a matter of a different character set encoding, but with the same application use of that component group.
2. Healthcare information systems will use specific scripts in one, two, or three of the Person Name component groups in accordance with local policy. Conformant DICOM Application Entities that receive name attributes must accept multiple name component groups. An Application Entity that is configurable to allow the use of local script for names in either the first or the third component group, and a transliteration script in the other, would support all these typical representations.
3. The transliteration (from a local script) may be a non-Latin script, e.g., Cyrillic. The same principles apply, and the Cyrillized name might be encoded in the first component group and the local script (which may in fact be a Latin-derived script) in the third component group.