I know the Index and LoS will not have titles according to your DTD, but the way that Arbortext works is that it will generate temporary elements during publishing. These temporary elements will include the Index and LoS (including the titling that appears on the page - presumably that is the same text you need in the TOC).
Another way to look at it is that you start with the actual document then, during publishing, Arbortext creates an enriched version of your document which it will then PDF. My suggestion was to grab that enriched content for your TOC, after the Index and LoS have been expanded and generated, but before PDF. It is likely the Index and LoS are generated as UFEs, in case that helps you.
It is hard to provide precise advice without seeing the gentext code you have, but I don't really have time to spend on the code anyway. If you need more help or training then our company GPSL (www.gpsl.co) would be happy to assist.