I took some time to think about your question proposed here. There may be a logic that satisfied both points of concern....
If we stick to the same logic as lets say CD releases, we rely on the branding, or what we are told a release is. Meaning, that 2 CDs with a different barcode on the artwork counts as 2 different releases, even though the CD inside may be identical. We do this by not assuming anything and looking at what we are presented to make a determination.
What if we do the same for digital? In your example, the KISS thing, we do not consider it a remaster, because it is not called a remaster. But, we do consider it "mastered for iTunes" because it was labeled as such (like a barcode or a packaging label). There is further support for this as often the artwork will also include mastered for iTunes in some form on it, and that is packaging. In this method, we look at the facts presented to us, but do not make assumptions. What are your thoughts?
All of the KISS specific details outlined would be irrelevant to the decision, the decision is based solely on the fact that iTunes tells us the release is different by the packaging. Leaving out all other arguments, like MP3 release vs M4A release, mastered for iTunes would be considered a different release on its own merit, regardless of anything else. It still leaves the digital releases a mess, but it does address the specific proposal here on how to address the mastered for iTunes version.
To specifically share on your #1 or #2 question, I would think it is better off being a type of medium. Meaning, on the list we only have "Digital Media" as an option. But under vinyl, cd, etc we have a good list of types. Digital media also has types, just like CD vs enhanced CD. We have "Mastered for iTunes" which means the master was not a downsampled to 16 bit for CD production. Further (different discussion though), we would have the 16 bits of M4A, MP3, FLAC, etc. Those file types are containers, kind of like a cassette to CD, it is delivered on/in a different medium.