I’ve seen several attempts at optical monitors. The first I became aware of was Agamatrix in early 2000s.

I doubt this is possible without some kind of catalyst. There’s just nothing all that unique about the composition or phonon modes of sugar or insulin.

There are other fluids aside from blood though. I think those have promise. Google X was working on a CGM in a contact lens. Maybe somebody is working on a CGM in dental roots. Gastric sensors may be just as good.