With all those studies marking the ability of dogs to identify very specific diseases I'm curious as to why I've never heard of a causal biochemical mechanism b

With all those studies marking the ability of dogs to identify very specific diseases I'm curious as to why I've never heard of a causal biochemical mechanism being investigated. What are they really smelling? I find it hard to believe that they can differentiate between two different respiratory viruses purely by some kind of trace metabolite ratio or by sensing protein structure. Perhaps there's a flaw in methodology or a correlational relationship or bias in the sampling. It would be very impressive if it's true and it definitely warrants a lot of funding to finally determine how this works.