So the original "sniffer dog" study (The Lancet 8640)[0] pretty much covers a dog sniffing out a melanoma BEFORE it got to a later stage, and saved a woman's life.
I also dont understand how training a machine would have selectivity issues. Machines dont "get trained" on one thing.
That is not a study. That is a case report written as a letter to the editor. That meets no scientific rigor at all and no generalizable conclusions can be drawn from it, it provides an anecdote for further research. It also doesn’t suggest that sniff tests would fare better compared to any other screening system for melanoma.
I also dont understand how training a machine would have selectivity issues. Machines dont "get trained" on one thing.
0: https://www.sciencedirect.com/science/article/pii/S014067368...