Diagnostics is getting revolutionised by research on biochemical markers of diseases that carry their characteristic smells
Animals “sniffing out” diseases much before symptoms manifest made us think if diseases have biochemical markers that make our body smell a certain way. But how has it changed diagnostics? Without doubt, early diagnosis can make a big difference in certain cases, especially where chances of successful treatment diminish with progression of the disease. Joy Milne, a Scottish nurse, made news because she could smell volatile organic compounds off those suffering from Parkinson’s. In one case, there were no visible signs of the disease—indeed, early diagnosis has traditionally been thought to be near impossible as symptoms at these stages mirror those of other brain disorders—Milne sniffed a person out who was clinically diagnosed only many months after.
There are a fair number of diseases—various cancers, diabetes, cirrhosis, TB, etc—for which smell-based diagnosis is being studied. ‘Smell-based’ here doesn’t specifically mean an olfactory test; it also includes various assays of the specific chemicals produced by a person that give a disease its “characteristic” smell. Of course, Milne-like olfactory capabilities must be quite rare; else we would have heard of many such sniffer men and women. So, the dependence has to be on translating the smell tests into a format intelligible for, say, artificial intelligence.
In January, Chinese researchers published a paper in ACS Omega that shows the potential of biochemical tagging. Smell-based Parkinson’ diagnosis is possible because of compounds typical of the disease contained in sebum. The Chinese researchers used fast gas chromatography-mass spectrometry. In layman terms, signature patterns were generated from sound waves interacting with gases emanating from sebum. Given each compound influences the waves uniquely, the researchers taught an AI unit to recognised interference patterns that are a tell-tale sign of the Parkinson’s compounds. The result? A small (read portable) machine studying 43 patients and 44 control participants could recognise Parkinson’s in a patient 70% of the time and identify healthy subjects as such four times out of five.
Korean scientists meanwhile have devised a way of using a certain species of worms and chips together to detect lung cancer (shouldn’t surprise that the worm prefers odours emitted by cancerous cells in the lungs), with 70% accuracy. Diagnostics will enter a brave new world with the olfactory x factor that many diseases carry.