How, exactly, does a run-of-the-mill rescue dog render a diagnosis of cancer or other human illness in a split second, outshining seasoned human specialists, million-dollar machines, and complicated testing sequences? It’s really pretty amazing, but not a total mystery. Living things and their various components (organs, urine, respiratory tissues, for example) release molecules into the air called volatile organic compounds, or VOC’s. The mixture of molecules released from each material functions like an airborne fingerprint. As it turns out, that fingerprint can be used to determine whether the patient (or even a tiny sample of his tissue or blood) is healthy or diseased. We humans can not detect them, obviously, and—let’s be honest—we never would have had a clue except for the actions of one random border collie mix who let us in on the secret 30 years ago.

It’s obvious to the dogs because of their noses. These things look pretty unassuming, but they’re densely packed with 50 or 60 times as many scent receptors as your nose. Each and every receptor is way more sophisticated than anything you’re packing, and that’s not all. Dogs have an intricate latticework of sensor-coated surfaces inside their nose to break every inhaled scent into its separate components for analysis. They can huff tiny volumes of air into their noses 5 times a second to wring every possible piece of information out of thin air. They even have a lateralized communication system which allows them to determine which nostril detected a particular odor.

With such advanced equipment, why haven’t dogs taken over the medical field? They diagnose Parkinson’s, cancer, and COVID much more quickly than a human or a machine and will do practically anything for the low, low price of a Scooby Snack. What could go wrong?

Well, first off, these dogs cannot be mass-produced because they need 6-12 months of intensive work with a highly skilled trainer, and because the tissue samples—or “study guide”—needed for their education are pretty hard to come by. Upon graduation from the program, each dog can identify a single specific condition. To screen a patient for twenty common conditions, you’d need… twenty different dogs. Not every dog has a good enough nose or the right personality for the job. They don’t respond to paychecks, promotions, or fame; they just want a biscuit after doing a good job. It’s extremely difficult, however, to determine when they’ve done a good job when we ourselves don’t possess the answer key to the tests they’re performing each day.

None of that matters, though! These dogs have already advanced medicine light years ahead by showing us that tissue abnormalities do emit unique compounds that can be harvested from the air. We have the technology to collect those particles and name them; we just need a few patient souls to teach our machines which ones go with which illness. That collaboration between our dogs and our devices will propel the diagnosis of disease into a new era.

Dr. M.S. Regan