Are Medical Detection Dogs Winning the Diagnosis Game?

Your dog is currently lying on the sofa, dreaming about squirrels and whether today might finally be the day someone drops a sausage. Meanwhile, his cousins — the highly trained, four-legged medical professionals working out of the UK charity Medical Detection Dogs — are busy sniffing out cancer, Parkinson’s disease, and bacterial infections with a level of accuracy that would make some diagnostic labs genuinely nervous.

This is not science fiction. It is not a particularly elaborate metaphor. Dogs are, right now, being trained to detect diseases in human beings, and the results are making scientists do a very satisfying double-take.

Here is the thing that makes dogs so extraordinary at this: their noses. A dog’s sense of smell is at least 10,000 times stronger than ours, and inside that nose are hundreds of millions of scent receptors, along with a highly developed olfactory bulb that processes chemical signals with exceptional precision. When your body develops a disease, it changes the cocktail of chemicals you release — those volatile organic compounds, or VOCs, that drift invisibly off your breath, your sweat, your urine. To us, these changes are completely imperceptible. To a well-trained Labrador, they are essentially a neon sign.

So what can they actually detect? Quite a lot, as it turns out. Dogs have demonstrated the ability to identify toxigenic Clostridium difficile in stool, lung and breast cancers in breath, four different bacteria causing urinary tract infections, malaria through patient clothing, prostate cancer in urine, ovarian cancer in blood, type 1 diabetes, and Parkinson’s disease through sebum. That last one is particularly extraordinary. By the time Parkinson’s is typically diagnosed, significant brain cell damage has already occurred. Researchers have discovered, however, that the disease may have a distinct odour — one that trained dogs can detect years before clinical symptoms appear.

Cancer detection is where the headlines tend to cluster, and understandably so. In some cases, such as prostate cancer, dogs achieved a 99% success rate detecting the disease from urine samples. A 2024 double-blind study published in Scientific Reports went further still, pairing trained detection dogs with artificial intelligence to screen for four cancer types — breast, lung, prostate, and colorectal — via breath samples. The bio-hybrid platform demonstrated overall sensitivity and specificity of 93.9% and 94.3% respectively. That kind of accuracy, achieved non-invasively, by breathing into a tube, is the sort of thing that makes oncologists sit up very straight.

Then there is the COVID-19 chapter, which accelerated everything. Studies confirmed that trained dogs can detect SARS-CoV-2 infections from sweat samples with sensitivity and specificity comparable to rapid antigen tests — but much faster. At transportation hubs and large events, detection dogs can screen hundreds of people per hour. When researchers tested this at actual concerts, 70% of participants preferred canine testing before the event, and 72% still preferred it afterwards — which says something rather charming about both people’s trust in dogs and their enthusiasm for avoiding a cotton swab up the nose.

Of course, none of this is quite as straightforward as “dog sniffs man, man is cured.” There are real complications. Results in published cancer studies have varied significantly, with sensitivities ranging from 19% to 99% depending on cancer type, sample material, training methodology, and the dog itself. Diet, medication, inflammation, and environmental pollution can all affect VOC profiles, potentially muddying the results. Standardisation remains a genuine challenge — you cannot exactly roll out the same quality control protocol you would for a blood test when one of your diagnostic instruments is having a particularly distracted day because someone nearby has cheese.

Trainer bias is also a known issue. Dogs are extraordinarily good at reading human cues — it is, frankly, one of the things that makes them such remarkable companions — but in a diagnostic setting, that sensitivity can become a liability. If the handler unconsciously signals which sample is positive, the dog may follow that cue rather than the scent. Rigorous double-blind methodology is therefore essential, and not all early studies applied it.

There is also the question of personality. New research from the University of Bristol, published in April 2025, found that dogs displaying more “optimistic” responses in judgment bias tests tend to perform better overall in detection tasks, while more “pessimistic” dogs achieve higher specificity — meaning they produce fewer false positives. Essentially, the canine equivalent of the cautious, methodical colleague who never files anything prematurely turns out to be extremely valuable in a diagnostic setting. Who knew pessimism would be a professional asset.

The regulatory path remains another hurdle. Before any dog-based diagnostic tool can be adopted into mainstream medicine, it would need approval from national health authorities — and rightly so. A medical device or health technology requires approval by national health organisations before permission for usage is granted. This is not cynicism about the science; it is the entirely reasonable demand that anything used to tell someone they might have cancer be held to an exceptionally high standard.

What makes this genuinely exciting, rather than just a charming curiosity, is where it is all heading. Researchers at MIT have been working to miniaturise the principles behind canine detection into a compact device. They have developed a system that can detect the chemical and microbial content of an air sample with even greater sensitivity than a dog’s nose, coupled to machine learning that identifies the distinctive characteristics of disease-bearing samples. The long-term vision involves building something like this into a smartphone. A sniff-based cancer screen, in your pocket, powered by everything we have learned from watching dogs work.

In the meantime, the dogs themselves continue doing what they do best — working with focused, joyful precision, motivated by treats and the approval of people who adore them. They cannot write up their findings or apply for a grant. They do not attend conferences or publish in journals. What they do, with a reliability that keeps surprising the scientific community, is smell things that we cannot — and then sit, very pointedly, until someone pays attention.

Given that early detection remains one of the most powerful factors in cancer survival, it turns out that paying attention to a sitting Spaniel might be one of the more medically significant things you can do.