Breeds originally selected for scent work, such as Beagles, Labrador Retrievers, German Pointers, and Bloodhounds, are the dogs most commonly trained to detect cancer. But the truth is more nuanced than a single breed list: any dog with a long snout and proper training can potentially learn to identify cancer samples, and the training method matters at least as much as the breed itself.
Breeds With the Strongest Scent Ability
A study published in PLOS One tested olfactory performance across different breed groups and found that dogs originally bred for scent work significantly outperformed other groups. These “scent breeds” included Beagles, Basset Hounds, German Pointers, Wire-Haired Vizslas, and several lesser-known breeds like the Bracco Italiano and Transylvanian Hound. Not only were they more successful overall, they were the only group to perform above chance at the hardest difficulty level.
Labrador Retrievers and German Shepherds also appear frequently in cancer detection research, largely because they’re highly trainable and already widely used in professional scent work like bomb and drug detection. Labradors in particular have featured in studies on prostate, lung, and bladder cancer detection.
On the other end of the spectrum, short-nosed (brachycephalic) breeds like Pugs and Bulldogs performed worst. Breeding for flat faces has come at a cost to olfactory anatomy, reducing airflow and the surface area available for scent processing. If you’ve ever wondered whether a Pug could do the job of a Bloodhound, the answer is a clear no.
What Dogs Are Actually Smelling
Cancer changes your body’s metabolism in ways that produce specific chemical byproducts called volatile organic compounds, or VOCs. These compounds evaporate into your breath, seep into your urine, and escape through your skin. Dogs aren’t detecting “cancer” as a single smell. They’re recognizing a distinct chemical fingerprint made up of multiple VOCs that healthy people don’t produce in the same pattern.
Research into prostate cancer, for example, has identified several categories of these compounds in urine: certain ketones produced when tumors ramp up fatty acid breakdown to fuel their growth, terpenes linked to a metabolic pathway that cancer cells hijack, and aromatic compounds that appear at elevated levels. Specific molecules like pentanal, 3-octanone, and dimethyl disulfide have all been flagged as potential prostate cancer markers. Dogs appear to detect the overall pattern rather than any single molecule, which is part of what makes their ability so difficult to replicate with technology.
Which Cancers Can Dogs Detect?
Peer-reviewed studies have demonstrated that trained dogs can identify at least seven types of cancer: lung, breast, prostate, ovarian, bladder, melanoma, and colorectal. Cervical cancer and a form of bladder cancer in dogs themselves (transitional cell carcinoma) have also been studied. The sample types vary. Dogs have been trained to sniff exhaled breath, urine, skin lesions, and even thawed frozen tumor tissue.
The accuracy varies dramatically depending on the cancer type, the sample used, and the training approach. Lung cancer detection using exhaled breath has produced some of the most impressive numbers: one early study reported 99% sensitivity and 99% specificity, meaning the dogs correctly identified nearly every cancer sample and almost never flagged a healthy sample as cancerous. A later study using breath-based training found 91.7% sensitivity and 85.1% specificity.
Prostate cancer results have also been strong. In one study published in European Urology, a trained dog correctly identified cancer-positive urine samples in 30 out of 33 cases, reaching 91% sensitivity and 91% specificity. One of the three “misses” turned out to be a correct call: the patient was later rebiopsied and did have prostate cancer.
Other cancers have been harder to detect reliably. Bladder cancer detection from urine hit only 41% accuracy in one study, and breast cancer detection via urine was just 22% sensitive, though the same cancer reached 88% sensitivity when dogs sniffed exhaled breath instead. Ovarian cancer detection using thawed tumor samples reached 100% sensitivity and 97.5% specificity, but that study used a sample type that wouldn’t be practical for screening.
Why Training Matters as Much as Breed
The same dog can perform brilliantly or barely above random chance depending on how it was trained and what samples it practiced with. A lung cancer study illustrates this perfectly: dogs trained on exhaled breath samples achieved 91.7% sensitivity, while dogs trained on lung cancer tissue samples and then asked to evaluate breath hit only 50.4%, essentially a coin flip. The training sample needs to match what the dog will encounter in practice.
Handler influence is another variable. If the person directing the dog unconsciously signals which sample is positive (through body language, hesitation, or gaze), the dog may respond to the handler rather than the scent. The most rigorous studies use double-blind designs where neither the dog handler nor anyone in the room knows which sample contains cancer.
Why Dogs Aren’t Used in Clinics Yet
Despite the impressive accuracy numbers, no country has approved dog-based cancer screening for clinical use. The barriers are practical more than scientific. Dogs have off days. They get tired, distracted, or bored. Their performance can shift between sessions, and there’s no way to standardize a biological detector the way you can calibrate a machine. Training a single cancer-detection dog takes months, and each dog can only process a limited number of samples per day before fatigue sets in.
Reproducibility is also a challenge. A dog that performs at 91% accuracy in one lab may not hit the same numbers in a different environment with different handlers and a different patient population. These aren’t flaws in the dogs so much as realities of working with living, variable detection instruments.
The more lasting contribution of canine cancer detection may be in guiding technology. Researchers are developing “electronic nose” devices that use chemical sensors to detect the same VOCs dogs respond to. These devices aim to match canine accuracy with the consistency and scalability of a machine. The dog research has been essential for proving that cancer produces a detectable scent signature and for narrowing down which compounds matter most. In that sense, Beagles and Labradors have already made a significant contribution to cancer diagnostics, even if they never set paw in a screening clinic.