Researchers have trained dogs to detect certain cancers from breath and urine samples, sometimes with astonishing accuracy. Other experiments have produced results no better than chance. Both findings matter, because the question isn’t simply whether dogs can smell illness, but what they’re detecting and how reliably they can do it.
Perhaps you’ve noticed your dog becoming unusually interested in someone who isn’t feeling well. He might repeatedly sniff their hands or linger near their clothing. When that person later becomes sick, it’s natural to wonder whether your dog noticed something first.
The impressive part isn’t that dogs somehow sense illness. It’s that illness can change the chemistry available to smell. Scientists are discovering how much information those changes contain, while learning why turning a dog’s remarkable nose into a dependable medical test is so difficult.
How Illness Changes the Way Someone Smells
Your body constantly produces chemical compounds as cells use energy, tissues repair themselves, and microorganisms interact with your skin. Some leave through your breath, sweat, urine, and other biological material.
Illness can alter those processes. Infections may change microbial activity and trigger inflammation. Cancer can affect cellular metabolism, potentially changing the compounds released by affected tissues.
Many substances researchers investigate are volatile organic compounds, or VOCs. These chemicals can evaporate into the surrounding air, making them available to a sensitive nose.
But diseases don’t necessarily produce unique chemical fingerprints. Similar compounds can appear in healthy people or those with unrelated illnesses. Diet, medication, and age also influence body odor. Researchers therefore investigate patterns of chemical differences rather than searching for one unmistakable disease molecule.
Where Dogs May Detect Illness-Related Odors
We rarely consider how much chemical information leaves our bodies throughout the day. To a dog, even familiar sources of human odor offer plenty to investigate.
Where Disease-Related Odor Changes May Appear
| Sample source | Potential chemical information | Detection-dog research examples |
|---|---|---|
| Breath | Volatile compounds influenced by metabolism and respiratory processes | Lung and breast cancer studies |
| Skin and sweat | Compounds shaped by metabolism and skin microorganisms | COVID-19, malaria, and Parkinson’s research |
| Urine | Metabolic waste products and other compounds | Bladder and prostate cancer studies, with mixed results |
Consider how naturally your dog sniffs your breath. He might be curious about food or unfamiliar odors. That everyday behavior isn’t medical screening, but it illustrates why breath can be an interesting source of chemical information.
Why Dogs Can Detect Differences We Miss
Dogs have specialized nasal structures that direct odor-bearing air toward sensory tissue. Their olfactory receptors respond to many different chemical compounds, giving them access to information people may never notice. That’s a major reason why a dog’s sense of smell so remarkable. But medical detection requires more than just sensitivity. Dogs must learn which differences matter and recognize them across unfamiliar samples.
Similar principles appear in research on whether dogs can smell human stress. Trained dogs have distinguished certain stress-related odor samples from baseline samples, demonstrating how physiological changes can become detectable through scent.
How Medical Detection Dogs Learn
Picture a trained dog investigating several identical containers holding carefully prepared breath samples. He pauses at one and performs his trained indication. During training, a correct response earns a reward.
With practice, dogs learn to distinguish target samples from controls. The challenge is proving they’ve learned the intended disease-related difference rather than an unrelated clue.
Suppose every cancer sample comes from one hospital and every healthy sample from another. The dog might learn differences in collection materials or storage instead of cancer itself. If those same differences remain in the test samples, he could pass the test with impressive accuracy while detecting the wrong thing.
Good studies therefore require safeguards:
- Unfamiliar samples: Dogs must recognize targets they haven’t encountered during training.
- Suitable controls: Testing should account for unrelated illnesses and other odor influences.
- Consistent collection: Containers and handling shouldn’t reveal which samples are positive.
- Blinded testing: Handlers shouldn’t unknowingly guide dogs toward correct answers.
These precautions aren’t a criticism of dogs. They’re necessary because dogs are excellent learners, including when the clues they learn aren’t the ones researchers intended.
What Research Has Demonstrated
In a 2004 bladder-cancer study, six trained dogs correctly identified urine samples 41% of the time, compared with approximately 14% expected by chance. Each trial included one cancer sample among seven choices, so random guessing would succeed about once in seven attempts.
A 2006 study reported much higher performance using breath samples. Five trained dogs achieved approximately 99% sensitivity and specificity for lung cancer, while breast-cancer results were approximately 88% sensitivity and 98% specificity. Those exceptional lung-cancer figures haven’t been consistently reproduced at that level in subsequent research.
The difference doesn’t necessarily mean dogs detect lung cancer better than bladder cancer. The studies involved different diseases, samples, training methods, and testing procedures.
A 2014 prostate-cancer experiment offered an important counterexample. Two dogs that learned to distinguish samples during training failed to perform better than chance when tested with unfamiliar urine samples under double-blind conditions.
The researchers suggested the dogs may have memorized individual training samples rather than learned a generalizable cancer-related odor. That doesn’t prove prostate cancer is undetectable through scent. It demonstrates why unfamiliar samples are essential.
What Newer Research Has Added
Medical scent-detection research now extends beyond cancer. During the COVID-19 pandemic, researchers investigated whether trained dogs could distinguish samples from infected and uninfected people. A 2022 systematic review covering cancer and infectious-disease detection found promising results in some experiments, but substantial differences in methodology and study quality.
Researchers have also investigated malaria detection using odors collected from clothing worn by children. In 2025, a Parkinson’s disease study reported that two trained dogs distinguished skin swabs from affected and unaffected individuals, with sensitivities of 70% and 80% and specificities of 90% and approximately 98%.
These findings broaden the possibilities, but performance still depends on the condition, samples, dogs, and experimental design.
Why a Correct Alert Isn’t Automatically a Diagnosis
Sensitivity describes how often a test correctly identifies someone with a condition. Specificity describes how often it correctly identifies someone without it.
But even impressive accuracy can be misleading when a disease is uncommon.
Imagine screening 1,000 people for a condition affecting just 1% of them. With 90% sensitivity and 90% specificity, the test would correctly identify nine of the ten affected people. But approximately 99 healthy people would also receive positive results.
Only nine of 108 positive results would be correct. This hypothetical example shows why disease prevalence matters. Laboratory experiments often include a much higher proportion of positive samples than routine screening, so impressive sensitivity and specificity don’t directly reveal how many positive alerts would be correct in the general population.
Real patients also have overlapping medical conditions, medications, and chemical changes. Reliable screening requires independent replication and careful clinical validation.
What About Diabetes and Seizure Alert Dogs?
Dogs trained to alert people with diabetes or respond to seizures perform different tasks from laboratory disease-detection dogs.
Some diabetes alert dogs provide useful warnings, but controlled studies have found substantial variation in accuracy. Researchers are still investigating which cues contribute to alerts, including possible scent changes.
Owners sometimes report that their dogs behave differently before seizures, although reliable seizure prediction remains difficult to establish. Neither type of alert should replace prescribed monitoring equipment or medical care.
What If Your Dog Keeps Sniffing Someone Who’s Sick?
At home, your dog isn’t examining standardized samples. He’s encountering the changing scent of someone familiar, along with everything else happening around them.
Perhaps your partner started taking medication, using a different skin cream, or spending more time resting. Your dog might investigate their hands or clothing because something smells unfamiliar. He may also notice changes in movement, routines, and interactions.
Some dogs briefly investigate and move on. Others return repeatedly to an interesting scent. Neither response reveals exactly what they’ve detected.
A dog who acts protective when you’re sick may be responding to altered routines or unusual behavior rather than illness-related odor.
If you’re concerned about someone’s health, seek appropriate medical advice rather than relying on your dog’s sniffing.
What Medical Detection Dogs May Help Us Discover
Dogs may help researchers identify chemical differences associated with disease. When trained dogs repeatedly distinguish samples under rigorous conditions, scientists can investigate what makes those samples different.
That work could eventually contribute to better chemical analysis or detection technologies, although dependable medical screening requires much more validation.
Think back to the dog repeatedly sniffing someone’s hands. He may have noticed a genuine scent change or something entirely ordinary. We can’t know from the behavior alone. What makes medical detection research fascinating is that some of those subtle chemical differences really are detectable, even when their meaning remains difficult to establish.
Behavior Lens
Watch What Happens After the Sniff
Your dog encounters your scent every day through your skin, clothing, bedding, and belongings. When something changes, he may investigate more carefully simply because the smell is unfamiliar. A new medication, different soap, or change in daily activity could attract his attention.
Notice what happens afterward. Does your dog investigate briefly and return to his usual activities? Does he hesitate, move closer, or seem unsettled by something else? Those observations can help you understand his response without assigning a specific meaning to the odor. Sniffing is how dogs explore changes in their surroundings, and the behavior following the sniff often reveals more about how they’re feeling.
Related Behaviors to Explore
- Why Dogs Sniff the Air: What They’re Detecting
- Why Dogs Sniff Dirty Laundry: Why Your Scent Is So Interesting
- Why Dogs Sniff Your Hands When You Come Home: What Have You Touched?
- Why Dogs Sniff You More After You’ve Been Around Another Dog
Supporting Hub: Scent & Olfaction: How Dogs Use Their Nose to Understand the World
Master Hub: Dog Behavior Explained: Complete Guide to Understanding Your Dog
Sources & Further Reading
- Willis, C. M., Church, S. M., Guest, C. M., Cook, W. A., McCarthy, N., Bransbury, A. J., Church, M. R. T., & Church, J. C. T. (2004). Olfactory detection of human bladder cancer by dogs: Proof of principle study. BMJ, 329, 712. Read the study
- McCulloch, M., Jezierski, T., Broffman, M., Hubbard, A., Turner, K., & Janecki, T. (2006). Diagnostic accuracy of canine scent detection in early- and late-stage lung and breast cancers. Integrative Cancer Therapies, 5(1), 30–39. Read the study
- Elliker, K. R., Sommerville, B. A., Broom, D. M., Neal, D. E., Armstrong, S., & Williams, H. C. (2014). Key considerations for the experimental training and evaluation of cancer odour detection dogs: Lessons learnt from a double-blind, controlled trial of prostate cancer detection. BMC Urology, 14, 22. Read the study
- Guest, C., Pinder, M., Doggett, M., Squires, C., Affara, M., Kandeh, B., Dewhirst, S., Morant, S. V., D’Alessandro, U., Logan, J. G., & Lindsay, S. W. (2019). Trained dogs identify people with malaria parasites by their odour. The Lancet Infectious Diseases, 19(6), 578–580. Read the study
- Bauër, P., Leemans, M., Audureau, E., Gilbert, C., Armal, C., & Fromantin, I. (2022). Remote medical scent detection of cancer and infectious diseases with dogs and rats: A systematic review. Integrative Cancer Therapies, 21, 15347354221140516. Read the review
- Rooney, N., Trivedi, D. K., Sinclair, E., Walton-Doyle, C., Silverdale, M., Barran, P., Kunath, T., Morant, S., Somerville, M., Smith, J., Jones-Diette, J., Corish, J., Milne, J., & Guest, C. (2025). Trained dogs can detect the odor of Parkinson’s disease. Journal of Parkinson’s Disease, 15(6), 1111–1115. Read the study
- Gonder-Frederick, L. A., Grabman, J. H., & Shepard, J. A. (2017). Diabetes alert dogs (DADs): An assessment of accuracy and implications. Diabetes Research and Clinical Practice, 134, 121–130. Read the study
- Wilson, C., Campbell, K., Petzel, Z., & Reeve, C. (2022). Dogs can discriminate between human baseline and psychological stress condition odours. PLOS ONE, 17(9), e0274143. Read the study