A pack of beagles outside Bengaluru just outperformed most lab tests at spotting cancer before it spreads.
Quick Take
- Bengaluru startup Dognosis trains dogs to sniff human breath samples for signs of cancer.
- A Phase 2 trial reported over 90% sensitivity and specificity across seven cancer types, including early stages.
- The company pairs canine noses with sensors and artificial intelligence to turn dog reactions into readable data.
- Results were published in the Journal of Clinical Oncology, and the company has raised funding to build an at-home test kit.
Dogs With Jobs, And A Lab Coat
On a two-acre farm outside Bengaluru, three beagles named Billy, Jessie and Banu clock in for work each day. Their job is not fetching balls. They sniff breath samples looking for early signs of cancer, using noses far more sensitive than any current medical scanner. Dognosis, the startup that trains them, was founded in 2023 by Akash Kulgod and Itamar Bitan.
The idea builds on something scientists already knew. Cancer cells release volatile organic compounds, tiny chemical traces that escape the body through breath. Dogs, with up to 300 million scent receptors compared to a human’s 6 million, can detect these compounds at levels no lab machine can match right now.
How A Breath Test Becomes A Data Point
Patients breathe into a special mask for ten minutes. The mask traps the volatile compounds, then gets shipped to a Dognosis lab. There, trained dogs sniff the sample while wearing a wearable brain-computer interface that reads their neural and behavioral responses in real time. Machine-learning software then translates those canine reactions into a measurable, repeatable signal instead of just a bark or a nudge.
That combination, dog plus sensor plus algorithm, is what makes the approach different from older “cancer-sniffing dog” demonstrations. Instead of relying on a single dog’s judgment, Dognosis runs multiple dogs per sample and feeds their combined responses into a Bayesian model built to reduce human and animal error alike.
The Numbers Behind The Headline
Dognosis says its model achieved sensitivity and specificity above 90%, with an area under the curve of 0.962, in a trial covering 1,502 patients. Performance held steady across cancer types and at Stage 1 and Stage 2, the earliest and most treatable phases of disease. The company calls it the largest study of its kind published in a leading cancer journal, and results appeared in the Journal of Clinical Oncology.
The finding fits a broader trend in breath-based diagnostics. A separate peer-reviewed study using dogs and artificial intelligence together detected four cancer types with 94% accuracy, reinforcing that canine olfaction paired with machine learning produces results worth taking seriously. Researchers have documented dogs identifying more than two dozen diseases through scent alone over the past two decades.
Why Cost And Access Matter Here
Most cancer screening tools, from colonoscopies to mammograms to blood-based tests, require expensive equipment, trained specialists, or invasive procedures. Dognosis is chasing something different: a cheap, non-invasive test that could reach patients in rural India or low-income countries where advanced screening infrastructure barely exists. That mission matters more than the novelty of using dogs at all.
Early cancer detection saves lives largely because it saves time. Catching a tumor at Stage 1 instead of Stage 4 can mean the difference between routine treatment and a terminal diagnosis. A test that is cheap enough for mass screening, even if it is not perfect, could shift outcomes for millions who currently have no access to screening at all.
What Comes Next For The Company
Dognosis has raised $1.6 million in funding and plans to ship an at-home cancer screening kit next year, letting patients mail in a breath sample without visiting a clinic first. The company frames its work as building on more than 40 double-blind trials of canine disease detection conducted over the past two decades, adding scale and standardization that earlier proof-of-concept studies lacked.
Whether this model becomes a mainstream screening tool will depend on further trials, regulatory review, and real-world rollout. But the underlying science, dogs detecting disease through scent, is not new. What is new is turning that instinct into a scalable, data-driven product that could put early cancer screening within reach of people who have never had it before.
Sources:
enterpriseai.economictimes.indiatimes.com, dognosis.tech, yourstory.com, fenado.ai, linkedin.com
