
TL;DR: Canos detects and tracks cancer in dogs from a single blood draw by reading the whole cell-free DNA genome. In a benchmark of 896 samples from 341 dogs, our model reached 91% sensitivity and 95% specificity, and it gets better with every new sample. At New York Vet last week, 160+ veterinary professionals from 130 clinics/hospitals came to our booth to learn more in just two days, and we’re now onboarding 10 founding clinics/mobile vets in New York City and the Bay Area.
The problem
1 in 4 dogs will develop cancer in their lifetime, and most cases are caught late, because dogs can’t tell us when something feels wrong. Routine bloodwork rarely gives an answer, and advanced imaging like CT is costly and requires sedation. Vets need tools that not only detect cancer, but also show where it is and what type it is, so they know exactly what to do next.
The solution
We’re bringing the standard of human liquid biopsy to veterinary medicine. Instead of measuring a single biomarker or a fixed panel of mutations, we read the whole-genome cfDNA signal and analyze 500+ multimodal features, including methylation, one of the most robust signals for cancer detection and the key to telling where a cancer comes from. Vets draw blood at a routine visit, and the same test tracks how a dog responds to treatment over time.
The bigger picture
Every test builds something rare: records of the same dogs from health, to cancer, to treatment. Dogs age roughly 7 times faster than humans and develop cancer naturally in our homes, so what takes decades to observe in people unfolds in a few years. This data can unlock breakthroughs in cancer and ageing research for dogs and humans alike, and opens the door to more personalised treatments, such as vaccines.
The team
I’m an Oxford zoologist (PhD), and Bin is an Oxford machine learning scientist (PhD). We’re lifelong animal lovers who have built in pet cancer before: at our previous company, ZYTCA, we built an AI pathology platform for pet cancer in 6 months, and ZYTCA went on to raise $3.3M and launch across Europe, Asia and Latin America. We took Canos from idea to working prototype in 3 months, validating the workflow on our own blood, and have already shared our first scientific paper.
Our ask
Vets and clinics in the Bay Area and New York who want earlier answers for their patients
Dog owners whose dog was recently diagnosed with cancer and want to track treatment response
Intros to veterinary oncologists, hospital groups and pet insurers
Researchers and drug developers interested in comparative oncology or longitudinal cfDNA/immune cell data
Reach us: mingshan@canos.ai | https://calendly.com/ms-tsai-canos/30min | www.canos.ai