Researchers launch largest open-source dog and cat tumor database

A UK-Spain research collaboration says it has created the world’s largest open-source tumor database for dogs and cats, bringing together more than 1 million records in a single resource aimed at veterinary cancer surveillance and comparative oncology. The database was developed by the University of Liverpool’s Veterinary Data Science Group and the University of Las Palmas de Gran Canaria, and the team says its scale should make it easier to detect patterns in uncommon cancers and in specific breeds that smaller datasets can miss. (liverpool.ac.uk)

The announcement builds on years of registry work through SAVSNET, Liverpool’s surveillance network, which already collects large volumes of routine diagnostic data from veterinary laboratories. In a 2021 Scientific Data paper, the group described a pathology-based tumor registry created from 180,232 canine and feline electronic pathology records submitted by three UK labs between April 2018 and June 2019. That earlier dataset yielded 109,895 tumors and was described at the time as the largest national-level animal pathology-based tumor registry, surpassing the roughly 85,000-case Swiss registries cited for comparison. (nature.com)

What appears to have changed now is scale and usability. According to Liverpool’s March 2026 announcement, the registry has grown past 1 million records and is being positioned as an open-source platform for researchers worldwide. The team said one element of the work, focused on dogs, is already reflected in a newly published Veterinary and Comparative Oncology paper, “Epidemiology of Four Major Canine Tumours in the UK: Insights From a National Pathology Registry With Comparative Oncology Perspectives.” In the university repository summary, the authors say the registry can identify breed-specific and demographic risk factors and serve as a foundation for comparative cancer epidemiology. (liverpool.ac.uk)

That matters because veterinary oncology has long faced a data fragmentation problem. The 2021 registry paper noted that most veterinary tumor registries have been hospital-based or pathology-based, often limited in duration, incomplete in coverage, or difficult to compare across systems. It also emphasized that routine pathology reports contain rich information, but much of it sits in free text and isn’t easily searchable unless it’s normalized. The SAVSNET approach used text mining to classify diagnoses, locations, grades, and related descriptors, turning routine lab submissions into structured data for surveillance and research. That need for broader, more standardized surveillance is echoed elsewhere: a 20-year California hospital study covering 150,063 dogs and cats found 26,883 cancer cases, showed age was the strongest predictor in both species, and concluded that a dedicated companion-animal cancer registry is needed to provide a more complete epidemiologic picture. (nature.com)

Industry reaction so far has centered on access and practicality. A veterinary news report on the launch highlighted a new tumor risk dashboard tied to the registry, framing it as a way for veterinarians to explore cancer patterns in dogs and cats rather than relying only on isolated institutional datasets. That framing is consistent with how the Veterinary Cancer Society describes oncology-pathology collaboration more broadly: standardized pathology and integrated data are central to improving reporting quality, prognostic interpretation, and cancer research. (vetsurgeon.org)

The broader backdrop is comparative oncology, and that context has sharpened considerably in the past year. A large international Science study generated the first large-scale genetic map of feline cancer by sequencing 493 tumor-normal pairs across 13 feline cancer types collected in five countries. Researchers examined feline orthologs of roughly 1,000 human cancer-associated genes and found recurrent alterations in genes already familiar in human oncology, including TP53, FBXW7, CTNNB1, PTEN, and TRAF3. TP53 was the most frequently mutated gene overall, appearing in about a third of tumors, while mammary carcinoma stood out for strong parallels to human breast cancer: FBXW7 was mutated in more than half of feline mammary tumors, PIK3CA alterations were also common, and some FBXW7-mutant samples showed increased sensitivity in laboratory testing to certain chemotherapy drugs. The study also reported potentially actionable alterations in a subset of feline tumors, reinforcing the idea that naturally occurring cancers in pets can inform both veterinary and human precision oncology. (liverpool.ac.uk)

That comparative promise is real, but it also comes with practical limits. A recent review in Life argued that dogs and cats can mirror human cancers in histopathology, molecular alterations, tumor microenvironment, and even some treatment-response patterns better than traditional rodent models, while also sharing environmental exposures with people. At the same time, the review stressed that species differences in drug metabolism, enzyme expression, tolerability, and dosing remain important, meaning translational insights still need veterinary-specific validation rather than simple borrowing from human oncology. (liverpool.ac.uk)

There are also examples of how registry-scale epidemiology could eventually connect with more targeted clinical tools. In canine lymphoma, researchers at Tufts and UMass Chan recently reported that blood-based immune gene signatures in pet dogs with diffuse large B-cell lymphoma were associated with treatment response in a clinical trial setting, suggesting a future role for less-invasive biomarker testing to identify poor responders earlier. And in species-specific pathology, a long-term institutional cohort from Italy found that mammary tumors in male dogs and cats, though uncommon, were predominantly malignant in both species and tended to be higher grade in cats—another reminder that large, well-annotated datasets can help clarify patterns that are easy to miss in routine practice. (liverpool.ac.uk)

Why it matters: For veterinary professionals, a dataset this large could improve both research and day-to-day conversations in practice. On the research side, it may help generate stronger signals around breed predispositions, age-related risk, tumor distribution, and rare neoplasms that are difficult to study in single-center cohorts. On the clinical side, those signals can eventually sharpen screening awareness, referral decisions, and how veterinarians counsel pet parents about relative risk and diagnostic next steps. The California epidemiology study offers a useful example of the kinds of patterns these databases can clarify: cancer risk rose strongly with age in both dogs and cats, incidence trends differed by species, and some associations varied by sex and neuter status depending on tumor type. There are still limits: pathology-based registries don’t capture every cancer case, can include duplicate tumors from repeat sampling, and may reflect submission bias rather than true population incidence. But even with those constraints, they can be highly useful for trend detection, hypothesis generation, and eventually for linking population-level risk with genomic and biomarker findings. (nature.com)

The broader backdrop is comparative oncology. Large canine and feline cancer datasets are increasingly valuable not just for veterinary medicine, but also for studying naturally occurring cancers that mirror human disease. Liverpool’s team explicitly positions the registry in that One Health context, and the recent feline genomics work strengthens the case by showing substantial overlap between feline and human tumors across mammary, lung, skin, blood, gastrointestinal, bone, and central nervous system cancers. That doesn’t make this registry a clinical tool on its own, but it does make it a potentially important infrastructure asset for future translational studies—and a useful complement to the growing push toward precision oncology in companion animals. (liverpool.ac.uk)

What to watch: The next step is likely a steady stream of tumor-specific and breed-specific papers from the registry, along with possible expansion of public-facing tools such as risk dashboards; the key question will be how quickly this research infrastructure translates into practice-level guidance that veterinarians can actually use. Just as important will be whether large registries can be linked more systematically with genomic profiling, pathology standardization, and treatment-response biomarkers, creating a more connected evidence base for veterinary oncology rather than separate islands of epidemiology, genomics, and clinical trials. (liverpool.ac.uk)