Oxford Startup LatusPet Publishes Study on AI Blood Test for Detection of Multiple Diseases in Dogs

LatusPet, a University of Oxford biotechnology startup, has announced the publication of a peer-reviewed study demonstrating that its proprietary SINO diagnostic platform can detect multiple diseases in dogs—including cancer and cardiovascular disease—from a single blood sample. The findings, published in Frontiers in Veterinary Science, represent one of the first demonstrations of combining nuclear magnetic resonance (NMR)-based metabolomics with artificial intelligence to simultaneously screen companion animals for multiple disease conditions.

The company believes the technology could fundamentally change preventive veterinary medicine by enabling routine health screening during regular veterinary visits, allowing clinicians to identify serious diseases before clinical signs become apparent.

Single Blood Sample, Multiple Health Insights

The SINO platform combines advanced NMR metabolomic analysis with machine learning algorithms to identify disease-specific metabolic signatures circulating in a dog’s blood. Rather than searching for a single biomarker, the platform analyses hundreds of small molecules that collectively reflect the animal’s physiological state.

According to the published study, the system successfully detected cancer, cardiovascular disease, and differentiated healthy dogs from those with underlying disease using one serum sample, achieving an overall diagnostic performance exceeding 90% (ROC AUC).

Real-World Clinical Validation

The study evaluated blood samples from 139 dogs presented at veterinary centres across Italy. Importantly, researchers designed the validation to reflect real clinical practice.

Instead of comparing diseased dogs only against healthy controls—a common limitation in diagnostic research—the investigators challenged the algorithm by distinguishing each disease from a mixed population containing both healthy animals and dogs affected by other illnesses. This approach more closely mirrors the diagnostic uncertainty encountered in everyday veterinary practice.

The research was led by Professor Riccardo Finotello of Link University in Rome in collaboration with scientists from the University of Oxford and veterinary clinical partners.

Moving Beyond Reactive Medicine

LatusPet says its objective is to shift veterinary healthcare from diagnosing disease after symptoms appear to identifying risk during routine wellness examinations.

Chief Veterinary Officer Dr. Simone Scoccianti said the platform could enable veterinarians to detect disease during annual health checks, allowing intervention before clinical deterioration becomes evident. Chief Scientific Officer Professor Daniel C. Anthony added that disease leaves measurable biochemical signatures in blood which the platform has been trained to interpret simultaneously for multiple conditions.

Lead investigator Professor Finotello also suggested the technology may eventually extend beyond early diagnosis to monitoring treatment response and disease remission through repeated non-invasive blood testing.

Expansion to Additional Diseases

The company indicated that the current publication represents the first stage of a broader diagnostic programme. LatusPet plans to validate the platform in larger clinical cohorts while progressively expanding its disease library to include additional canine disorders. Development work is also underway to adapt the technology for other companion animal species.

Growing Role of AI in Veterinary Diagnostics

Artificial intelligence is increasingly being incorporated into veterinary diagnostics, ranging from digital imaging and pathology to clinical decision support and laboratory analytics. The publication of the SINO study highlights another emerging application—integrating metabolomics with machine learning to identify complex disease patterns that may not be detectable using conventional laboratory tests alone.

Unlike single-analyte diagnostic assays, metabolomic profiling evaluates broad biochemical changes associated with disease, potentially enabling earlier detection of multiple conditions through a single laboratory workflow.

Industry Significance

Companion animal healthcare continues to move toward preventive and personalised medicine, driven by advances in molecular diagnostics, genomics and artificial intelligence. If further validated in larger populations, multi-disease blood screening platforms such as SINO could become valuable additions to routine veterinary health programmes, particularly for ageing dogs and breeds predisposed to chronic disease.

For veterinary practitioners, the technology offers the prospect of earlier intervention and improved clinical decision-making. For pet owners, it could provide a non-invasive method of identifying serious health conditions before they progress to advanced disease.

With publication of its first peer-reviewed clinical evidence, LatusPet has taken an important step toward translating advanced metabolomic and AI technologies from research laboratories into everyday companion animal practice.