Leptospirosis can be lethal in dogs, and may result in kidney failure, liver disease, bleeding lungs, and death. However, early detection can enable veterinarians to nip the disease in the bud. Now there’s an AI model for detecting leptospirosis in the early stages.
Created by UC Davis, the model was developed by analyzing data from more than 400 canine patients tested for leptospirosis at the university’s facility. AI was deployed that used intricate statistical methods to look for patterns associated with a leptospirosis outcome in the dogs’ blood work. A system was then developed that could be applied to new lab work for making a prediction about the presence of infection. Traditionally, Leptospiratesting lacks sensitivity in the early phase of the infection.
Leptospirosis tests also involve the quantification of an increasing number of antibodies in blood samples over time. It usually takes over two weeks to detect the disease — which is unfortunate because of the severe pathogenic potential of the bacterium.
The UC Davis leptospirosis prediction AI model removes all the diagnostic roadblocks and has shown groundbreaking results so far. In a test group of new dogs, nine were identifiedas positive. All the predictions were correct, translating to 100% sensitivity.
The developers’ goal for the AI model is to make it a readily available online resource for veterinarians so they can easily reach an early diagnosis in cases of leptospirosis.
This recent initiative highlights the value of AI-based diagnostics. “AI-based, clinical decision-making is going to be the future for many aspects of veterinary medicine,” says MarkStetter, Dean of the UC Davis School of Veterinary Medicine.