Announcing the Symposium Artificial Intelligence in Veterinary Medicine at @cornellvet Join us at @Cornell for this multidisciplinary event! Dates: April 19-21, 2024 Abstract Submission Deadline: January 15, 2024 #vetmed #ArtificialIntelligence web.cvent.com/event/1992516f…

These results yield even tougher questions that our lab is now focused on, such as 1/ Would computational pathology help? 2/ Would better US imaging help? 3/ Are cats diagnosed with IBD at greater risk of Lymphoma later? These are exciting new questions we hope to tackle! eof

Some of this could be explained by the possible continuum IBD to lymphoma. Moreover, histopathologically, it is challenging to draw a line between the two diseases. So we may be seeing mixing of expressions. 10/n

… but performance dropped when discriminating IBD from lymphoma. Performance did not improve when combining CBC/serum data with radiomics data 9/n

… and models (particularly the radiomics features only) to identify cats that benefit from a biopsy also performed well: 8/n

We found we could develop models to predict normal from abnormal with very high accuracy (~90%). 7/n

We used two types of feature selection methods: a wrapper method and one based on decision trees. We capped the maximum number of features to 20, but also used 3, 5, and 10. 6/n

We developed 4 different models: Normal vs Abnormal; Does the cat require or not require a biopsy; and models that could discriminate IBD/Lymphoma/Other (+/- Healthy) 5/n

Our study developed a number of models using different combinations of data 4/n

This study combined US data from 3 institutions and even then we could only obtain ~160 biopsy confirmed diagnoses of cats with a clear transverse US image and blood serum/chemistry panels. 3/n

Our goal was to develop models to improve the sensitivity of ultrasound for discriminating IBD from lymphoma in cats. Details of our research plan were published on the @OSFramework website several years ago -> osf.io/auf5c/ 2/n

Our latest paper on machine leaning via #ultrasound #feline abdominal disorders. 1/n onlinelibrary.wiley.com/doi/…

May the Forth be with you! #vetmed

We are also really excited that a figure we created is on the cover of the journal! (That makes 2 journals our lab's work has made the cover page of a journal this year!).

Special thanks to Del Leary, Ian Porter, @Youshan_, @MircVet, and the tireless Editor @ApplebyDVM for making this happen. These articles act as a great starting point for anyone wishing to learn more about AI, medical imaging, and radiation oncology. 2/3

The lab is pleased to have submitted invited reviews and papers in Veterinary Radiology & Ultrasound for their special edition Supplement on #ArtificialIntelligence onlinelibrary.wiley.com/toc/… 1/3

Our latest publication on the use of #MachineLearning to distinguish from #Thoroughbred #racehorses that suffered catastrophic fractures from those that did not using micro-CT imaging. Open access article here -> mdpi.com/2076-2615/12/21/303…