Emerging technologies such as artificial intelligence (AI), machine learning, and deep learning, which generates data with numerous levels of abstraction, have the potential to alter how veterinary medicine is done. They were created to facilitate decision-making when practitioners examine medical images by enhancing predictive analytics and diagnostic performance. However, veterinary medicine does not require premarket testing of AI systems, in contrast to human medicine.
As a result, it becomes even more crucial for the veterinary profession to develop best practices to safeguard care teams, patients, and clients, especially when it comes to conditions with a poor prognosis where such interpretations may result in a decision to euthanize.
The North Carolina State College of Veterinary Medicine’s clinical professor of diagnostic imaging, Dr. Eli Cohen, says that. He gave the presentation for the AVMA Axon webinar, “Do No Harm: Ethical and Legal Implications of A.I.,” which made its premiere in late August.
He examined how AI might improve radiology’s efficiency and accuracy during the lecture, but he also highlighted its flaws and risks.
The potential uses
According to a Currents in One Health paper published in JAVMA in May 2022, the use of AI in clinical diagnostic imaging practice will keep expanding, in large part because a lot of the data—radiographs, ultrasound, CT, MRI, and nuclear medicine—and their related reports are in digital form.
The paper’s author, Dr. Ryan Appleby, an assistant professor at the University of Guelph Ontario Veterinary College, noted that task-speeding with artificial intelligence can be quite beneficial.
AI can be used, for instance, to develop hanging protocols—instructions on how to arrange photos for best viewing—or to bring up report templates based on the body parts involved in the study. It can also be used to automatically rotate or position digital radiographs.
More generally, AI can triage processes by performing a first pass on various imaging studies and pushing more urgent patients to the front of the line, according to Dr. Appleby, chair of the Artificial Intelligence Committee of the American College of Veterinary Radiology (ACVR).
To ensure that patients are treated correctly and for it to be beneficial, AI must read radiographs in a way that not only identifies frequent cases of a disease but also flags border cases.
Dr. Cohen, a radiologist and co-owner of Dragonfly Imaging, a teleradiology company, said, “As a specialist, I’m there for the subset of times when there is something unusual.” “AI isn’t perfect, but it will become better. When it doesn’t function properly, we need to be able to troubleshoot it.
