Dr. Liam Broughton-Neiswanger, a pathologist at the Washington Animal Disease Diagnostic Laboratory, peered at his computer screen where a crisp, high-resolution image of cancerous cells was displayed.
“This is from the jaw of a dog,” he said before pointing to an area of cells highlighted with dark red dye. “This is a tumor growing between the tooth and the bone, and it’s taking over the mouth.”
Until recently, pathologists at WADDL, part of the College of Veterinary Medicine at Washington State University, did the bulk of their work under a microscope, manually viewing cells on glass slides. Broughton-Neiswanger, an assistant professor in the Department of Veterinary Microbiology and Pathology, has led an effort to improve efficiency and accuracy by digitizing all diagnostic pathology slides using a specialized glass slide scanning device.
WADDL is among only a handful of animal disease diagnostic labs to fully implement this technology into its diagnostic workflow. The switch has opened new educational opportunities and possibilities for collaboration with researchers all over the world.
But what comes next — the addition of artificial intelligence — will be an even bigger game-changer for the laboratory. The plan is to develop computer algorithms designed to examine and analyze the digital slides and to flag samples that warrant closer examination by a pathologist.
“For a case like this, you could apply image analysis algorithms to do things like count mitotic figures or to apply published grading schemes, which can help with prognostication of the tumor,” Broughton-Neiswanger said.
Broughton-Neiswanger received USDA grant funding to develop the technology to detect chronic wasting disease, a neurodegenerative prion disease affecting farmed and wild cervids, an animal group that includes deer and elk. Other prion diseases include mad cow disease of cattle and scrapie of sheep. From July 2022 through June 2023, pathologists at WADDL examined 2,830 scrapie test samples alone.
For surveillance of these diseases, pathologists examine specific sections of the brain and lymph node to detect the presence of prion proteins. The process can be tedious and time-consuming, and almost all samples are negative.
“This is a repetitive task performed on a specific set of tissues, so I want to develop an image analysis program to automate the screening process to flag slides for pathologists to examine when something is detected,” Broughton-Neiswanger said.
The goal, he added, is not to replace pathologists, but to augment their work and allow for more efficient use of time.
Detecting prion diseases is just the beginning, as the technology can be expanded for use in many of the pathology tests performed at WADDL. Broughton-Neiswanger expects the technology to be widely used in veterinary diagnostic medicine in the coming years.
“As far as I am aware, the use of artificial intelligence in veterinary diagnostic anatomic pathology is limited to non-existent. In human medicine, it is also still pretty limited,” he said. “But over the next five to 10 years, I bet it’s going to become an integral part of a pathologist’s job.”
Broughton-Neiswanger was named as the head of WADDL’s histology section in the fall of 2022. Soon after, he began the process of converting the traditional pathology workflow to a complete digital system.
“The digital images give you a lot of flexibility,” Broughton-Neiswanger said. “You can zoom in and out; you can drop a pin; you can measure margins with ease and asking for a digital second opinion is a snap. All this allows freedom for the WADDL pathologists to do their work when and where they want to. This really is the way of the future for pathology.”
Digitization has also opened the door to opportunities for collaboration with researchers within and outside the university. Broughton-Neiswanger has been working to digitize slides for a project involving hair follicles led by Ryan Driskell, the head of the Fibroblast and Skin Regeneration Laboratory in WSU’s School of Molecular Biosciences.
“One of my goals is to get it out there to the college that we have this slide scanner, and we can help researchers with their projects that use histology slides,” Broughton-Neiswanger said. “This technology, along with image analysis algorithms, really has the potential to help advance histology-based research. I, for one, am excited for AI to revolutionize the field of anatomic pathology.”
