Augmenting prion surveillance by immunohistochemistry using artificial intelligence-based image analysis

Authors: BROUGHTON-NEISWANGER, LIAM - Washington State University; Schneider, David; SMITH, JODI - Animal And Plant Health Inspection Service (APHIS); LEHMKUHL, AARON - Animal And Plant Health Inspection Service (APHIS); HOLDER, LAWRENCE - Washington State University

Submitted to: Veterinary Pathology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/22/2026
Publication Date: N/A
Citation: N/A

Interpretive Summary: Chronic wasting disease (CWD) and scrapie are prion diseases of cervid species and small ruminants, respectively. Surveillance for these diseases requires standardized and highly repetitive manual review of tens-of-thousands of microscope slides per year by a limited number of trained federal and state veterinary pathologists. To automate the initial review, whole-slide digital images were made of slides and artificial intelligence models were trained to recognize prion-specific quality-control parameters and staining features. Using a unique set of diagnostic slides, the automated results from the models were in very high agreement with the evaluations by a trained veterinary pathologist. Furthermore, the automated initial review significantly reduced the time needed for evaluation by a specialized veterinary pathologist. Implementation of this technology should enhance diagnostic consistency, improve efficiency, and provide scalable capabilities essential for comprehensive prion surveillance throughout the veterinary diagnostic laboratory network.

Technical Abstract: Chronic wasting disease (CWD) and scrapie are transmissible spongiform encephalopathy diseases caused by prions, infectious forms of the prion protein. Within the past year, novel CWD detections have occurred in several western states, including Washington, Idaho, and California. Currently, immunohistochemistry (IHC) is the sole approved diagnostic method for confirming these prion infections in formalin-fixed tissues. Evaluation of prion IHC requires specially trained veterinary pathologists to assess multiple quality-control parameters as well as detect characteristic chromogen staining patterns. This manual slide review creates a significant bottleneck for veterinary diagnostic laboratories needing to rapidly scale up surveillance during periods of increased demand, as currently experienced in these western states. Given the repetitive and standardized nature of prion IHC slide review, this assay was considered a candidate for computer-assisted diagnostics. To address this challenge, we developed a deep learning image analysis approach specifically tailored to review slides from large-scale veterinary prion disease surveillance. Our training dataset included 144 prion IHC whole-slide images (WSI) containing a total of 3,296 manual annotations. Annotated images were segmented into non-overlapping tiles and then used to fine-tune a pretrained convolutional neural network, enhancing the model’s ability to recognize prion-specific quality-control parameters and staining features. When tested on a separate, blinded dataset of 50 CWD IHC slides, the model achieved 100% concordance for tissue classification (brain vs. lymph node), 94% concordance for identifying relevant anatomical structures (lymphoid follicles and dorsal motor nucleus), and 100% concordance for chromogen staining when compared to evaluation by a trained veterinary pathologist. The overarching objective of this project is to automate the initial review of prion IHC slides using deep learning image analysis models to substantially reduce the time needed for evaluation by specialized veterinary pathologists. Implementation of this technology should enhance diagnostic consistency, improve efficiency, and provide scalable capabilities essential for comprehensive prion surveillance throughout the veterinary diagnostic laboratory network.