Dr. Daniel AthertonThe emergence of Artificial Intelligence (AI) in healthcare has shown robust potential for future utilization. One common application is its ability to provide quality assurance for practitioners in various specialties. This direction is the vision of Daniel Atherton, M.D., an associate professor in the Department of Pathology’s Division of Forensics, who worked with colleagues from the Institute for Biomedical Innovation to develop WoundX, a software tool for gunshot wound classification.
“Forensics is very image-focused,” Atherton said. “It aligns well with a lot of what AI has to offer in terms of image classification.”
WoundX was designed to augment forensic pathologists’ decision making by providing quality assurance in catching human error in the classification of gunshot wounds.
“Errors are rare, but they may occur for various reasons,” Atherton said. “Any opportunity we have to explore ways to minimize or reduce errors can be very impactful.”
“One of my responsibilities as a forensic pathologist is to determine information about gunshot wounds. For example, we ascertain the possible distance from the weapon to the victim, and we identify whether gunshot wounds are classified as entrance or exit gunshot wounds. You could imagine how necessary it is to determine whether a bullet entered a victim’s frontside or backside, as those findings have important legal implications.”
WoundX began in 2023 when Atherton learned that pathologists and radiologists were using AI to classify images. He decided to pursue a Master of Clinical Health Informatics at UAB to learn more.
Atherton approached Sandeep Bodduluri, Ph.D., director of AI programs in the Heersink Institute for Biomedical Innovation, with the idea for using AI to help classify gunshot wounds, his proposed graduate capstone project. The two decided to collaborate, with Atherton overseeing clinical components, Bodduluri advising the project, and Stephanie Marie Aguilera Cueto, a Ph.D. student working with Bodduluri, managing technical logistics. These efforts paved the way for the development of the WoundX program.
“Dr. Bodduluri and Stephanie have been the brains behind the interface of this software,” Atherton said. “While a few groups are working on AI models like this, we’re the first to show what a real user interface could look like for gunshot wound classification.”
Atherton says in an ideal world, this would be a software that medical examiner offices could employ daily to quickly automate AI classification and let pathologists know when there’s a disagreement between their interpretation and AI’s.
“We have shown that we can use AI to differentiate gunshot wounds accurately and reliably, and now we’re improving the model.”
Atherton earned his graduate degree in May 2025. His project was selected as part of the Blazer App Accelerator, funded and administered by the Bill L. Harbert Institute for Innovation and Entrepreneurship, to help potential UAB software and app creators validate ideas that are ready for further funding via the Blazer Bridge Fund application process or to launch as potential startup companies. In addition, Atherton and Cueto have been presenting their data at national conferences, showing how WoundX can be integrated into daily forensic workflows.
“The next step in our process is increasing the dataset size and continuing to improve the program’s accuracy,” Atherton said. “We are improving the model in the hope that our interface could make a real difference for forensic pathologists one day.”