A technique used to detect land mines could soon help rheumatologists and radiologists diagnosis osteoporosis and cervical cancer, thanks in part to the efforts of a University of Missouri-Rolla researcher.
For the past five years, Dr. Joe Stanley, assistant professor of electrical and computer engineering at UMR, has been using image analysis and processing techniques to study human cervical and lumbar spine X-rays. The research is funded by the National Library of Medicine.
“I’m identifying key attributes that can be used to characterize osteoarthritis,” Stanley says. “The goal is to develop a general search engine that rheumatologists, radiologists, and other heath care professionals could access to query different X-rays to use in their own practice.”
Stanley is working with radiologists at Phelps County Regional Medical Center in Rolla, Mo., and other specialists in St. Louis to ascertain what features they look at when determining if patients have osteoarthritis, such as disk narrowing or bony growths on the vertebrae. His work will provide a semi-automated approach for segmenting vertebrae in the X-rays, allowing radiologists to select points of interest to highlight.
Ultimately, the tool could serve as a teaching aid, giving heath care professionals X-ray examples to compare to their own patients and helping to diagnose and guide treatment plans.
Stanley’s work with image analysis began as a Ph.D. student at the University of Missouri-Columbia, where he used spatially distributed features to identify human chromosomes and to find chromosome anomalies. After joining UMR as a faculty member in 1999, Stanley customized his image-processing techniques to help detect land mines in a project for the U.S. Army.
“It’s funny how you can take techniques that work with one problem and apply them to solve another one,” Stanley says. “Sometimes it doesn’t work, but sometimes it does.”