By Matthew Chin
Building on an advanced imaging technique, researchers at the UCLA Henry Samueli School of Engineering and Applied Science have developed algorithms that could be used to identify foot ulcers associated with diabetes weeks before they are visible to the eye.
Physicians could use this technique, called hyperspectral imaging and common in many applications, to diagnose and suggest treatment options before the ulcers have formed.
Diabetes affects 194 million people worldwide and is expected to increase to 439 million in 20 years. Ulcers on the foot are a major complication of the disease affecting 15-25% of diabetes patients, and in 2006, was the cause for more than 88,000 limb amputations on diabetes patients.
The researchers, led by principal investigator Laurent Pilon, a UCLA Engineering associate professor affiliated with the Biomedical Engineering InterDepartmental Program, have conducted several studies on incorporating hyperspectral imaging into medical care for diabetic patients.
Hyperspectral imaging is a series of images, each taken in a narrow band of the light spectrum. This series of images is then combined into a three-dimensional composite image called a hypercube. Imagine using a camera with a series of color filters and taking photos of the same object, then layering those filtered photos, one on top of the other, to produce a very detailed photo. The hyperspectral imaging algorithm developed at UCLA Engineering can analyze the hypercube of a diabetic patient’s foot and determine a map of the epidermal layer thickness; blood volume; and blood oxygen levels over the patient foot – all of which are associated with foot ulcer formation and healing. The algorithm also correct for light absorption by melanin which give our skin its color and protect it from ultraviolet radiation.
The team also includes UCLA Engineering graduate student Dmitry Yudovsky, who was the lead author on a series of research papers; Dr. Aksone Nouvong, a podiatric clinician at UCLA Olive View Medical Center’s Department of Surgery; and Kevin Schromaker, of HyperMed Imaging, Inc., a biomedical equipment company based in Burlington, Mass.
“Combining our optical models and algorithms with HyperMed’s hyperspectral imaging hardware could really improve the care of diabetic patients and reduce complications associated with the disease,” Pilon said.
In one study, the UCLA researchers examined hyperspectral images that Nouvong collected at Olive View Medical Center from 66 diabetic patients – with varying degrees of foot ulcers. They used that data to develop an algorithm that predicts where ulcers would occur based on the hyperspectral images.
In their most recent study, published in the Journal of Biophotonics, the research team, used hyperspectral imaging to monitor two patients over several months following the formation, and healing of their foot ulcers. They found that thickening of the skin and a decrease in blood-oxygen concentrations can be detected by hyperspectral imaging and analysis and therefore identify pre-ulcer sites on the foot before an ulcer is formed. They also developed an ulcer healing prediction index based on the data.
“Using the system, we can diagnose whether a diabetic foot ulcer will heal or not and determine if a new ulcer will be forming and where it might form,” Nouvong said. “This is an important tool to have for clinicians who treat diabetic patients who are at risk for ulceration, which may lead to infection and often times amputation.”
Nouvong added that the improved system can both help doctors prescribe treatments at an earlier stage, depending whether or not a wound is likely to heal, and if a certain area is likely to ulcer.
Hyperspectral imaging technology is several decades old, first used by U.S. military surveillance planes in the Vietnam War to look for tanks hidden in the dense jungle, said HyperMed Imaging, Inc.Consultant, Rick Lifsitz. After the technology was declassified, HyperMed Imaging started to look into medical diagnostic applications for the technology a little more than 10 years ago. They connected with Nouvong, who subsequently connected with Pilon and Yudovsky to improve the system.
“The work that was done gave us further validation that we’re ready to move toward new applications, and it’s clear that the approach that was taken by Dmitry and Professor Pilon was well-thought out and the correct approach to the problem,” Lifsitz said.
Lifsitz said that the UCLA team’s work may be included in further development of commercial hyperspectral imaging applications for diabetes ulcer detection applications, as well as possibly other areas in medical diagnostics.
The research was funded in part by a grant from the National Institute of Diabetes and Digestive and Kidney Diseases.
For more information, please visit’s Professor Pilon’s Web site:
Main Image Credit: Journal of Biophotonics, http://onlinelibrary.wiley.com/doi/10.1002/jbio.201000117/abstract