WORCESTER – Bodily wounds that persist or don’t go away – also known as “chronic wounds” – can be potentially deadly.
Diabetics can develop sores that can become infected. Nursing home patients develop bed sores from sitting or lying in one place for too long. Other sores are linked to cancer, poor circulation or vascular problems.
What starts out as a small opening can grow larger, develop into an infection and, in the worst case scenario, result in amputation and complications that can lead to death.
Now, there’s an effort by a team of researchers at Worcester Polytechnic Institute and UMass Chan Medical School that could potentially identify infected wounds before they become problematic.
The team is working on what’s called the “Deep Infected Wound Detector,” a smartphone app that uses photographs, thermal imaging and artificial intelligence to identify wounds that are infected and in need of advanced care.
“This is extremely important,” said Emmaneul Agu, a professor of computer science at WPI and one of the co-leaders of the app’s development. “The app has the potential to be as good as a world-class medical expert.”
Chronic wounds that do not heal account for millions of visits to hospital emergency departments annually, and some studies estimate that these visits cost Medicare more than $30 billion annually.
see how it works
To develop the application, hundreds of photographs are collected from patients at the Wound Center at UMass Memorial Medical Center. The photos are fed into computer algorithms that Agu called “neural networks” to detect infections.
The main advancement in this application is thermal imaging.
The app is on smartphones equipped with thermal cameras to detect temperature changes under the skin that might go unnoticed during a routine check of a wound. These temperature changes can signal the onset of the infection.
So when doctors leave the hospital to perform regular examinations of a patient’s wound, the app can identify the infection and the need to see a specialist.
For example, when a nurse visits a patient at home or in a nursing home to do a regular check on a wound, the application can detect whether it is healing properly, showing signs of infection, or infected. If infected, the patient is referred to a specialist for more advanced care.
In addition to offering quality medical care to patients, increasing the efficiency of medical resources is another objective of the application.
Dr. Giorgio Giatsidis, assistant professor of plastic surgery at UMass Chan and co-lead of app development, explained that some patients come to the Wound Center weekly for a checkup.
But some of those same patients don’t need to visit as often.
“(The app) supports everything we want to achieve. Offering efficient and safe home care and booking hospital care for those in need,” said Giatsidis.
More than a decade of work
The research team started developing the app 11 years ago.
The initial version included photos of patients’ wounds, focusing on diabetes, measuring factors such as blood sugar and the patient’s exercise level. A $1.2 million grant from the National Science Foundation (NSF) supported the work.
Wound detection was another feature of this app, and it caught the attention of the National Institutes of Health (NIH) with a grant of $1.6 million. The NIH recently provided a grant of $2.4 million over the next four years to support development, including the thermal imaging component.
Application performance will be compared to medical examinations of wounds. Researchers are looking to the app to get a correct diagnosis at least 80% of the time. If that happens, Agu said there is potential to bring the app to the commercial market.
A possible spin-off deal at WPI with investors contributing $2 million to $5 million is a possibility, Agu said.
Father’s illness is a motivator
Agu’s father, Phillip Nnaemeka-Agu, suffered from diabetes and complications from bed sores towards the end of his life.
His father died in 2011 and Agu said this motivated him to continue his research and see that the app helps patients get the care they need to safely treat their chronic wounds.
“We invested 11 years in our research”, said Agu. “The ultimate goal is to get this app into the hands of nurses to reduce poor outcomes.”
Contact Henry Schwan at [email protected] Follow him on Twitter @henrytelegram