UW researcher creates AI-enabled technology to detect bone fractures

UW researcher creates AI-enabled technology to detect bone fractures
UW researcher creates AI-enabled technology to detect bone fractures

A University of Waterloo (UW) engineer helped develop AI-powered technology to improve the way doctors detect a broken bone.

Omar Ramahi, a professor at the Department of Electrical and Computer Engineering, is the principal investigator on the project and claims it could be a breakthrough for the medical field.

He has developed an instrument that can detect bone fractures without the need for radiation.

“Early diagnosis allows doctors to assess what needs to be fixed,” Ramahi said.

It can take some time for a patient to get an X-ray, CT scan or MRI scan. The equipment is not readily available in ambulances or primary care centers. Patients often have to wait for an X-ray or scan in the hospital.

Ramahi said his idea offers a more portable and cost-effective alternative to what currently exists.

Ramahi uses a cow bone for the study. He combines low-cost wireless communication antennas with artificial intelligence. Two antennas are placed on either side of the suspected fracture site. One antenna sends low-frequency microwaves through the bone to the other. A deep neural network AI model, trained on extensive datasets of human body parts and bone fracture types, then analyzes the data.

“So it’s actually information, rather than an image or an impression,” Ramahi said.

Another advantage of the tool is interpretation. Conventional medical imaging methods require expert interpretation, but this system can provide a preliminary fracture diagnosis in emergencies and avoids the use of radiation.

“The energy is much less than what we use in our cell phone. So it’s a very benign, non-invasive, non-ionizing, non-harmful substance,” Ramahi said.

The system is still in the lab phase, but portability is key. The researchers hope to develop a wearable device, such as a cuff that wraps around the injured area, and one day get the technology into the hands of medical professionals.

“We can come up with a prototype within a year to 18 months because the system is very simple,” Ramahi said.

Ramahi said it could be used in a variety of medical sectors, such as paramedics, doctors for elite sports teams, nursing homes and emergency room patients.

Ramahi said it is not intended to replace medical imaging, but to assist medical professionals. He said it could be used in a variety of areas, from paramedics to elite sports teams, or nursing homes and emergency room patients.

“Something that has 100 percent prediction never exists in the biomedical world. But if we can have something that can compete with what’s available at a much lower cost, I think we’re making good progress,” Ramahi said.

According to the University of Waterloo, the system was developed in collaboration with an international research team. It is believed to be the first system to use microwave AI to detect bone fractures without using imaging techniques.