New project aims to predict bone healing and reduce complications from fractures

Every year, nearly 10 million Americans experience a broken bone. A quarter of patients with lower leg fractures face delayed healing, and one in 10 patients will develop a nonunion, a break that requires additional major surgery to heal.

The consequences of nonunions are serious and exact a significant physical, mental, and financial toll on patients' quality of life.

There are often complications in fracture healing, and it's difficult to eliminate all of them. Some patients have really severe injuries. Some have biological comorbidities, like Type 1 diabetes, that you can't overcome. The challenge with fracture healing is recognizing when somebody is going to have problems with healing and knowing when to intervene."

Hannah Dailey, Lehigh University researcher, associate professor and associate chair of mechanical engineering and mechanics, P.C. Rossin College of Engineering and Applied Science

Dailey and her team recently received funding as part of an international collaboration with Switzerland's AO Research Institute Davos (ARI) that puts them a step closer to identifying that critical point of intervention. The four-year project is supported by the U.S. National Science Foundation and the Swiss National Science Foundation and aims to create computational models that can predict how bones will heal over time. 

Both mechanical and biological factors influence how a bone heals. Mechanical factors-such as the rigidity of the implant meant to stabilize the fracture, the distance between bone ends, and the loading pattern on the limb-have been studied for decades, says Dailey, and existing models show how these physical and structural conditions set the stage for healing. Less well understood, however, are the biological factors-the cellular, molecular, and systemic processes that actually rebuild the bone. 

"If you and I broke our legs in exactly the same place, in exactly the same way, we would not have identical healing responses because we have different biologies," says Dailey. "And up to this point, those differences aren't something the model could account for. So we're going to change the framework to incorporate these biological differences and make the model more probabilistic."

To do that, the team will use a library of imaging data provided by ARI, one of the world's leading institutes for orthopedic research. The data tracks fracture healing in sheep over time, a process that closely mimics that in humans. (The project uses only existing data, so no new animal studies are required.)

"The richness of this dataset is what's really exciting," she says. "Instead of getting just one picture at the end of the healing process, we have images taken over many months, which allows us to measure what's happening as healing progresses. We can then use that data to feed and tune these predictive models. The model will then inform the physician how healing will progress based on both mechanics and biology. Nobody's done that before."

Another innovative aspect of the project is the team's plan to add their model to ARI's online training platform OSapp. As one of the world's premier organizations for surgeon education, ARI supports interactive simulations available to professionals around the world. The modules provide instruction on topics including instrumentation; manipulating implants, such as plates and screws, and determining their proper length and number; and discussing recovery options with patients. 

"We all know that biology is really hard to control, but surgeons can control mechanics," says Dailey. "For example, they can change the way they use implants or what they tell a patient about how to rehab. Our model will help them visualize how those mechanics can change the biological response."

The long-term goal, she says, is to develop a patient-specific simulation that predicts how an individual's bone will heal based on their biology and the implant used. 

"We want to reduce the complications from poor healing fractures," she says. "Having that predictive capability will be like having a smart crystal ball. Surgeons will be able to make better, earlier decisions about whether to operate again or prescribe a bone stimulator."