HPC-facilitated modeling of bone may be key to helping those affected by bone disease

04.19.16 -

by Susan Szuch

Professor Iwona Jasiuk has spent the past 10 years studying bones to understand why bones break and to assess bone quality. While her research has been looking at the problem both experimentally and computationally, she is currently focusing on the use of high scale computations to model bone by using patient specific medical images.

"If we can understand better why bones break due to bone diseases, then we can design a better treatment or diagnose disease earlier, or we can prevent disease-related factors or we can assess effectiveness of treatments," Jasiuk says.

By finding the differences between bone affected by disease and normal bone, she may find the key to helping people with osteoporosis.

Osteoporosis is a bone disease characterized by low bone mass and microstructural deterioration which leads to bone fragility. Age, poor diet, lack of exercise, and genetic and hormonal factors may lead to osteoporosis. For example, women after menopause, people on bed rest, as well as those who have spent time in space may be affected. Osteoporosis has no symptoms until bone fractures, and there is no cure but early treatments can slow down its progression. However, medications have side effects such as abnormal heart rhythm or loss of bone in the jaw, so accurate prognosis is important.

Currently, osteoporosis is diagnosed by measuring bone mineral density which provides information on bone quantity rather than quality and it does not always accurately predict the risk of fracture. Thus, better diagnostic tools are needed.

Bone has a highly complex, spatially changing structure, consisting of minerals and proteins (collagen) which is assembled in a hierarchical way from the atoms to the whole bone level. Thus, bone fracture is a complex process which cannot be fully explained by only measuring a mineral content.

Due to the complexity of bone structure and the fact that fractures initiate at atomic level and progress across structural scales, multiscale models of bone are needed. Such computations are possible using the computing power of Blue Waters which allow to create patient-specific models of bone fracture.

Seid Koric, adjunct professor in mechanical science and engineering as well as technical program manager for the Private Sector Program at NCSA, has been working with Jasiuk for the past two years. He sees the benefits for the biomechanical field that the research offers, and also notes that things learned about the bone structure and its fracture behavior can benefit other fields.

"It's not only for biomechanical research, fractures and osteoporosis, but actually we can learn from the structure of the bone how to design new materials with bone-like structure which optimize bearing loads," Koric says. "We can design better fracture resistant materials and structures by just emulating nature."

The experimental research done in Jasiuk's lab has been on pigs rather than human bone. Due to limited access to human bones and transmissible diseases, it's safer for researchers to work with pig or other animal bones.

However, they do utilize human bones in the computational research, by way of images from experiments the Mayo Clinic has done on cadaver human bones that involve fractures.

Koric also views this interdisciplinary collaboration as emblematic of how NCSA is engaging with the faculty at the University of Illinois at Urbana-Champaign campus, a university known for its research and innovation.

National Science Foundation

Blue Waters is supported by the National Science Foundation through awards ACI-0725070 and ACI-1238993.