Non-invasive measurement of bone

Bone adaptation is understood to be driven by mechanical stimuli that are related to bone strain. Logically, if bone strain is a driver of adaptation and osteogenic response, then knowing how human bone is strained in vivo during certain activities might help us to identify specific strategies that could be used to help people naturally gain bone strength. Our laboratory is working to develop:

• Non-invasive measurements of bone macro- and micro-structure
• Non-invasive methods to measure subject-specific bone strain

Projects Related to Bone Structure

Quantitative Computed Tomography Analysis

Computed tomography (CT) is a medical imaging technique that generates three-dimensional x-ray based data. The gray-value of the image pixels, measured in Hounsfield Units, is linearly related to the x-ray attenuation of the object at a specified location. In bone, x-ray attenuation is directly related to the density of the bone mineral (hydroxyapetite), so by including a calibration phantom with known hydroxyapetite densities, calculations can be made about the bone itself. Typical calculations include bone density, bone mineral content, bone volume, and mechanical measures such as moments of inertia. This technique can be used on any clinical CT scan, so long as calibration information is known.

High-Resolution Peripheral Quantitative Computed Tomography (HR-pQCT)

Specialized high-resolution CT scanners have recently become available for use in human subjects. Our laboratory uses HR-pQCT to quantify bone micro-structure in volunteers who are participating in various clinical research studies for us. Typical sites are the distal radius (forearm) and tibia (ankle), although we have developed imaging protocols for other sites as well. The principle of HR-pQCT analysis is identical to QCT analysis, with the major difference being the scan resolution. Whereas the clinical scans we analyze are typically 231 x 231 x 625 microns per pixel, the HR scans are 82 x 82 x 82 microns per pixel. This high level of detail makes it possible to measure parameters related to bone micro-structure such as trabecular connectivity and thickness. The trade-off is that only a very small region of the bone can be measured (typically 9 mm).

Projects Related to Bone Strain

Subject-specific Finite Element Modeling

Finite element (FE) modeling is a technique that allows us to predict how complex structures behave under various mechanical loads. Bone is a complex structure whose behavior depends not only on bone geometry and size, but also the distribution of mineral within the bone. Just as people come in different sizes and shapes, so do their bone, and this can affect how their bone behaves as a structure. We use computed tomography (CT) to non-invasively measure bone structure, and these medical images also can be used to build FE models that are based on a person’s specific anatomy. Our modeling techniques have been validated with cadaveric testing and physical measurements such as fracture strength and surface strain. Using these models, we have found that individual variation in bone strain is more than three times larger than variation in bone density! This is important because it suggests that changes in bone density – a typical clinical outcome measure – may underestimate the change in bone mechanical behavior.

Multi-scale finite element modeling to apply physiologically relevant boundary conditions to micro-FE

Boundary conditions, the exact forces and constraints applied to a model, are an important determinant of predicted mechanical behavior. HR-pQCT data can be used as the basis for creating micro-FE models, in which microstructure is explicitly modeled. Typically, micro-FE models are highly detailed, but are subjected to very simple loading conditions such as a between-platen compression. However, we have shown that simplified boundary conditions to not result in physiologically relevant bone strain calculations. Therefore, we are working to develop multi-scale modeling methods to combine clinical-resolution and micro-FE models so that micro-structural behavior can be simulated when subjected to realistic in vivo loads.