Biomedical Engineering Projects

Creating 3D models of Bones from 2D X-ray Images
We have been developing methods for creating 3D models of bones from 2D X-ray images, and determining their position and orientation with respect to the X-ray sensor. This is needed for computer-aided orthopaedic surgery (CAOS) techniques, in which the computer guides the surgeon in placing surgical tools and verifying implant alignment and other critical factors. Patient-specific computer models are essential to this process, which are usually created manually from CT scan data. Creating the models automatically from one or two X-ray radiographs would be much faster, less expensive, and impose less radiation hazard to the patient.

Our method is based on fitting a deformable shape model of the bone to the X-ray image data. The deformable bone model is created from a collection of previously segmented models of the same bone from different patients, combined to form a statistical anatomical atlas.

Development of a Pelvic Reference Frame for Image-Guided Computer-Assisted Total Hip Arthroplasty
Correct orientation of the acetabular shell component within the pelvis is believed to be a significant factor in determining the risk of dislocation, aseptic loosening of the implant within the surrounding bone, component wear, and osteolysis in patients undergoing Total Hip Arthroplasty (THA). This study aimed to find a set of repeatably identifiable bony landmarks from two-dimensional digitally reconstructed radiographic (DRR) images that would define a novel pelvic reference frame to orient acetabular components during THA. Twenty orthopaedic surgeons identified twelve bony pelvic landmarks on DRR images taken of one pelvis from varying views. Analyses of standard deviation data from landmark location data recorded from these DRR images were used to determine the optimal landmarks and DRR views to create a new pelvic reference frame.

Implant Wear Simulation
The Computational Biomechanics Group studies the function of orthopaedic devices for joint reconstruction. Ultra-high molecular weight polyethylene is the most common bearing material in joint replacement implants, and polyethylene wear is still an important factor in the long term success of these devices. We use finite element methods coupled with other computational techniques to estimate the wear of poly bearings after years of exposure to activities of daily living. The goal of this research is to improve our understanding of the relationship between the subtleties of implant design and long term biomechanical function.

Methods to Track Artificial Joint Implant Components
We have developed methods to track artificial joint implant components in X-ray fluoroscopy video. With each image, we determine the 6 degree of freedom position and orientation (pose) of the artificial joint components. Our technique has been used to analyze the kinematics of several different knee types and to quantify the effect of several kinematic phenomena, including sliding and edge lift-off.

Modular Hip Implant CAOS System
The goal of this project is to develop an algorithm that can be used in the operating room to optimize the size and location of the neck during a total hip arthroplasty using Encore’s modular hip system. In using this algorithm, it will determine the best components to use in order to mimic a patient’s natural range of motion and leg length prior to surgery. Using this method will also limit the time taken in the operating room performing trial and error to determine neck sizing and placement.

Optimal Flexion Axis for Femoral Implant Positioning
The objective of this research was to compare and contrast axes used clinically for femoral implant positioning in total knee arthroplasty (TKA). Through this comparison, the goal was to ascertain which axis best represents the true functional flexion-extension axis of the knee and therefore the ideal axis to be used in TKA alignment of single radius of curvature implants. The four axes investigated include: the transepicondylar (TEA), posterior condylar (PCA), cylindrical (CA), and Whiteside’s line (also known as anteroposterior). Kinematic data for all six degrees of freedom were collected on nine cadaveric knees for both the transepicondylar and cylindrical axes. The relative position of all four axes were also determined on a CT based model for each of the nine knees. Each axis was then compared in the frontal and transverse planes and the angular variance was measured.

Spine Biomechanics
The Computational Biomechanics Group uses advanced finite element methods to study the mechanics of the human spine before and after surgical reconstruction. We use statistical methods to account for differences in patient anatomy as well as normal variability in surgical and medical device parameters. The long term goal of our spine modeling work is to improve clinical outcomes from spine arthroplasty procedures through a deeper understanding of the variables that affect clinical success.

Telemetric Hip Implant
For this project we are modeling an antenna to transmit data to the surface of the skin where it can be received and analyzed. The data the antenna transmits originates in sensors placed in the artificial hip socket and stem, where factors such as friction, stress, torque and position are measured. If a wireless system can be left in the body long term, the data can be used to personalize physical therapy and alert the patient when dangerous amounts of stress or torque are placed on the joint. Currently, antennas are used in total knee and hip implants, but a system has not yet been designed to carry a large amount of data out of the body.