The overall goal of the virtual functional anatomy (VFA) project is to greatly advance the clinical diagnosis and treatment of musculoskeletal impairments as they relate to joint function. The current focus is to develop and ultimately validate a combined set of tools (VFA toolbox) that will enable the accurate and precise measurement, analysis and visualization of three-dimensional (3D) static and dynamic musculoskeletal anatomy (i.e., bone shape, skeletal kinematics, tendon and ligament strain, muscle force, and joint space). We plan to merge and extend our existing dynamic MRI imaging and analysis capabilities for the development and implementation of a highly accurate, imaging-based measurement and analysis technique for the non-invasive quantification of complete joint anatomy and tissue dynamics during functional movements. In short, we plan to develop a method for creating 3D digital images of loaded and moving joint tissues (bone, cartilage, and connective tissues) that reveal joint contact patterns and tissue loads. In conjunction with building this tool, we will evaluate the variability of bone shape across subjects, the sensitivity of defined joint posture (translation and rotation of one bone relative to another) to osteo-based coordinate system definition, and the ability to ultimately use these tools to document and evaluate the function of normal and impaired joint under simulated conditions experienced during activities of daily living. Over the past year our efforts have focused primarily on developing the backbone for the VFA toolbox and began to explore the issues surrounding the dynamic MR scanning of muscle and tendon. The key focal points for the algorithm development were the image registration process along with continued progress in improving the integration algorithms. An overview of this entire project was given at National Institutes of Health Research Festival (October 2003). An exciting milestone was the completion of a preliminary registration algorithm. Dynamic MRI can provide 3D musculoskeletal kinematics information, but this information cannot be readily applied to 3D models of the bones, created from static high-resolution scans of the joint being studied (e.g., knee and ankle). In order to apply the kinematics from the dynamic MRI to the static models, the static images must be ?registered?, the word used to refer to the alignment of images, to one of the 24 discreet positions of the dynamic MRI data. Visualization is made possible by programs that have been written in-house using Matlab?s scripting language. Programs have also been developed for extracting 3D kinematics from image registration results. The image registration methodology has been validated by depending on the redundancy of information generated when the static images are repeatedly registered to each of the 24 dynamic positions. Registration has demonstrated an accuracy that ranges from 1.3 mm for the translations of the femur to 4.3 mm for the translations of the tibia and from 1.5? for the rotations of the femur to 7.2? for the rotations of the patella. This work was presented at The International Society of Biomechanics? Eighth International Symposium on the 3-D Analysis of Human Motion (Tampa, FL. April 2004). In addition, work is ongoing in the area of improving the integration algorithms. To this end, we have created numerous simulated data sets in which to test the fidelity of the new algorithms. Plans are being developed to create a motion phantom that will allow us to directly test the accuracy of these algorithms. Additionally, we overhauled the kinematics analysis algorithms. These algorithms use the output from the integration routines in order to calculate rigid body (e.g., patellar, femur and tibia) attitudes throughout the motion cycle. This process has enabled us to correctly handle oblique images, improved the compatibility between the kinemmatics analysis outputs and the registration outputs, and enhanced the user interface of the VFA toolbox. The MR data collection continued throughout the year with a key focus on developing the imaging protocols needed to quantify forces in the quadriceps muscles, patellar tendon, the anterior cruciate ligament and the cartilage during an extension/flexion cycle of the knee joint. Since the forces in the muscles are being calculated by measuring the strain in the tendons, it is imperative that errors be minimized during this measurement. Thus, we have decided to maximize the strain within the tendons by maximizing the load being raised in extension, through the use of non-magnetic ankle weights. We are currently conducting a study to test the maximum weight that can be used without disrupting the repeatability of the motion. Lastly, we have established the required image plane location and orientation needed to capture the insertion of each quadriceps muscle into its tendon. In total 25 subjects were seen under this protocol in the last year. Another key area in which advances were made was in quantifying the moment arm of the patellar tendon. A required element of this was developing algorithms to define the finite helical axis (FHA). Thus far we have analyzed data from 5 subjects and are planning on continuing the analysis for all available data sets (currently there are 17). All 5 subjects had similar values for the patellar tendon moment arm, which increased in value as the knee extended. In all subjects the FHA shifted posteriorly and slightly inferior, relative to the femoral origin, during extension. The screw home mechanism (sharp external rotation in terminal extension) was evident at maximum extension for all subjects reaching a minimum of 6? of knee extension. This work has been submitted the 2005 meeting of the Orthopeadic Research Society. In order to define a force model for the knee joint, the contact force on the cartilage must be quantified. Thus, work has begun in modeling the cartilage and its contact with an opposing surface. As this issue was investigated the more complicated issue of defining the shape of an object (e.g., cartilage, bone) came to light. Thus, a study into mathematically quantifying bone shape was begun. A secondary area of study was the ankle joint. Using the tools currently developed for the VFA, we looked at the kinematics of the hindfoot (tibia, talus and calcaneous). To date 8 subjects have been studied and the data from these studies demonstrated that talus and calcaneous do not move as a single rigid body as often modeled. Two invited presentations were given on this topic (American Physical Therapy Association Foot and Ankle Research Retreat II, April 2004 and the New Horizons in Clinical Treatment and Innovative Technology Conference, November 2003). An additional presentation was given at the American Society of Biomechanics (September 2003). The papers from the project that evaluated the precision of defining anatomically based patellofemoral coordinate systems have been finalized. One has been published (Journal of Magnetic Resonance in Imaging) and one has been conditionally accepted and is in final review (Clinical Orthopeadics and Related Research).
| Sheehan, Frances; Derasari, Aditya; Fine, Kenneth et al. (2009) Q-angle and J-sign: Indicative of Maltracking Subgroups in Patellofemoral Pain. Clin Orthop Relat Res : |
| Sheehan, Frances T (2007) The 3D patellar tendon moment arm: quantified in vivo during volitional activity. J Biomech 40:1968-74 |
| Shibanuma, Nao; Sheehan, Frances T; Stanhope, Steven J (2005) Limb positioning is critical for defining patellofemoral alignment and femoral shape. Clin Orthop Relat Res :198-206 |
| Shibanuma, Nao; Sheehan, Frances T; Lipsky, Peter E et al. (2004) Sensitivity of femoral orientation estimates to condylar surface and MR image plane location. J Magn Reson Imaging 20:300-5 |
| Rebmann, Andrea J; Sheehan, Frances T (2003) Precise 3D skeletal kinematics using fast phase contrast magnetic resonance imaging. J Magn Reson Imaging 17:206-13 |
