Knee osteoarthritis (OA) is a major cause of pain and disability in the U.S., and its prevalence is steadily increasing as the population increases in age and weight. KellgrenLawrence (KL) grading has long been the clinical standard for assessing tibiofemoral osteoarthritis severity from features such as joint space narrowing that are apparent on plain weight-bearing radiographs. The qualitative nature of KL grading introduces inter- and intraobserver variability, which along with challenges in interpreting a 2D projection of the 3D joint surface as seen on plain radiographs, limits the reliability for assessing joint space narrowing. Recognizing these limitations, a variety of MRI-based methods have been developed to evaluate OA severity, but MRI is expensive relative to radiography and usually obtained with a subject laying on a table (i.e., with the joint in a functionally dubious pose). More reliable and affordable metrics for measuring OA are clearly needed. The increased clinical use of weight bearing CT (WBCT) provides an opportunity to obtain more sensitive metrics of joint preservation, as it images three-dimensionally in a functionally relevant loaded pose.

The goal of this research is to develop an objective and fully automated method to quantify joint space narrowing from WBCT that allows for a reliable, robust assessment of the 3D joint space width (JSW). Based on previously developed semi-automated methods, we developed a fully automated process to directly map the tibiofemoral 3D JSW from WBCT images. The process involves analyzing image intensity profiles generated along surface normals of nominal subchondral bone segmentations. The reliability of the automated method was evaluated using existing test-retest scan data. The performance of the fully automated methods was further assessed by analyzing tibiofemoral 3D JSW in a larger collection of WBCT scans from subjects whose knees spanned a spectrum of OA severity