Fully automated 3D joint space width analysis of the tibiofemoral joint from weight bearing CT

Knee osteoarthritis (OA) is a major cause of pain and disability in the U.S. The prevalence is steadily increasing as the population increases in age and weight. More reliable metrics are needed for measuring OA because current radiographic measures lack the sensitivity for tracking progression, and MRI-based metrics are relatively expensive and obtained in an unloaded joint position. The increased clinical use of weight bearing CT (WBCT) provides an opportunity for the development of new, more robust metrics because it images the knee three-dimensionally in a functionally relevant loaded pose.

This study aimed to develop a fully automated system to map joint space width (JSW) of the knee to use as a marker for OA. Prior studies presented semi-automated JSW measurement methods, but these were more time-consuming, and the requisite user interaction introduced inter- and intra-observer measurement variability. The reliability of the fully automated method was evaluated on repeat scans of the same subjects obtained within two weeks of each other. The feasibility of using the automated measurement of JSW to monitor OA was then analyzed in a large collection of WBCT scans from subjects whose knees spanned a spectrum of OA severity.