Biomechanics and Motion Analysis

Automated Digital Analysis of Joint Space Width in Knee Radiographs

Principal Investigator: Kenton R. Kaufman, Ph.D.

Figure 15: Typical program output displaying the edge detection (red) and the location of the minimum JSW (green) on both plateaus.

Radiographs are commonly used to confirm diagnosis and track the degeneration of cartilage in osteoarthritic patients. Specifically, the minimum joint space width (JSW) has been important in monitoring the severity and progression of osteoarthritis. While studies have looked at different methods of obtaining radiographic images to best analyze JSW, large variability has been reported in both intra- and inter-observer measurements of JSW. Therefore, the purpose of this study was to develop a computer program to automate the analysis of knee JSW.

The radiographs were analyzed using a custom program written in Matlab (MathWorks, Inc., Natick, MA) using Canny edge detection and an algorithm based on pixel intensity. The algorithm next calculated the JSW in both the medial and lateral compartments and determined the location of the minimum JSW for each (Figure 15).

Previously reported work has not reported comparison of automated measurements of JSW with manual measurements. The current analysis algorithm shows average mean JSW accuracy to within two pixels, with high reliability and reproducibility, making it a useful tool for measuring minimum JSW in longitudinal studies of OA using radiography.