Background Joint space narrowing is an important outcome measure in rheumatoid arthritis, linked tightly to function and disability. High-resolution peripheral quantitative computed tomography (HR-pQCT) allows detection of bone margins with high precision. Based on this capability, we have devised a software script to quantify volumetric joint space width based on the method of “fitting maximal spheres”1. The position-dependence of this method needs to be determined to be reliably applied to longitudinal studies.
Objectives To assess the reliability of 3D volumetric joint space measurements with 5 degree variations in hand position, between 0 and 30 degrees of flexion, at the metacarpophalangeal (MCP) joints.
Methods Six cadaver hands were imaged with HR-pQCT, with image acquisition of the 2nd and 3rd MCP joints (n=12 joints) using a SPECTRA-endorsed imaging protocol2. Using a positioning device, the MCPs were placed at 7 different angles of flexion (0, 5, 10, 15, 20, 25 and 30 degrees), and actual angles were verified from the acquired images in the y-z plane post-hoc. Descriptive statistics were used to calculate the mean, median, minimum, and maximum joint space widths and total volume measurements, by degree of angulation. The root-mean-square coefficient of variance (RMSCV) was calculated for individual joints to determine the variability caused by angulation.
Results The differences in positioned vs measured angle were small for 0, 5, and 10 degrees, but significant for >15 degrees of flexion, with the actual measurement ranging from 20.8 to 32.5 at 30 degrees. Despite this, the volumetric joint space width measurement gave robust results in line with accepted metrics for HR-pQCT, with a RMSCV for mean joint space width of 6.2% (SD 2.8), median joint space width of 6.5% (SD 2.4), maximum joint space width of 5.3% (SD 2.9), and 5.8% (SD 3.6) for total volume.
Conclusions Our 3D volumetric measurement method of joint space width for images acquired with HR-pQCT technology is reliable for MCP flexion angles of 0 to 30 degrees. Performance is optimized between 0 and 10 degrees of flexion, and care should be given to remaining in this parameter for longitudinal or repeated measures studies.
Barnabe C, Buie H, Kan M et al. Med Eng Physics 2013;35:1540-1544.
Barnabe C, Feehan L. J Rheumatol 2012;39:1494-5.
Disclosure of Interest None declared