Background Bone erosions scored on classical 2D radiographs (CR) are a hallmark for diagnosis, disease prediction and evaluation of drug effects in patients with rheumatoid arthritis (RA). However, on a population level, more erosions are found using 3D imaging than with 2D CR.1
Objectives To compare the presence of bone erosions on CR with the presence of cortical breaks using high-resolution quantitative computed tomography (HRpQCT) in cadaver human fingers.
Methods Five human female cadaveric index fingers (aged 81,4 ± 11 years) were radiographed using CR and scanned with a HRpQCT device as used in vivo (Xtreme CT, Sanco, Swiss, spatial resolution 82 µm) and with µCT (µCT 80, Scanco, Swiss, spatial resolution: 18 µm). For each CT measurement, three regions of interest were measured; an 18 mm area covering the metacarpal phalangeal (MCP) joint and two 9 mm areas covering the proximal interphalangeal (PIP) and the distal interphalangeal (DIP) joints. The CR images were assessed by an experienced rheumatologist for the presence of erosions. In the CT data, cortical bone breaks with a diameter >0.5 mm were manually counted throughout the transversal CT slices of both the HRpQCT and the µCT images.
Results In total, 15 interphalangeal joints were assessed. Using CR, two bone erosions were found in two joints, whereas in HRpQCT images, 35 cortical breaks with a diameter >0.5 mm were counted in 13 joints (chi square 13.3, p<0.001 versus CR). Using µCT, a total of 46 cortical breaks >0.5 mm wascounted in 14 joints (chi square 16.2, p<0.001 versus CR, no significant difference versus HRpQCT).
Conclusions This pilot in vitro study in human cadaver finger joints shows that high resolution CT imaging enables to identify significantly more cortical breaks than CR. HRpQCT is therefore a more sensitive technique for the evaluation of cortical breaks in rheumatic diseases in interphalangeal joints. The results indicate that more small individual cortical breaks can be identified with µCT with respect to HRpQCT due to a higher resolution and less signal noise.
Stach et al. A&R, Vol. 62, No. 2, February 2010, pp 330–339
Disclosure of Interest None Declared