Poster Presentation 29th Australian and New Zealand Bone and Mineral Society Annual Scientific Meeting 2019

Does the degree of radiographic erosive disease affect the interreader agreement of erosion score using HR-pQCT? (#146)

Anne-Birgitte Blavnsfeldt 1 2 , Rasmus Klose-Jensen 1 2 , Jesper Thygesen 3 , Kresten Krarup Keller 1 4 , Ellen-Margrethe Hauge 1 2 , Bente Langdahl 2 5
  1. Department of Rheumatology, Aarhus University Hospital, Aarhus N, DANMARK, Denmark
  2. Department of Clinical Medicine, Aarhus University, Aarhus, Denmark
  3. Department of Procrurement andClinical Engineering, Aarhus University Hospital, Aarhus N, Denmark
  4. Diagnostic Centre, Silkeborg Regional Hospital, Silkeborg, Denmark
  5. Department of Endocrinology Aarhus University Hospital, Aarhus , Denmark

Background

Bone loss is common in rheumatoid arthritis.  High Resolution peripheral Quantitative Tomography (HR-pQCT) is used in osteoporosis research and also enables three-dimensional visualization of arthritic bone erosions at a high resolution. However, the degree of erosive disease may influence the reliability of HR-pQCT evaluation.

 

Purpose
To assess the inter-reader variability of identification of erosions and their distinction from cysts in the metacarpo-phalangeal (MCP) joints, using HR-pQCT in healthy controls and patients with Rheumatoid Arthritis (RA), stratified according to the conventional radiographic erosion score (Sharp- van der Heijde (SvdH)).

 

Methods

We analyzed HR-pQCT images from 77 patients with RA and 25 healthy controls. Patients were allocated to three groups of mild, moderate and severe disease according to SvdH score of MCP joints 2 and 3. The randomized HR-pQCT images were analyzed twice by three experienced readers, blinded to group distribution. The number of cortical interruptions according to the SPECTRA definition and their classification as either erosions  or cysts according to predefined criteria were recorded. Interclass correlation coefficients (ICC) for cortical interruptions, erosions and cysts were calculated for each group using a two-way random effects model.

Results

The ICC remained good to moderate for cortical interruptions and moderate for erosions. The ICC for the detection of cysts deteriorated with increasing degree of erosive disease. See table.

Conclusion

Reliability of the detection of cortical interruptions remains good to moderate. The distinction between erosions and cysts is challenging and should be further investigated.

 

5d1b289c652b4-table+icc+2.jpg