Session Information
Date: Monday, November 9, 2015
Title: Imaging of Rheumatic Diseases Poster II: X-ray, MRI, PET and CT
Session Type: ACR Poster Session B
Session Time: 9:00AM-11:00AM
Background/Purpose: Conventional radiography is the main imaging modality to assess damage in the spine in ankylosing spondylitis. Modified Stoke Ankylosing Spondylitis Spine Score (mSASSS) has been endorsed by OMERACT as being more sensitive to change, therefore more commonly used in clinical trials. We aimed to compare the reliability of mSASSS on status scores by using different approaches to understand how different readers agree on overall scores and on the presence of individual findings included in mSASSS.
Methods: : The SPAR (The Spondyloarthritis Radiography) training module was used for the purpose of the study. After a standardization exercise, 4 investigators scored 68 plain radiographs of the cervical and lumbar spine of 34 patients at 2 time points with 2-year intervals. The anterior vertebral corners in lateral images were scored for erosions, sclerosis, squaring, syndesmophytes and ankylosis and a total mSASSS score was calculated. The reliability of the 4 readers was compared with 2 gold standard readers using ICC (intraclass coefficient). Additionally, the reproducibility of each finding in 1632 vertebral corners was calculated by kappa analysis and positive agreement rates.
Results: The reliability analysis using ICC revealed that all readers had excellent agreement with both gold standards with a range of ICC between 0.858-0.948. The kappa analysis showed worse agreement within lesions (table): For syndesmophytes agreement was between 0.394-0.677. The highest kappa values were detected for ankylosis, in a range between 0.624-0.889. The positive agreement rates showed that erosions were never detected at the same vertebral corner by 2 readers (positive agreement rate of 0%). The positive agreement rates for other score 1’s were also very low (sclerosis mean for agreement (range): 5.2% (0-15); squaring: 22.6% (0-59.7)). Ankylosis had the highest positive agreement rate, between 77.8 % (66.1-89.6%) and syndesmophytes had a mean positive agreement of 48.9% (39.4-63.9).
Conclusion: Our results show that there is a poor agreement on grade 1’s between readers which has a risk of increasing measurement error. The currently used definitions of reliability have a risk of overestimating the reproducibility (such as high ICC values) and the results of mSASSS should be interpreted with precaution.
|
vs gold standard |
Sclerosis |
Squaring |
Erosions |
Syndesmophytes |
Ankylosis |
||||||
|
kappa |
positive agreement |
kappa |
positive agreement |
kappa |
positive agreement |
kappa |
positive agreement |
kappa |
positive agreement |
||
|
Reader 1 |
I II |
0.554-0.680 0.525-0.623 |
0 7.5 |
0.462-0.575 0.33-0.584 |
0 0 |
0.416-0.501 0.649-0.728 |
0 0 |
0.463-0.535 0.394-0.416 |
39.4 28.3 |
0.664-0.681 0.690-0.739 |
71.4 75.1 |
|
Reader 2 |
I II |
0.678-0.690 0.676-0.684 |
0 13.4 |
0.594-0.612 0.634-0.649 |
13.7 10.8 |
0.483-0.510 0.796-0.808 |
0 0 |
0.621-0.649 0.580-0.592 |
56.5 51.1 |
0.765-0.776 0.814-0.842 |
80.4 85.2 |
|
Reader 3 |
I II |
0.675-0.697 0.645-0.672 |
0 15.0 |
0.625-0.628 0.590-0.681 |
27.3 21.2 |
0.705-0.726 0.486-0.517 |
0 0 |
0.570-0.580 0.495-0.547 |
51.1 48.1 |
0.624-0.684 0.637-0.666 |
68.7 66.1 |
|
Reader 4 |
I II |
0.713-0.736 0.682-0.730 |
0 5.6 |
0.696-0.711 0.689-0.771 |
59.7 48.7 |
0.492-0.503 0.785-0.877 |
0 0 |
0.670-0.677 0.599-0.604 |
63.9 52.4 |
0.801-0.802 0.839-0.889 |
86 89.6 |
Table: The kappa values and positive agreement rates (%), given separately for all lesions and all readers, compared to gold standards I and II.
To cite this abstract in AMA style:
Aydin SZ, Kasapoglu-Gunal E, Kurum E, Akar S, Mungan E, Alibaz-Oner F, Marzo-Ortega H, McGonagle D, Lambert RG, Atagunduz P, Maksymowych W. Assessing Radiographic Damage in Ankylosing Spondylitis: Variability in Reliability of the Modified Stoke Ankylosing Spondylitis Spine Score According to the Statistical Method [abstract]. Arthritis Rheumatol. 2015; 67 (suppl 10). https://acrabstracts.org/abstract/assessing-radiographic-damage-in-ankylosing-spondylitis-variability-in-reliability-of-the-modified-stoke-ankylosing-spondylitis-spine-score-according-to-the-statistical-method/. Accessed .« Back to 2015 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/assessing-radiographic-damage-in-ankylosing-spondylitis-variability-in-reliability-of-the-modified-stoke-ankylosing-spondylitis-spine-score-according-to-the-statistical-method/