Session Information
Session Type: ACR Concurrent Abstract Session
Session Time: 4:30PM-6:00PM
Background/Purpose: Evaluation of sacroiliac joints (SIJ) on pelvic radiographs according to the modified New York criteria (mNYc) is considered the gold standard for classification of axial spondyloarthritis (SpA). However, agreement among trained readers was consistently limited with kappa values around 0.5 [1-4]. The goals of this study in chronic back pain patients suspected of having SpA were: to determine the reproducibility of radiographic SIJ evaluation among 7 readers with varying levels of experience; to identify potential drivers to disagreement in classification among 5 predefined radiographic lesion types.
Methods: The study sample comprised 104 (38% male) consecutive patients aged 18-40 years with low back pain ≥3 months’ duration, who met the Assessment of Spondyloarthritis International Society (ASAS) definition for a positive SIJ MRI (n=92) or were HLA B27 positive (n=12) and had ≥1 SpA related clinical/laboratory feature according to the ASAS classification criteria for axial SpA. All readers (2 musculoskeletal (MSK) radiologists; 2 junior and 3 senior rheumatologists) were calibrated by reference images covering all mNYc grades. The rheumatologists additionally had 3 training sessions comprising independent evaluation of unrelated pelvic radiographs. 7 blinded readers classified pelvic radiographs according to the mNYc and then recorded 5 lesion types in both SIJ: erosion, sclerosis, ankylosis, joint space widening and narrowing. Reproducibility of the mNYc classification among 21 reader pairs was assessed by Cohen’s kappas and percent agreement. Potential drivers of disagreement were identified as proportions of concordant lesion types among patients with discordant classification. Finally, a generalized linear mixed logistic regression model was computed to explore to which extent discordance in lesion type was associated to discrepant mNYc classification.
Results: Kappas (percent concordance) for classification by mNYc were: 0.39 over 7 readers; 0.46 (79.8%) between 2 MSK radiologists; 0.55 (86.5%) and 0.36 (77.9%) among the most experienced rheumatologist and the 2 MSK radiologists.
Figure 1 Proportion of concordant lesion types among patients with discordant classification among 21 reader pairs
Table 1 Odds ratio for disagreement in classification in relation to lesion type among 21 reader pairs
|
Discordance in lesion type |
Odds ratio for disagreement in mNYc |
95% confidence interval |
P-value |
|
Erosion |
13.5 |
9.1 to 20.1 |
<0.0001 |
|
Sclerosis |
0.9 |
0.6 to 1.3 |
0.49 |
|
Ankylosis |
4.8 |
2.7 to 8.3 |
<0.0001 |
|
Joint space widening |
5.6 |
3.4 to 9.2 |
<0.0001 |
|
Joint space narrowing |
3.0 |
2.0 to 4.6 |
<0.0001 |
Conclusion: Reproducibility of SIJ classification by mNYc was fair to moderate among 7 readers with varying experience in assessing pelvic X-rays. Erosion was the main driver of discordant classification.
References: ARD 1987;46:139. ARD 2003;62:519. A&R 2012;64:1412. A&R 2014;66:2403.
To cite this abstract in AMA style:
Ashouri Christiansen A, Hendricks O, Kuettel D, Hørslev-Petersen K, Jurik AG, Nielsen S, Rufibach K, Loft AG, Juhl Pedersen S, Thuesen Hermansen L, Østergaard M, Arnbak B, Manniche C, Weber U. Evaluation of Sacroiliac Joint Radiographs in Patients with Chronic Low Back Pain: Is Erosion the Main Driver of Interreader Disagreement? [abstract]. Arthritis Rheumatol. 2015; 67 (suppl 10). https://acrabstracts.org/abstract/evaluation-of-sacroiliac-joint-radiographs-in-patients-with-chronic-low-back-pain-is-erosion-the-main-driver-of-interreader-disagreement/. Accessed .« Back to 2015 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/evaluation-of-sacroiliac-joint-radiographs-in-patients-with-chronic-low-back-pain-is-erosion-the-main-driver-of-interreader-disagreement/