Background/Purpose: The modified Stoke Ankylosing Spondylitis Spine Score (mSASSS) is commonly used to assess disease severity and progression in patients with AS; however, the reliability of scoring change in mSASSS is still challenging.  The SpondyloArthritis Radiography (SPAR) online training module is aimed at providing explicit instructions for scoring the mSASSS, reference cases in DICOM format, and raw expert reader scores aimed at training and calibration of readers for detecting change in mSASSS¹. However, it has not been validated beyond the group of experts that originally developed the module.

Methods: The SPAR module is comprised of 4 parts: 1) slide review of mSASSS scoring methodology, 2) reference DICOM image set of 34 cases with pairs of radiographs with two year time frames and a schematic of the spine for direct online data entry, 3) raw data scores for reference DICOM image set from two expert readers and unblinding data, and 4) schematic of the spine illustrating the mSASSS scoring module for calibrated reviewers to conduct direct online data entry for test cases. The details are available at http://tools.carearthritis.com/spar/. In exercise 1, 2 inexperienced readers independently assessed the 34 pairs of cervical and lumbar spine radiographs of the SPAR reference case DICOM image set blinded to time point (baseline, 2 years).  Exercise 2 was conducted after review of the SPAR module that included assessment of expert reader raw scores for the 34 case reference image set. Exercise 2 included 21 pairs of scans from exercise 1 that were randomly scored with an additional 34 new pairs of scans to assess consistency of reliability. This nested study design addresses the limitation for assessment of reliability posed by differences in progression rates between different study samples, which may significantly impact ICC scores when between-patient variability in progression is small. Inter-observer reliability of mSASSS for status and change scores was assessed by intra-class correlation coefficient (ICC). Inter-observer reliability of each type of lesion was assessed by kappa and percentage of agreement.

Results:

There was consistent reliability of mSASSS for status score and the reliability of the change score was improved in exercise 2 after reviewing the online module (Table 1). Improved reliability in detecting specific lesions was also observed.

	Exercise 1		Exercise 2
	All cases (n = 34)	Ex 1 Subset (n = 21)	All cases (n = 55)	Ex 1Subset (n = 21)	New cases (n = 34)
Status ICC (95% CI)	0.94 (0.88, 0.97)	0.97 (0.92, 0.99)	0.96 (0.92, 0.97)	0.97 (0.93, 0.99)	0.94 (0.89, 0.97)
Change ICC (95% CI)	0.38 (0.04, 0.63)	0.53 (0.13, 0.78)	0.67 (0.50, 0.80)	0.56 (0.19, 0.79)	0.71 (0.50, 0.85)

Table 1.

		Squaring		Sclerosis		Syndesmophyte		Ankylosis
	Cases	Kappa	Agreement	Kappa	Agreement	Kappa	Agreement	Kappa	Agreement
Exercise 1   (n = 34)	All	0.17	84.9 %	0.12	94.1 %	0.49	92.0 %	0.76	94.6  %
	N = 21*	0.09	78.7 %	0.11	91.2 %	0.51	90.3 %	0.83	95.0 %
Exercise 2   (n = 55)	All	0.65	90.9 %	0.47	97.6 %	0.59	90.7 %	0.87	97.0 %
	N = 21*	0.64	91.4 %	0.61	97.6 %	0.63	90.2 %	0.86	96.2  %

Table 2. *Cases scored in both exercises

Conclusion: The reliable scoring of the mSASSS by inexperienced readers can be improved by using the SPAR online training module.

1. Maksymowych et al. Ann Rheum Dis 2012. 71(Suppl3): 405

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/external-validation-of-the-spondyloarthritis-radiography-online-reference-module-for-training-readers-to-score-the-modified-stoke-ankylosing-spondylitis-score/