Background/Purpose:  The modified Stoke Ankylosing Spondylitis Spine Score (mSASSS) is used to assess progression in AS based on the lateral spinal radiograph but the methodology is not well standardized and the role of the lumbar antero-posterior (AP) radiograph is unclear. In the Spondyloarthritis Research Consortium of Canada (SPARCC) and the SPondyloArthritis Research and Treatment Network (SPARTAN) conjoint imaging group (SPAR) we aimed to develop and validate a reference image module to calibrate readers using the mSASSS.

Methods: The group readers comprised 5 rheumatologists and 3 musculoskeletal radiologists with special expertise in AS. We conducted the following: 1. Systematic review of the literature to identify aspects of the mSASSS requiring methodological clarity. 2. Independent assessment by 6 readers of baseline and 2 year radiographs from 25 patients with AS (exercise 1). 3. Consensus development of an imaging module (SPAR module) which clarifies definitions, scoring methodology, and a set of extensively annotated reference images. In exercise 2 the same 6 readers assessed radiographs from 39 patients with AS, which included 15 from exercise 1 (Subgroup 1), where baseline and 2 year radiographs were scored blinded to time point. Readers first scored only the lateral radiographs of the lumbar (LS) and cervical spine (CS), then only the AP radiograph, and then both radiographs simultaneously. Inter-observer reliability of the mSASSS was assessed by the intraclass correlation method (ICC). 

Results:

The first exercise demonstrated excellent reliability for status scores (ICC for 6 readers (range) = 0.92; Median (range) ICC for 15 reader pairs = 0.92 (0.84-0.96)) but poor reliability for change scores (ICC for 6 readers = 0.46; Median (range) for 15 reader pairs = 0.52 (0.11-0.66)). In particular, ICC for change score for the radiologist reading pair was only 0.46. In exercise 2, the ICC for change score for the radiologist reading pair improved substantially to 0.62 while improvement from 0.49 to 0.57 was also noted for the overall group in the subgroup of patients scored in both exercises. Reliability was not further enhanced when lateral and A-P radiographs were assessed simultaneously for either status or change mSASSS score. There was substantial variation between readers in the contribution of the AP radiograph to staging (mean (range) 27.2% (21.1-42.1)) and progression (mean (range) 15.4% (5.3-31.6)) and was most consistent for staging.

Conclusion: Calibration according to the standardized methodology developed for the SPAR module led to improved reliability in the scoring of the mSASSS even for expert readers. The contribution of the AP radiograph affects a substantial proportion of patients and requires further systematic study.

ACR Meeting Abstracts - https://acrabstracts.org/abstract/validation-of-a-reference-imaging-module-for-calibration-of-readers-scoring-with-the-modified-stoke-ankylosing-spondylitis-spine-score/