Background/Purpose: Axial spondyloarthritis (axSpA) is typically radiographically confirmed on the sacroiliac joint (SIJ) as assessed by the modified New York (mNY) criteria. In clinical trials, the mNY criteria is used to support determination of subject eligibility for study enrollment. Thus, it is important to analyze the mNY scores most likely to result in disagreement between reviewers and to better understand the impact on enrollment of the target axSpA population.

Methods: Data from central independent review of anterior-posterior (AP) pelvis radiographs in a double read model with adjudication for a total of 1,648 subjects from 3 clinical trials (phases 2 and 3) was analyzed for mNY +/- and individual left (L) and right (R) SIJ score discordance. The mNY scores were categorized per joint (R and L) as follows: 1) independent reviewers agreed on SIJ scores, 2) independent reviewers did not agree on individual joint scores, but mNY +/- was not impacted, and 3) independent reviewers did not agree and mNY +/- was impacted (i.e., adjudication triggered). To identify the score combinations which most often resulted in mNY +/- adjudication discrepancy, the frequency of each combination of scores that led to a discrepancy in categories 2 and 3 was determined.

Results: In 69% of the cases with individual score discrepancies per joint (R or L), the mNY assessment was not impacted; whereas, in 31% of these cases, adjudication was triggered (i.e., the mNY assessment was impacted by the R/L score discrepancy; Fig 1a). The score combinations most often associated with at least one joint for a timepoint where adjudication was triggered were: 1 vs. 2, 2 vs. 3, and 1 vs. 3. The score combination most often discrepant, but did not impact the mNY assessment (i.e., did not trigger adjudication) was 0 vs. 1. Overall, the score of 2 was associated with the most ambiguity.

Conclusion: Discrepancies in SIJ scores per joint (R and L) were not significant in approximately 2 out of 3 (69%) of the cases, considering impact on subject population enrollment. However, in approximately 1 out of 3 (31%) of the cases, the SIJ score discrepancies were associated with a disagreement in subject eligibility for study enrollment resulting in an adjudication.

Based on the data presented (Fig 1), a score of 2 was the most frequently involved score in inter-reader 1-step discordance. This may be related to the definition of a score of 2, which is open to a wider scope of interpretation. These results support the use of a double read and adjudication model for eligibility in clinical trials and potential use of other non-radiographic assessments (e.g. MRI) to improve accuracy of classification and appropriate inclusion of subjects into axSpA clinical trials.

« Back to 2019 ACR/ARP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/understanding-mny-radiograph-score-discordance-in-axial-spondyloarthritis-clinical-trials-using-imaging-criteria-for-subject-eligibility/