Detecting radiological changes in rheumatoid arthritis that are considered important by clinical experts: influence of reading with or without known sequence

Abstract

Objective: To evaluate whether knowledge of the chronological sequence influences the sensitivity and specificity of the Sharp/van der Heijde (SvH) and Larsen/Scott (LS) scoring method to detect clinically important progression of joint damage caused by rheumatoid arthritis (RA) in the individual patient and assess whether scoring in chronological order leads to better sensitivity at the cost of lower specificity.

Methods: For both scoring methods, progression scores obtained with (chronological) and without knowledge of the sequence of the films (paired) were compared with the judgment of an international expert panel. This panel assessed whether progression of joint damage seen on films with 1 year intervals was clinically relevant (defined as progression of joint damage that would make clinicians change therapy). The applied thresholds for clinical relevance were (1) the progression scores with the highest accuracy by receiver operating characteristics analyses for the expert opinion, and (2) the smallest progression score that can be detected apart from interobserver measurement error by the scoring method, i.e., the smallest detectable difference (SDD).

Results: Progression scores that detected clinically relevant progression most accurately (chronological: 3.0 SvH units and 2.0 LS units; paired: 2.5 SvH units and 1.5 LS units) were smaller than the SDD (chronological 5.0 SvH units and 5.8 LS units; paired 13.8 SvH units and 9.7 LS units). With the SDD as threshold, the sensitivity to detect clinically relevant progression increased significantly from 20 to 60% for the SvH method and from 23 to 33% for the LS method if the sequence of the films was known. The specificity remained good when scoring chronologically: 88% for the SvH and 100% for the LS.

Conclusion: Our results suggest that knowing the chronological sequence leads to an increase in detecting clinically relevant changes in the patient without serious overestimation of nonrelevant differences. Analyzing a clinical trial should be done preferably by reading films in chronological order.