LAUSR

Background Musculoskeletal ultrasound (MSUS) is a popular assessment tool, but its contribution to diagnostic practice over and above standard clinical and laboratory parameters has proved difficult to quantify. Objectives A published 7-joint ultrasound algorithm1 has been adapted for pragmatic application during 15 min screening appointments forming part of initial patient assessments in the Newcastle Early Arthritis Clinic (NEAC). Its additive contribution to diagnostic classification in this routine setting was appraised. Methods Detailed baseline clinical and laboratory parameters were recorded. Semi-quantitative MSUS scoring (0–3, grey scale and power Doppler) of the most symptomatic wrist (midline and ulnar dorsal longitudinal views), 2nd/3rd MCPs and PIPs and 2nd/5th MTPs (all longitudinal views) was recorded by sonographers, along with the ‘sonographer’s scan impression’ (‘definitely inflammatory,’ ‘possibly inflammatory’ or ‘non-inflammatory’). All MSUS findings were available to rheumatologist diagnosticians during subsequent consultations, and persistent inflammatory arthritis (PIA) was classified only where patients were started on ≥1 disease modifying anti-rheumatic drug (DMARD). Stepwise multiple logistic regression was employed to identify clinical variables that independently predicted IA diagnosis; the additive contribution of MSUS parameters to resultant models was assessed by comparing areas under receiver operator characteristic curves (AU ROCs).Abstract OP0151 – Figure 1 ROC curves depicting additive discriminatory utility of MSUS ‘scan impression’ with clinical parameters alone (red), with respect to PIA diagnosis amongst seronegative early arthritis patients. Δ AU AOC=0.08; p<0.001 Results 847 patients were enrolled (17% seropositive for anti-citrullinated peptide autoantibody, ACPA); final outcomes of PIA were recorded in 29% and 18% of the overall and ACPA-seronegative cohorts, respectively. SJC, CRP, age and ACPA status were non-redundant clinical/laboratory predictors of a PIA diagnosis by consulting rheumatologists in the overall cohort (AU ROC 0.85; 95% CI: 0.81 to 0.88), their discriminatory utility being diminished in the seronegative sub-cohort (AU ROC 0.78; 95% CI: 0.72 to 0.82). Although the additive contribution of summed parameters from the 7-joint MSUS algorithm to the model was statistically significant (p<0.001) it was numerically small (delta-AU ROC 0.03 and 0.05 in the overall and seronegative cohorts, respectively). The ‘sonographer’s scan impression’ was a potentially more useful diagnostic tool, its additive contribution to diagnostic outcome over clinical parameters alone being most evident in ACPA-negative patients where it increased the AU ROC by 10% (delta-AU ROC 0.08; p<0.001; figure 1). Conclusions In this large, un-blinded observational study, the clinical utility of a 15 min MSUS screen for diagnosing PIA requiring DMARDs was particularly evident amongst ACPA negative patients attending an EA clinic. Reference [1] Backhaus, et al. Arthritis Rheumatol61:1194–201. Acknowledgements National Institute for Health Research Newcastle Biomedical Research Centre Disclosure of Interest None declared