Background Musckuloskeletal ultrasound (US) complements clinical assessment of inflammatory arthritis, being more sensitive than x-ray in finding erosions and allowing assessment for subclinical inflammation. In practice not all rheumatologists are competent in or have access to US. In these circumstances US assessment is performed by radiologists and sonographers who, while being experienced in technique, do not have the detailed clinical context to form a final assessment the patient.
Objectives To compare sonographers' and rheumatologists' assessment of synovitis.
To assess how prior treatment affects US results.
To assess how US results inform a prior clinical judgement.
Methods Results from a 15 week period in 2012 were reviewed retrospectively. All scans of patients hands with or without other regions were included. Results were recorded as positive, negative or equivocal. Clinic correspondence was then reviewed to record examination findings prior to ultrasound, clinical impression (likely, unlikely or unsure), NSAID or steroid therapy prior to US and influence of result on subsequent management.
Results 96 scans were performed, including 50 new patients, 3 re-referrals and 43 follow up patients. Median time to US was 42 days (IQR 36-45). 15 rheumatologists requested US (range 1-21), which were performed by 14 radiologists/sonographers (range 1-28).
Results are tabulated below:
Agreement between rheumatologists who formed a clinical impression and radiologists/sonographers was fair (Kappa 0.3471). The US findings were similar between cases that were “unsure” and “unlikely” (Fisher's exact=0.698).
Overall, therapy at US did not appear to influence findings. However, in those patients where synovitis was thought likely or unsure, patients on no treatment appeared more likely to have a positive scan result, which was signficicant when compared to patients using NSAIDs (Fisher's exact=0.037) but not using steroids, likely due to low numbers.
Of 60 negative scans, all except 2 resulted in observation or de-escalation of treatment. Of 27 positive scans, 2 did not have escalation treatment. In these 4 cases clinical judgement and examination findings were trusted over US results. Of the 9 equivocal results, 6 had treatment escalated, while 3 were observed or had treatment withdrawn.
17 scans were performed when definite clinical synovitis was documented. Interestingly only 9 of these scans were positive. In the remainder, US results influenced management in 4 cases.
When the clinical impression was unsure, the 12 negative scans all resulted in observation or de-escalation of therapy, while 3 positive and 2 equivocal scans led to escalation.
Conclusions There is some agreement between a rheumatologists and radiologists/sonographers examinations, but frequent disagreement both when rheumatologists find synovitis and when they do not. US examination, being more sensitive than clinical examination, is expected to disagree in the latter case, but not the former. This suggests that either US negativity for synovitis does not definitely exclude synovitis, or that rheumatologists' assessment may at times overestimate disease.
A prospective blinded study could further investigate these issues. Larger numbers of patients would be needed to clarify if drug therapy influences US result.
Disclosure of Interest None declared