Background/Purpose: The lack of tools to accurately measure and quantify joint inflammation is a challenging problem in the study of inflammatory arthritis. No gold standard method for identifying and quantifying the presence of inflammation exists.  The joint examination has relatively poor inter-rater reliability and ultrasound (US) can identify inflammation but is operator-dependent.  Joint inflammation can also be seen on 18F-Flurorodeoxyglucose (FDG) Positron Emission Tomography/Computed Tomography (PET/CT) but little is known about how this compares to other modalities.  The objective of this study was to examine the agreement among three methods of assessing joint inflammation: physical examination (swelling/tenderness), US (grayscale/Doppler), and FDG-PET/CT visual assessment.  We hypothesized that there would be good agreement among the measures.

Methods: Ten patients with active inflammatory arthritis underwent physical examination, US and FDG-PET/CT. Exclusion criteria included pregnancy, elevated fasting glucose (>150) or diabetes.  The examination was performed by an attending rheumatologist.  US was performed by an US-trained rheumatology fellow or a trained ultrasonography technician.  US images were read by a US-trained rheumatologist or musculoskeletal radiologist blinded to the clinical examination.  PET/CTs were read by a rheumatology fellow blinded to the examination and US results. Previously reported grading scales for US synovitis, doppler and PET/CT were used but for analysis these variables were converted to binary measures (yes/no inflammation) for analysis.  Agreement among the modalities was examined using Cohen’s kappa. Confidence intervals were generated using a bootstrapping method to account for clustering by individual.

Results: Four patients with rheumatoid arthritis and 6 with psoriatic arthritis were enrolled with mean age 52.7 years and 80% male.  All patients were on therapy at the time of the assessments and had relatively mild disease activity.  Among the 10 patients, 300 joints were included in the analysis (340 examined, 300 PET/CT, 180 US).  Of these, 46 were swollen, 50 were tender, 56 had FDG uptake on PET/CT, 26 had signs of inflammation on US grayscale and 14 were Doppler positive.  Agreement among the imaging modalities and examination was low (Table), with κ ranging from 0.01-0.22.  Restricting the analysis to only large peripheral joints did not substantially change agreement.

Conclusion: This is the first study to examine agreement among PET/CT, US and physical examination in the assessment of joint inflammation.  In this small pilot study, we found low agreement among PET/CT with US and both PET/CT and US with physical examination findings among patients on therapy with mild disease activity.    It may be that each modality measures a distinct property of inflammation, all equally valuable in measuring the construct of inflammation.