Author: Victoria Konzett
Di Matteo et al. (OP0237) investigated the utility of the OMERACT ultrasound elementary lesions for enthesitis in spondyloarthritis (SpA) in a multicenter cohort of 413 patients with psoriatic arthritis or axial SpA, as well as 282 controls with osteoarthritis or fibromyalgia. They found varying discriminatory potential for the different features, with highest associations observed for power doppler signals and bony erosions, especially in the achilles tendon.
Hügle et al. (OP0112) presented a deep learning model for automated detection of calcium pyrophosphate deposition on hand radiographs. The authors trained a convolutional neural network in a dataset of 926 labeled radiographs, and achieved sensitivity and specificity of 0.86 and 0.72, respectively, indicating potentials of the model as screening tool for calcium deposition in large-scale records or databases.
Two-year data from the SPACE (SPondyloArthritis Caught Early) cohort, a multicenter cohort of 548 patients <45 years of age with chronic back pain (≥ 3 months and <2 years) was presented by Marques et al. (OPO310). In this study, half of the patients with chronic back pain referred to rheumatologists, and especially those with imaging features on MRI, HLA-B27 positivity and peripheral arthritis, received a diagnosis of axial SpA within 35 months from symptom onset.
Ramming et al. (OP0075) showed fibroblast activation protein (FAP) uptake in early stages of inflammatory arthritis, therefore indicating an important role of fibroblasts in early disease pathogenesis, as well as a potential use of FAP-imaging in early disease detection.
Muscle ultrasound was investigated by Galuzzo et al. (OP0024) in a prospective, single-blind, monocenter cohort of 40 patients with inflammatory myopathies. The authors demonstrated strong agreements between US and MRI, warranting the use of US as readily-available and cost-effective imaging modality, albeit limitations were detected in the specific differentiation of various myopathy subtypes.
Shamonin et al. (OP0190) trained a deep learning model to automatically quantify MRI based detection of wrist tenosynovitis, synovitis and bone marrow edema in 28 rheumatoid arthritis patients. The model was fast compared to visual scoring, and performed well for tenosynovitis and synovitis detection, but needs further improvement for detection of bone marrow oedema.
Molina-Collada et al. (OP0294) investigated the usability of the OMERACT Giant cell arteritis Ultrasonography Score (OGUS) for flare prediction after initial treatment response. In a retrospective analysis of 76 patients, significant OGUS improvements were found in non-relapsing patients through month 6, while the absence of US improvements was associated with non-response or subsequent flares in this study.
ABOUT THE AUTHOR
Victoria Konzett
Victoria is a resident for internal medicine and PhD fellow at the Division of Rheumatology at the Medical University of Vienna, Austria.
Her major research interests are clinical and translational research projects in rheumatoid and psoriatic arthritis, where she focuses mainly on epidemiology, outcomes, research and predictive modelling, including the use of imaging in this context.
Victoria is a member of the Newsletter Sub-committee and the Young Division of the Austrian Society of Rheumatology.