Author: Diego Benavent
Molina Collada et al (OP0171) examined the added value of FDG-PET/CT for detecting aortitis in patients with giant cell arteritis (GCA). They found that FDG-PET/CT can detect aortic involvement in one-third of patients with ultrasound proven GCA. However, half of the patients with ultrasound confirmed large vessel GCA may have negative FDG-PET/CT. Patients who were younger, female, and had a higher level of platelets were more likely to present aortitis by FDG-PET/CT.
Dumoulin et al (OP0261) analyzed 185 serial MRIs from clinically suspect arthralgia patients, to assess the order in which subclinical inflammation in different joint tissues progresses to RA. They found that increasing osteitis during the final 4-months before RA diagnosis followed a period of increasing tenosynovitis and synovitis during the 1-year to 4-months period before RA diagnosis. During inflammation resolution, tenosynovitis decreased first in the initial 4-months after symptom onset, preceding decreases in synovitis and osteitis during 4-12 months (p=0.02; p<0.01).
Hammer et al (OP0296) presented a two-year study of 209 patients with gout showing that the amount of monosodium uric acid crystal depositions in the joints assessed by ultrasound (US) and dual energy computer tomography (DECT) decreased during urate lowering treatment (p<0.001). The number of gout flares also reduced from a median of 2 (range 0-14) in the first year to flares, occurring in only 26% of patients in the second year. Baseline total US sum scores successfully predicted a flare during the initial 3 months (DC; p=0.031, β=0.113, total sum score; p=0.047, β=1.026), while DECT was able to predict a flare both at 3 and 12 months (p=0.014, β=1.086 and p=0.010, β=1.193, respectively).
Portier et al (POS0303) evaluated the impact of pregnancy on SIJ imaging in patients with early axial spondyloarthritis (axSpA), for which they analysed data from 381 women with early axSpA. Sacroiliitis on MRI and X-ray was more common in nulligravidae women as compared with women with past pregnancy (16.9% vs 9.9%, p = 0.05 and 33.8% vs 19.4%, p < 0.01, respectively). Among nulligravidae women, 38 (10%) had an incident pregnancy during the study, with a slight increase in the New York score only of the left SIJ, and an unexpected trend toward a reduction in MRI sacroiliitis and SPARCC score after pregnancy.
Li et al (OP0002) developed an artificial intelligence method for predicting early rheumatoid arthritis (RA) from extremity MRI scans, by using deep learning to analyze pre- and post-processed images. The system obtained a mean area under the receiver operator curve (AUC) for predicting RA of 0.683 in patients in an early arthritis clinic, and 0.727 in patients with clinical suspected arthralgia.
ABOUT THE AUTHOR
Diego Benavent
@DiegoBenavent
Diego is a consultant rheumatologist at Hospital Universitario La Paz in Madrid, where he is doing his PhD on axial spondyloarthritis.
He also collaborates as medical expert in Savana, a medical company working on artificial intelligence in medicine.
Diego is the chair of the Newsletter Sub-Committee