Treatment of patients with rheumatoid arthritis (RA) is rarely personalized, since predictors of disease course are lacking. The
severity of RA can be measured objectively by radiographic progression. The most reliable way to measure radiographic progression is in
a longitudinal cohort with serial time points, scoring on a quantitative scale, with a validated scoring method and trained readers.
Current models used to predict radiographic progression are based on C-reactive protein and anti-citrullinated protein antibodies. Other
biomarkers could increase the prognostic ability of these models. In this review, we evaluated the published (and partly nonpublished)
data on genetic, serologic, and imaging biomarkers for the severity of joint destruction in RA.
We evaluated variants in 10 genes (CD40, IL2RA, IL4R, IL15, OPG, DKK1, SOST, GRZB, MMP9, and SPAG16). In 5 variants (IL2RA,
DKK1, GRZB, MMP9, and SPAG16), we found evidence of an association at the functional level. We evaluated several serological biomarkers,
namely, autoantibodies (RF, ACPA, anti-CarP), markers related to inflammation (ESR, CRP), and proteinases or components of
the extracellular matrix of bone and cartilage (MMP3, CTX-I, CTX-II, COMP, TIMP1, PYD, RANKL/OPG, CXCL13). Finally, we
evaluated markers that can be visualized by ultrasound or MRI, including erosions, bone marrow edema, synovitis, and tenosynovitis.
Several studies showed that bone marrow edema and synovitis on MRI are robust predictors of radiographic progression. Some studies
showed that inflammation detected with ultrasound predicted radiographic progression.
Future studies will reveal whether adding and combining all these different biomarkers will increase the accuracy of risk models predicting
radiographic progression in RA.