Search

Background Joint inflammation is the common feature underlying juvenile idiopathic arthritis (JIA). Clinicians recognize patterns of joint involvement currently not part of the International League of Associations for Rheumatology (ILAR) classification. Using unsupervised machine learning, we sought to uncover data-driven joint patterns that predict clinical phenotype and disease trajectories. Methods and findings We analyzed prospectively collected clinical data, including joint involvement using a standard 71-joint homunculus, for 640 discovery patients with newly diagnosed JIA enrolled in a Canada-wide study who were followed serially for five years, treatment-naïve except for nonsteroidal anti-inflammatory drugs (NSAIDs) and diagnosed within one year of symptom onset. Twenty-one patients had systemic arthritis, 300 oligoarthritis, 125 rheumatoid factor (RF)-negative polyarthritis, 16 RF-positive polyarthritis, 37 psoriatic arthritis, 78 enthesitis-related arthritis (ERA), and 63 undifferentiated arthritis. At diagnosis, we observed global hierarchical groups of co-involved joints. To characterize these patterns, we developed sparse multilayer non-negative matrix factorization (NMF). Model selection by internal bi-cross-validation identified seven joint patterns at presentation, to which all 640 discovery patients were assigned: pelvic girdle (57 patients), fingers (25), wrists (114), toes (48), ankles (106), knees (283), and indistinct (7). Patterns were distinct from clinical subtypes (P < 0.001 by χ2 test) and reproducible through external data set validation on a 119-patient, prospectively collected independent validation cohort (reconstruction accuracy Q2 = 0.55 for patterns; 0.35 for groups). Some patients matched multiple patterns. To determine whether their disease outcomes differed, we further subdivided the 640 discovery patients into three subgroups by degree of localization—the percentage of their active joints aligning with their assigned pattern: localized (≥90%; 359 patients), partially localized (60%–90%; 124), or extended (, Using unsupervised machine learning, Rae Yeung and colleagues uncover data-driven joint patterns that predict clinical phenotype and disease trajectories in JIA., Author summary Why was this study done? Juvenile idiopathic arthritis (JIA) is a heterogeneous childhood disease for which joint inflammation is the common denominator. The current classification system, the International League of Associations for Rheumatology (ILAR) subtypes, categorize patients according to the number of active joints (four or fewer versus five or more affected joints in the first six months of disease) with little evidence to support this cut-off. Grouping frequently co-involved joints into joint patterns may help to better classify patients and predict disease course. What did the researchers do and find? We analyzed baseline joint involvement data from 640 treatment-naïve patients from 16 Canadian centers participating in the Research in Arthritis in Canadian Children, Emphasizing Outcomes (ReACCh-Out) study between 2005 and 2010. We identified seven joint patterns that grouped frequently co-involved joints using multilayer non-negative matrix factorization (NMF), an unsupervised pattern recognition technique. We named these patterns as follows: pelvic girdle, fingers, wrists, toes, knees, ankles, and indistinct. These seven patterns described more homogenous groupings of joints than the ILAR subtypes and further stratified them. Throughout disease course, we found that patients with joint involvements largely overlapping their baseline patterns, i.e., those having localized involvement, continued to have the same active joints, whereas patients with nonlocalized involvement did not. Patients with localized joint involvement reached zero joint involvement faster than those with nonlocalized involvement. What do these findings mean? Patterns of joint involvement represent prognostic features that should be incorporated into a comprehensive JIA disease classification. Patients with nonlocalized joint involvement are at high risk of poor outcome. Grouping patients by joint pattern and degree of localization may help clinicians tailor treatments based on predicted disease trajectories.