Toward Establishing a Standardized Magnetic Resonance Imaging Scoring System for Temporomandibular Joints in Juvenile Idiopathic Arthritis

Objective. The temporomandibular joints (TMJs) are frequently affected in children with juvenile idiopathic arthritis (JIA). Early detection is challenging, as major variation is present in scoring TMJ pathology on magnetic resonance imaging (MRI). Consensus-driven development and validation of an MRI scoring system for TMJs has important clinical utility in timely improvement of diagnosis and serving as an outcome measure. We report on a multi-institutional collaboration toward developing a TMJ MRI scoring system for JIA. Methods. Seven readers independently assessed MRI scans from 21 patients (42 TMJs, from patients ages 6-16 years) using 3 existing MRI scoring systems from American, German, and Swiss institutions. Reliability scores, scoring system definitions, and items were discussed among 10 JIA experts through 2 rounds of Delphi surveys, nominal group voting, and subsequent consensus meetings to create a novel TMJ MRI scoring system. Results. Average-measure absolute agreement intraclass correlation coefficients (avICCs) for the total scores of all 3 scoring systems were highly reliable at 0.96 each. Osteochondral items showed higher reliability than inflammatory items. An additive system was deemed preferable for assessing minor joint changes over time. Eight items were considered sufficiently reliable and/or important for integration into the consensus scoring system: bone marrow edema and enhancement (avICC 0.57-0.61, smallest detectable difference [SDD] 45-63% prior to redefining), condylar flattening (avICC 0.95-0.96, SDD +/- 23-28%), effusions (avICC 0.85-0.88, SDD +/- 25-26%), erosions (avICC 0.94, SDD +/- 20%), synovial enhancement and thickening (previously combined, avICC 0.90-0.91, SDD +/- 33%), and disk abnormalities (avICC 0.90, SDD +/- 19%). Conclusion. A novel TMJ MRI scoring system was developed by consensus. Further iterative refinements and reliability testing are warranted in upcoming studies.