Listening to the Jaw: TMJ Acoustic Emissions as a Digital Biomarker of Juvenile Idiopathic Arthritis

Abstract

Temporomandibular joint (TMJ) involvement is often silent in juvenile idiopathic arthritis (JIA). Untreated, it can cause pain, growth disturbances, malocclusion, and skeletal deformities. Currently, combined clinical examination and radiographic imaging (i.e., magnetic resonance imaging, MRI) are necessary for diagnosis and treatment. However, imaging is not always feasible due to its high cost, access, occasional need for sedation and potential contraindications.1 Acoustic emissions (AEs) are the sounds produced from articulation in a joint.2 To measure the AEs of the TMJ, we built a custom headset with 2 uniaxial Dytran accelerometers that is placed over bilateral TMJs. The goal of this project was to determine if this approach for documenting TMJ sounds can differentiate between patients with JIA and healthy controls. This study was approved by the Georgia Institute of Technology and Emory University institutional review board (IRB #00081670). Subjects were included if they were 1) 18 years old or younger, 2) did not have any congenital craniofacial syndromes, 3) did not have a history of maxillofacial trauma, and 4) did not have ongoing fixed orthodontic appliances. Each subject wore the AE recording headset and performed 10 open/close (OC) exercises. For each subject, 49 features were calculated for each open/close cycle from the recording of the right and left TMJ AEs. Feature outliers were identified and removed if they were more than 3 median absolute deviations from that feature’s median. These features were fit to a logistic regression classification machine learning model. This model ranked each cycle of movement and classified it as either JIA or healthy. Healthy group (n = 28, 15 F) had average age of 13.5 ± 2.7 years and JIA group (n = 23, 18 F) had average age of 12.75 ± 2.3 years. According to AEs produced by TMJs and associated analysis, we were able to accurately classify a subject as having JIA or being a healthy in 71.3% of cases. The results of the study showed that the AEs of subjects with and without JIA have potential for being used as part of a diagnostic work-up. These early findings support the use of AEs as an additional non-invasive method to contribute in the assessment of TMJ dysfunction in children.