Dynamic Mechanics and Longitudinal Changes in Temporomandibular Joint Structural Integrity.
Authors
Affiliations (4)
Affiliations (4)
- Department of Oral and Craniofacial Sciences, School of Dentistry, Oregon Health & Science University, Portland, Oregon, USA.
- Department of Oral Diagnostic Sciences, University at Buffalo School of Dental Medicine, Buffalo, New York, USA.
- LMG Engineering GmbH, Zürich, Switzerland.
- Oregon Health & Science University-Portland State University, School of Public Health, Portland, Oregon, USA.
Abstract
Longitudinal changes in temporomandibular joint (TMJ) structures may reflect joint contact mechanics and jaw loading behaviours (Mechanobehaviour Score, MBS). Test if baseline MBS correlates with longitudinal changes in TMJ structures. According to Institutional Review Board oversight, diagnoses of adult subjects at baseline and follow-up were determined by Diagnostic Criteria for Temporomandibular Disorders (DC/TMD) via clinical examination, cone-beam computed tomography and magnetic resonance imaging. At baseline, subjects recorded in-field jaw muscle electromyography to calculate diurnal and nocturnal muscle duty factors (DF, %) and TMJ energy densities (ED, mJ/mm<sup>3</sup>) during jaw closing were estimated using computer-modelled joint loads and dynamic stereometry. These data were combined to produce baseline MBS (ED<sup>2</sup>●DF, (mJ/mm<sup>3</sup>)<sup>2</sup>%). General linear hypothesis, Tukey's Honest Significant Difference, and machine learning models tested MBS and component variables for baseline vs. follow-up differences in groups where TMJ integrity stayed the same (Group A), got better (Group B), or got worse (Group C). Forty-three subjects (29 female, 14 males) participated. The average time between baseline and follow-up was 8.0 ± 2.5 years. Analyses of 85 TMJs (43 right, 42 left) showed significantly smaller MBS (all p < 0.002) for Group B (n = 18; 71 ± 6 (mJ/mm<sup>3</sup>)<sup>2</sup>%) than Group A (n = 42; 197 ± 7 (mJ/mm<sup>3</sup>)<sup>2</sup>%) and Group C (n = 25; 650 ± 24 (mJ/mm<sup>3</sup>)<sup>2</sup>%). Group prediction accuracy by machine learning modelling was 0.96 (sensitivity, 0.98; specificity, 0.92) for the combination of TMJ stress-field aspect ratio, cartilage volume under the leading edge of the translating stress-field, stress-field translation distance, and velocity of stress-field translation. Mechanobehaviour variables accurately predicted changes in TMJ integrity.