ABSTRACT
Magnetic resonance imaging is described as the reference standard for the evaluation of temporomandibular joint soft tissues; however, the literature shows conflicting results regarding the reproducibility of this method. Objective: this study aimed to assess the reproducibility of temporomandibular joint diagnoses using magnetic resonance imaging at 0.5 and 1.5 Tesla. Methods: a trained observer analyzed 212 temporomandibular joint images (134 at 0.5T and 78 at 1.5T) and diagnosed the presence or absence of nine conditions. Results: overall agreement was over 80% in both magnetic resonance units, with no significant differences (P > 0.05). Images at 0.5T and 1.5T provided excellent reproducibility for anterior disc displacement without reduction (κ = 0.82 and 0.80, respectively), hypermobility (κ = 0.84 and 0.90), and hypomobility (κ = 0.80 and 0.95), while fair to moderate values were obtained for anterior disc displacement with reduction (κ = 0.48 and 0.42) and disc shape changes (κ = 0.45 and 0.37). Conclusion: magnetic resonance imaging diagnoses at 0.5T and 1.5T presented good agreement. However, the lowest reproducibility for anterior disc displacement with reduction and disc shape change reveals the difficulty to diagnose these disorders.