Limitation of the diagnostic value of MR images for diagnosing temporomandibular joint disorders

Abstract

Many studies have shown that MRI findings are reliable when experienced calibrated observers work as a group. The hypothesis was that MRI findings can be used as the gold standard also when evaluation is made by single expert observers. Temporomandibular joint (TMJ) MRIs of 34 patients were evaluated independently by four reviewers with expert knowledge of radiology for the presence of 13 specified pathologic entities, as well as the quality of the images, on a 5-step scale from "Sure Yes" to "Sure No". Intraclass correlation coefficients were calculated to estimate the rating reliability of the examiners. A coefficient of at least 0.8 was deemed good, between 0.60 and 0.80 was deemed acceptable, and less than 0.60 was considered poor. Additionally, weighted kappa statistics were used for pair-wise comparisons across all four reviewers. The hypothesis was not supported by the results. None of the 13 correlation coefficients for comparisons between single examiner evaluations of pathologic entities was good and 8 were poor. A diagnosis of TMD based on MRI examination protocols made by a single examiner should not be accepted as a gold standard with regard to TMJ disorders.