RESUMEN

PURPOSE: The objective of this study was to evaluate the computational reconstruction of the articular eminence of the temporomandibular joint (TMJ) based on magnetic resonance imaging (MRI) and semiautomatic volumetric segmentation techniques for morphological classification of the TMJ structure. MATERIALS AND METHODS: A total of 36 MRI scans of TMJ individuals were selected and formatted by using the ITK-SNAP software, consisting of MRI segmentation and generation of 3D models. The TMJ articular eminences were also classified according to the morphology analysis of the articular eminence in 3D reconstructions. Two independent trained and calibrated investigators performed the image analysis, which was repeated after thirty days. RESULTS: There was no association between sex and eminence shape (p = 0.456). Fisher's test revealed no statistically significant association between disc classification and eminence shape on both sides (p = 0.629). Chi-square test showed a significant statistically association between disc classification and disc displacement (p = 0.000). Intra- and interrater correlation coefficients showed excellent reproducibility values. CONCLUSIONS: Anatomical variability of the sample investigated was found, with predominantly round shape and presence of correlation between this shape and normal disc position. The correlation of flattened and convex shapes with disc position reduction indicated that type of disc derangement is more prevalent.