Reliability and accuracy of a method for measuring temporomandibular joint condylar volume

Abstract

The aim of this study was to develop and validate a technique for mandibular condyle segmentation and volume determination by using cone beam computed tomography (CBCT). A dry skull was used to generate 3 dimensional (3-D)-printed mandible models that were then imaged by using CBCT. Semiautomatic segmentation of condyles was completed. The Frankfurt plane was established and translated to the most inferior point of the sigmoid notch, and the condylar volume superior to the plane was determined. This procedure was repeated on 3-D-printed mandibles by using physical landmarks and the water displacement method to obtain the physical volume. This was repeated 3 times to evaluate reliability. Sensitivity analysis was performed to demonstrate the effect of discrepancies in locating landmarks in the Frankfurt plane. Condylar volume measurements obtained from CBCT were compared with physical measurements through repeated-measures analysis of variance (ANOVA) to determine accuracy. Condylar volume obtained from CBCT and physical measurements resulted in an intraclass correlation coefficient of 0.988 (0.918, 0.998) (P < .01) with both modalities, demonstrating excellent intrarater reliability. The mean difference of volume measurements between the modalities was not statistically significant (P=.365). Potential discrepancies in porion coordinates had minimal impact on condylar volume change. The condylar segmentation technique proved to be a reliable and accurate method for evaluating condylar volume.