Aim of this work is to establish evaluation criteria for identifying endolymphatic hydrops in the vestibule and cochlea using a magnetic resonance imaging (MRI) scanner. This is a retrospective diagnostic study. We evaluated 70 ears of 35 unilateral Ménière's disease patients. We performed 3-T MRI 4 h after intravenous gadolinium injection. Otologists manually traced the outline of vestibule, cochlea, and endolymphatic space of the vestibule and cochlea on two-dimensional fluid-attenuated inversion-recovery (2D-FLAIR) images. The traced area was measured, and rates of endolymphatic space to the vestibule and cochlea were calculated. The same otologists judged whether the low signal intensity area of the cochlea was at the edge of the cochlea. For measuring the rate of endolymphatic space to the vestibule, when the cut-off value was 30%, the presence of endolymphatic hydrops was determined with sensitivity of 87.1% and specificity of 94.3%. In contrast, the rate of endolymphatic space to the cochlea produced low accuracy. Therefore, when the presence of endolymphatic hydrops in the cochlea was judged by whether the low signal intensity area in the cochlea was at the edge of cochlea, endolymphatic hydrops could be detected with sensitivity of 91.4% and specificity of 94.3%. We were able to identify endolymphatic hydrops in the vestibule when the rate of endolymphatic space to the vestibule was greater than 30%, and could detect endolymphatic hydrops in the cochlea when a low signal intensity area was located at the edge of the cochlea in 2D-FLAIR images. Level of evidence 4.
Keywords: Electrocochleography; Endolymphatic hydrops; Magnetic resonance imaging; Ménière’s disease; Receiver-operating characteristic analysis.