Validation of inner ear MRI in patients with Ménière's disease by comparing endolymphatic hydrops from...

Intravenous gadolinium-enhanced inner-ear magnetic resonance imaging (IV-Gd inner-ear MRI) has been used to visualize endolymphatic hydrops (EH) in clinical diagnosis of Ménière's disease (MD). However, lack of histological validation has led to several concerns regarding how best to interpret the resulting images. Here, we compared hydropic changes in temporal bone specimens with the results of IV-Gd inner-ear MRI in patients with MD. Histopathologic images of temporal bones from 37 patients with MD and 10 healthy controls were collected from the National Temporal Bone Bank of the Massachusetts Eye and Ear Infirmary in the United States. The EH ratios in the vestibule and cochlea were calculated from temporal bones using the methods used for IV-Gd inner-ear MRI, and the degree to which the saccular and utricular hydrops contributed to vestibular hydrops was measured. The presence of hydropic change in each semicircular canal was assessed using temporal bone images and compared with IV-Gd inner-ear MRI scans of 74 patients with MD. Based on human temporal bone imagery, the EH ratios in the cochlea and the vestibule on the affected side were 0.314 and 0.757, respectively. In the healthy control group, the ratio was 0.064 for the cochlea and 0.289 for the vestibule; these values were significantly different from those for the affected side of MD patients. The values for the affected ear were similar to the ratios from the IV-Gd inner-ear MRI scans in MD patients. In the vestibule, saccular hydrops were more common than utricular hydrops. The average EH ratios in the saccule and utricle were 0.513 and 0.242, respectively. No significant hydropic change from each of three semicircular canals was evident in temporal bone histopathology. However, herniation of otolithic organs (saccule or utricle) into the lateral semicircular canal was found in 44.4% of the patients, with saccular herniation (24.8%) more common than utricular herniation (16.7%). Although IV-Gd inner-ear MRI might not reflect fully the results of actual histopathology due to the limited resolution of MRI and image-processing techniques, the measured EH ratios from temporal bone specimens and IV-Gd inner-ear MRI scans were similar. Hydropic change in the three semicircular canals was not significant at either the ampullated or nonampullated end. Canal invasion of vestibular hydrops seen on MRI also appeared in temporal bone histopathology, and saccular invasion was dominant.