A 5-T MRI study has found that a specific segment of the cochlear nerve was smaller in ears affected by acquired sensorineural hearing loss than in control ears. The difference was seen in all three reported size measures at the cochlear aperture: its long diameter, short diameter and cross-sectional area. All three comparisons were reported as p < 0.001. The result shows an association between hearing-loss status and nerve size; it does not show that hearing loss caused the smaller measurements.
The researchers set out to determine whether 5-T MRI could visualize and measure the cochlear-aperture segment, compare nerve size between affected and control ears, and identify factors within the same person that were associated with segment size. The work was prospective and observational. Consecutive patients and healthy controls who underwent the scan were enrolled prospectively.
The scan showed a distinct pattern across nerve segments
At the cochlear aperture, the long diameter averaged 0.64 ± 0.11 millimetres in affected ears, compared with 0.73 ± 0.17 millimetres in control ears. The short diameter was 0.49 ± 0.09 millimetres versus 0.56 ± 0.13 millimetres. Cross-sectional area was 0.24 ± 0.09 square millimetres versus 0.33 ± 0.15 square millimetres. All three comparisons were reported as p < 0.001.
The images were also scored for how clearly the cochlear-aperture segment could be seen. Affected ears received an average score of 3.97 ± 0.84, compared with 4.18 ± 0.69 in control ears, with p = 0.047. The supplied analysis does not give a confidence interval for the score difference.
The pattern was not confined to that location. Affected ears also had smaller long diameter, short diameter and cross-sectional area at the internal acoustic meatus segment. The facial nerve's cross-sectional area did not show a statistically significant difference: 0.70 ± 0.20 square millimetres in affected ears versus 0.73 ± 0.23 in controls, with p = 0.320.
When the cochlear-nerve measurements were expressed as ratios to facial-nerve cross-sectional area, the cochlear-aperture-to-facial-nerve ratio was 0.38 ± 0.15 in affected ears and 0.47 ± 0.26 in control ears, with p = 0.008. At the internal acoustic meatus, the corresponding ratios were 0.60 ± 0.24 and 0.67 ± 0.36. That difference was not statistically significant, with p = 0.121.
A direct comparison of the two cochlear-nerve segments reinforced that uneven pattern. The cochlear-aperture-to-internal-acoustic-meatus ratio averaged 0.66 ± 0.20 in affected ears and 0.75 ± 0.35 in control ears, with p = 0.045. The authors interpreted the lower ratio as evidence of relatively greater atrophy at the cochlear-aperture segment. It remains an imaging comparison, not a clinical outcome measure.
A tightly specified imaging setup
Visualization assessment included 71 patients with acquired sensorineural hearing loss and 50 healthy controls, covering 242 ears. The measurement analysis included 66 patients and 49 controls, but the report gives conflicting control-ear totals: 98 in one passage and 109 in another. The later count includes 11 contralateral ears from people with single-sided hearing loss.
All scans were performed on a 5-T uMR Jupiter scanner with a 48-channel head coil. The protocol used a three-dimensional, heavily T2-weighted variable-flip-angle sequence, with an acquisition time of 7 minutes 21 seconds and interpolated isotropic resolution of 0.29 mm3. The study therefore reports results from this particular scanner and coil configuration at one institution.
Two blinded radiologists independently scored visibility and measured the long diameter, short diameter and cross-sectional area. One had seven years of experience and the other five; each repeated the measurements after two weeks. To account for measurements from the same person, group comparisons used a subject-level random-intercept mixed-effects model. Agreement was checked with weighted Cohen's kappa, Bland-Altman plots and intraclass correlation.
Agreement on the visibility scores was good. Cohen's kappa was 0.775 in patients and 0.752 in healthy controls, with reported intervals of 0.691 to 0.858 and 0.648 to 0.856. Measurement agreement was also reported as good.
A useful imaging signal, with important limits
Within-subject analyses looked at how differences between a person's ears related to the imaging measures, PTA and hearing-loss duration. The reported correlations for cochlear-aperture-related measures were negative, including coefficients of -0.268 (p = 0.030) and -0.256 (p = 0.038). Two PTA-related correlations were also negative, at -0.356 (p = 0.003) and -0.247 (p = 0.045). The analysis reported a moderate positive correlation between interaural PTA difference and hearing-loss duration, with a coefficient of 0.513 and p < 0.001.
These results are associations, not cause-and-effect evidence. The project was observational, and the within-subject correlations do not show that longer hearing-loss duration or worse PTA caused a smaller nerve segment. The study also did not test whether the measurements predict cochlear-implant speech or hearing outcomes.
Generalisability is also limited. The protocol was developed on one 5-T scanner and coil system at one institution, so it needs testing across other systems and settings. Two radiologists made the measurements, and automated segmentation was not used. The participants were cochlear-implant candidates with long disease durations and mostly progressive hearing decline, which narrows the relevance to broader populations with acquired sensorineural hearing loss.
The practical message is therefore focused. In this sample, 5-T MRI could visualize and measure the cochlear-aperture segment, and the size pattern pointed to relatively greater reduction there than at the internal acoustic meatus. Whether those measurements can guide prognosis or predict implant outcomes remains unanswered.
The project received partial support from the National Key Research and Development Program of China under grant 2023YFC2411100. The report says written informed consent and institutional review board approval 2024-P2-377-01 were obtained, and the authors declared no relevant company relationships. The manuscript was received on 7 January 2026, revised on 7 June and accepted on 8 July.
Paper data and sources
Original title: Atrophy of the cochlear aperture segment of the cochlear nerve in patients with severe-to-profound acquired sensorineural hearing loss: a 5-T MRI study.
Authors: Siwei Yang, Shengmei Ma, Yan Huang et al.
Journal/Repository: European radiology
Status: Peer-reviewed
First online: 2026-08-20
DOI: 10.1007/s00330-026-12808-w
Original paper · Full text