Semiautomatic Quantification of Cochlear Nerve Cross-sectional Area Predicts Speech Intelligibility in Cochlear Implant Users

OBJECTIVE: To apply a semiautomatic method for quantifying cochlear nerve cross-sectional area (CSA) based on 3D models from MRI picture and to evaluate the association between CSA at different nerve locations and postoperative speech perception in cochlear implant (CI) users. STUDY DESIGN: Retrospective cohort study. SETTING: University Medical Center, tertiary academic referral center. PATIENTS: Twenty adult patients with asymmetric hearing loss undergoing a MED-EL cochlear implant (Synchrony 2, FLEX electrode series; MED-EL, Innsbruck, Austria) between 2020 and 2024. INTERVENTION: Preoperative MRI images were used for manual segmentation for the cochlear nerve (CN) and facial nerve (FN) followed by 3D reconstruction. CSA of the CN (ACN) and FN (AFN) was automatically quantified along the entire nerve course. Postoperative speech perception was assessed using word recognition scores (WRS) at the first fitting. MAIN OUTCOME MEASURES: Semiautomatic CSA measurements showed excellent agreement with manual measurements (RANC=0.95±0.06). CSA near the cochlear aperture exhibited the strongest correlation with postoperative speech perception. Peak correlation for polysyllabic tests at approximately 10% of the normalized nerve length (Rmax​=0.74) and for monosyllabic tests at approximately 24% (Rmax​=0.56). CONCLUSION: This study establishes a reliable semiautomatic method for cochlear nerve CSA quantification and identifies the nerve region near the cochlear aperture as a promising imaging biomarker for predicting postoperative speech outcomes in CI patients.