Leveraging eCAP thresholds in a logistic model to predict speech outcomes after implantation in anomalous cochlea

Background: Pediatric cochlear implant (CI) recipients with cochlear malformations face challenges due to variable speech recognition outcomes.

Aims/objectives: This study assesses the predictive value of intraoperative electrically evoked compound action potential (eCAP) thresholds, residual hearing, age at implantation, Intelligent Quotient (IQ), and malformation type for speech recognition outcomes.

Material and methods: A prospective cohort of 52 children (aged 1-4 years) with cochlear malformations who underwent CI between 2016 and 2024 was analyzed. Intraoperative eCAP thresholds were recorded, and disyllable word recognition scores (WRS) were evaluated 24 months postoperatively.

Results: Patients included incomplete partition (IP-II, 46.2%), cochlear hypoplasia (30.8%), IP-I (13.5%), and common cavity (9.6%). IP-II cases had higher mean eCAP thresholds (18.85 µV) compared to cochlear hypoplasia and common cavity cases. Higher eCAP thresholds correlated with better WRS outcomes (p < 0.001). Logistic regression revealed eCAP thresholds, residual hearing, age, IQ, and malformation type as significant predictors, with an AUC of 0.92.

Conclusions and significance: Intraoperative eCAP thresholds and residual hearing strongly predict auditory outcomes in children with cochlear malformations. Tailored CI strategies based on individual profiles can optimize speech recognition outcomes.