Intraoperative Visual Identification of Extra-Cochlear Electrodes in Cochlear Implantation Using Stimulation Current-Induced Non-Stimulating Electrode Voltages

OBJECTIVES: Extra-cochlear electrodes (EEs) are a frequent complication in cochlear implant (CI) surgery, potentially leading to reduced speech-perception outcomes. Standard intraoperative tools such as contact impedances and electrically evoked compound action potentials often fail to detect these misplaced electrodes. Stimulation Current-Induced Non-Stimulating Electrode Voltages (SCINSEVs), or electrical field imaging (EFI) in the advanced bionics device, are presented here as a faster and more accurate intraoperative measure for identifying EE. Adopting this method clinically could reduce reliance on intraoperative or postoperative imaging and allow EE identification in the critical window of surgery, when immediate correction can occur. DESIGN: This prospective intraoperative study tested adult CI candidates or users undergoing new or revision surgery at a single institution. EFI measurements were recorded during electrode insertion, where the surgeon was asked to hold a randomly selected number of 1 to 6 electrodes outside of the cochlea before full insertion. Visual confirmation of the number of EE electrode positions was achieved with the surgical microscope. Recordings were taken simulating postoperative conditions (with EE surrounded by blood or fascia). Three expert reviewers blinded to surgical outcomes reviewed EFI data to identify the number of EEs. Agreement between reviewers and surgeons was analyzed using correlations, receiver operating characteristic plots, and sensitivity and specificity calculations. RESULTS: General linear models revealed significant effects of the number of EEs not inserted by the surgeon (0 to 6 EEs) and the reviewer (1 to 3) on the EE ranking error, and no effects of the surgery type (revision or first-time cochlear implantation), nor of the material used (blood or fascia). A correlation was observed between the number of not-inserted EEs and EEs ranked based on the EFIs (r = 0.63, p < 0.0001). For identification of the presence or absence of EEs based on visual inspection of SCINSEVs, an overall sensitivity of 95.9% and a specificity of 40.7% were achieved (97.9% and 65.7% with ±2 EE reviewer-surgeon agreement tolerance). CONCLUSIONS: SCINSEVs are a rapid, reliable tool for intraoperative EE detection. With high sensitivity and moderate specificity, they offer a non-radiative alternative to postoperative computerized tomography, with results available during the surgical procedure. Use of this tool may improve CI outcomes through immediate surgical correction. The authors recommend training of relevant clinical staff to record and interpret these measurements for use within clinical practice.