Evaluating the Safety and Efficacy of Robot-Assisted Cochlear Implant Electrode Array Insertion

Abstract

Importance

Robotic-assisted cochlear implantation has the potential to reduce surgical variability, enhance insertion precision, and minimize intracochlear trauma; however, real-world clinical evidence remains limited.

Objective

To evaluate the safety, accuracy, and early hearing outcomes of robot-assisted cochlear implant (CI) electrode array insertion using the OTODRIVE^® system.

Design

Retrospective cohort study.

Setting

A tertiary academic CI center.

Participants

Adult patients who underwent robot-assisted cochlear implantation between the dates of February 2025 and August 2025.

Intervention or Exposures

Robot-assisted electrode array insertion with preoperative planning using the OTOPLAN software.

Main Outcome Measures

Intraoperative safety outcomes; audiologic performance, including unaided and aided pure-tone averages (PTAs) and AzBio sentence scores; and imaging-based electrode insertion accuracy, including angular insertion depth (AID) and electrode contact insertion on postoperative cone-beam computed tomography.

Results

The review identified 39 patients, mean age of 59.5 ± 19.2 years, with 59% male participants. The most common etiology of hearing loss (HL) was idiopathic sudden sensorineural HL. Mean total surgical time was 122.2 ± 49.4 minutes. No intraoperative complications or robotic-related adverse events occurred. Audiologic outcomes demonstrated significant improvement, with a mean preoperative unaided PTA of 92.4 ± 16.7 dB HL, improving to a postoperative aided PTA of 31.6 ± 6.3 dB HL. Mean AzBio sentence recognition scores increased from 18.1% preoperatively to 75.7% postoperatively, representing a 57.6% absolute improvement. Mean cochlear duct length was 34.6 ± 1.6 mm, and the planned versus achieved AIDs were 582.7° ± 35.2° and 569.0° ± 38.9°, respectively. Full insertion was achieved in 89.7% of implanted ears, with no tip fold-overs or electrode malposition identified.

Conclusions

Robot-assisted cochlear implantation using the OTODRIVE^® system was safely integrated into clinical practice in this cohort, with reproducible electrode insertion and early postoperative audiologic outcomes.

Relevance

These findings support the feasibility and safety of integrating robotic systems into routine CI surgery.

Graphical Abstract

Key Messages

• Robot-assisted cochlear implantation electrode insertion can be performed safely with no observed robotic-related intraoperative complications.

• Robotic insertion enables accurate and consistent electrode placement, closely matching preoperative insertion planning.

• Early postoperative hearing outcomes following robotic-assisted implantation are robust and clinically meaningful.

Introduction

The cochlea is a spiral-shaped organ within the inner ear where acoustic vibrations are decomposed into different frequencies, with high frequencies represented at the base of the cochlea and low frequencies represented at the apex. ¹ These frequencies are then transduced by sensory cells to create electrical signals that are transmitted to the brain via the auditory nerve. When the cochlea is not functioning properly, as in the case of sensorineural hearing loss (HL), devices such as hearing aids or cochlear implants (CIs) can be used to restore sensation of sound. Hearing aids are typically used in cases of mild-to-moderate HL while CIs are used for severe-to-profound HL. A CI is a neural-prosthetic device that consists of an electrode array that is inserted along the length of the cochlea to bypass the damaged portion of the inner ear and electrically stimulate the auditory nerve directly.

While successful clinical outcomes are achieved for most CI recipients, audiological performance and user satisfaction vary based on factors such as age at implantation, duration of HL, and etiology of HL, as well as factors related to the cochlear implantation procedure itself. For example, it has been reported that deeper CI electrode array insertions are associated with improved outcomes, and insertional trauma should be minimized to preserve the patients’ residual hearing. ² Advancements in the preoperative, perioperative, and postoperative cochlear implantation procedure have led to improvements in outcomes. Patient-specific preoperative planning and postoperative programming, which account for individual anatomical variation, have been shown to improve outcomes.^1,3-7 Atraumatic surgical techniques and the use of flexible electrode arrays can minimize operative trauma and preserve residual hearing, while longer arrays can achieve increased angular insertion depth (AID) and tonotopic coverage.^8,9 Despite advancements, atraumatic placement of the CI electrode array remains difficult, and CI outcomes are subject to inter-surgeon variability. For example, studies have shown that the speed and force with which the surgeon inserts the electrode array into the cochlea can have implications on postoperative outcomes ¹⁰ and maintaining the optimal insertion speed is beyond human capability. ¹¹ Moreover, with expanding selection criteria, preservation of residual hearing by minimizing intracochlear trauma has become increasingly important. ¹²

Robotic-assisted measures for electrode insertion in CI surgery have been developed to minimize inter-surgeon variability, improve precision and safety, as well as overcome the limitations of manual dexterity.^8,13,14 After surgical exposure of the round window, surgeon-controlled robots can insert the electrode array into the cochlea at a slow, consistent speed. Postoperative imaging can then be used to confirm the accurate placement of the electrode. ² Systems currently in clinical use include RobOtol^® (Collin, France), iotaSoft^® (iotaMotion, Inc, Iowa City, IA, USA), and OTODRIVE^® (CASCINATION AG, Bern, Switzerland and MED-EL, GmbH, Innsbruck, Austria). ² In a retrospective review, the RobOtol system was found to be associated with significantly fewer scalar translocations than manually-inserted implantations. ¹⁵ The iotaSoft system was assessed in a cadaveric study, showing significantly less trauma, lower insertional force, and reduced variability than manual methods. ¹⁶ The OTODRIVE system has been evaluated both clinically and experimentally, showing significantly slower and more controlled atraumatic insertions, ¹⁷ significantly reduced force variation than manual techniques, ¹⁸ and a tendency toward improved hearing preservation and subjective benefit. ¹⁹ The safety and efficacy of the OTODRIVE system was evaluated in a cadaveric study using 15 human temporal bones, showing expected cochlear coverage (74% ± 7%) and a low average trauma score of 0.27. ¹⁷

The OTODRIVE robotic-assisted system has recently been integrated into the standard of care cochlear implantation procedure at London Health Sciences Centre (LHSC; London, ON, Canada). The OTODRIVE system is used together with the OTOARM, OTOARM Aligner, Connector OD, and Forceps OD, as illustrated in Figure 1. OTOARM is an electro-mechanical arm, which mounts directly on the surgical bed and can be flexibly oriented to provide coarse positioning. The OTOARM Aligner connects to the end of the OTOARM and provides precise movement along 5 degrees of freedom. The OTODRIVE system attaches directly to the OTOARM and OTOARM aligner. The sterile Connector OD and Forceps OD attach to the OTODRIVE and are used to mount the electrode array to the system. Together, the OTOARM and OTOARM Aligner are used to set the trajectory of the Forceps OD toward the round window prior to insertion. ¹¹

Figure 1.

Schematic overview of the robotic-assisted cochlear implantation setup adapted from MED-EL (Austria).

The purpose of the present study was to evaluate the safety and efficacy of robotic-assisted cochlear implantation using the OTODRIVE system through a retrospective review of the initial cohort of clinical cases performed at LHSC.

Materials and Methods

Study Design and Setting

A retrospective cohort study was conducted at the LHSC—University Hospital CI Program. Adult patients who underwent robot-assisted cochlear implantation between February 2025 and August 2025 were identified for inclusion. The study was approved by Western University’s Health Sciences Research Ethics Board (approval #127817).

Patient Selection

Eligible participants were adults (≥18 years) who underwent cochlear implantation with robot-assisted electrode array insertion using the OTODRIVE system. All patients met institutional candidacy criteria for cochlear implantation, which generally includes residual low-frequency pure-tone thresholds of ≤70 dB HL at 125 and 250 Hz and ≤90 dB HL at 500 Hz, as well as best aided preoperative speech recognition scores of ≤60% for monosyllabic words [consonant-nucleus-consonant (CNC)] and ≥10% for open-set sentence recognition (AzBio). Patients were required to be medically fit for surgery and without cognitive impairment precluding implantation. All consecutive adult cochlear implantations performed during periods when the robotic system was available were included. Cases in which the robot was unavailable were excluded. No patient selection was performed based on clinical characteristics or anticipated surgical difficulty.

Preoperative Imaging and Surgical Planning

All patients underwent preoperative cone-beam computed tomography (CBCT) imaging as part of routine clinical care. Preoperative imaging was imported into OTOPLAN (CASCINATION AG and MED-EL), dedicated imaging and planning software used to model cochlear anatomy as well as the insertion of various electrode arrays to target desired insertion depths. ²⁰ In all available cases, OTOPLAN was used to measure the predicted AID of the inserted electrode array.

Surgical Technique and Robotic-Assisted Insertion

All procedures were performed under general anesthesia by fellowship-trained otologic surgeons using a standard mastoidectomy with facial recess approach to access the round window. To implement the robotic-assisted insertion, the OTOARM system was first attached to the surgical bed, and then the OTOARM Aligner was attached to the OTOARM. The OTODRIVE system was then connected to the OTOARM and OTOARM Aligner, and the system was draped. The Connector OD was then attached to the OTODRIVE to form a sterile barrier, and the Forceps OD were magnetically coupled to the OTODRIVE. To set the insertion trajectory, the OTODRIVE (and magnetically-coupled Forceps OD) was advanced to its most forward/distal position (40/40 mm). Using the OTOARM and OTOARM Aligner, the surgeon set the precise trajectory of the Forceps OD through the facial recess and toward the round window. To optimize the view of the surgical field, the microscope working distance was set between 350 and 400 mm to allow for sufficient clearance. The OTODRIVE was typically placed superiorly in the field, and the Forceps OD were pointed inferiorly toward the round window. This allowed a clear view of the electrode throughout the insertion. After a desired trajectory was set, the OTODRIVE was reversed to its most proximal position (0/40 mm). The electrode array was then secured to the Forceps OD ~10 to 15 mm proximal to the marker ring.

The electrode array was inserted through the round window membrane at a constant speed of 0.1 mm/s via foot pedal control of the OTODRIVE. ¹⁷ Manual surgical instruments were used as necessary to assist with electrode positioning or visualization during insertion. If resistance was encountered and insertion could not be completed robotically, the electrode array was released from the OTODRIVE and gentle manual advancement was attempted. Following insertion, the electrode array was coiled into the mastoid, and the facial recess was packed with fascia or muscle.

Postoperative Imaging and Electrode Assessment

Postoperative CBCT imaging was obtained in all available cases to confirm accurate placement of the electrode array according to standard institutional practice. The OTOPLAN software was used to analyze the postoperative scans, and imaging-based metrics were collected, including apical contact position (C1), number of electrode contacts inserted, and AID. Imaging analysis was performed using standardized methods by investigators blinded to audiologic outcomes.

Data Collection

Clinical data were extracted from medical and audiology records and recorded under unique study identifiers. Collected variables included patient demographics (age, sex), etiology and duration of HL, implanted ear, electrode type, and surgical details. Surgical performance metrics included total operative time, insertion time, and intraoperative or postoperative complications, including scalar deviation, electrode tip fold-over, incomplete insertion, or other adverse events.

Audiologic Outcomes

Audiologic outcomes were assessed using standard clinical protocols. Unaided and aided pure-tone averages (PTAs) were calculated using thresholds at 500, 1000, and 2000 Hz. The maximum output of the clinical audiometer was 120 dB HL at 500 and 1000 Hz and 115 dB HL at 2000 Hz. When no response was obtained at the maximum output level, thresholds were recorded at the audiometer limit for the purposes of PTA calculation, consistent with standard clinical reporting practice. Speech perception was assessed preoperatively and postoperatively using AzBio sentence recognition in quiet and CNC word recognition testing, when available. Audiologic follow-up duration was defined as the time from surgery to the most recent postoperative audiologic assessment.

Statistical Analysis

Descriptive statistics were used to summarize patient characteristics, surgical metrics, imaging findings, and audiologic outcomes. Continuous variables are reported as mean ± standard deviation, and categorical variables as counts and percentages. Preoperative and postoperative comparisons of continuous audiologic measures were performed using paired t-tests for PTAs and Wilcoxon signed-rank tests for speech perception outcomes (AzBio and CNC), given the non-normal distribution and floor effects observed in speech scores. Statistical significance was defined as p < .05. Data were analyzed using IBM SPSS Statistics, Version 29.0 (IBM Corp, Armonk, New York, NY, USA).

Results

Patient Demographics

Thirty-nine adult patients underwent robot-assisted cochlear implantation between February 2025 and August 2025. The mean age at implantation was 59.5 ± 19.2 years, and 23 patients (59%) were male. HL was predominantly acquired in etiology (n = 36, 92%), with idiopathic progressive and sudden sensorineural HL comprising the most common causes. Three patients (8%) had congenital HL. Implantation was performed on the left ear in 21 cases, the right ear in 17 cases, and bilaterally in 1 case. MED-EL FLEX28 electrode arrays were used in 38 cases, and FLEXSOFT arrays were used in 1 case, as shown in Table 1.

Table 1.

Patient Demographics and Baseline Clinical Characteristics.

Characteristic	Value
Number of patients	39
Age at implantation (y)	59.5 ± 19.2
Sex, n (%)
Male	23 (59%)
Female	16 (41%)
Laterality of implantation, n (%)
Left ear	21 (54%)
Right ear	17 (43%)
Bilateral	1 (3%)
Etiology of hearing loss, n (%)
Acquired	36 (92%)
Congenital	3 (8%)
Most common acquired etiologies	Idiopathic progressive SNHL; idiopathic sudden SNHL
Electrode array type, n (%)
MED-EL FLEX28	38 (97%)
MED-EL FLEXSOFT	1 (3%)

Values are presented as mean ± standard deviation or number (percentage), as appropriate.

Abbreviation: SNHL, sensorineural hearing loss.

Surgical Characteristics and Safety Outcomes

The mean total surgical time was 122.2 ± 49.4 minutes. Robotic system setup involved attachment of the OTOARM to the operating table, mounting of the OTOARM Aligner and OTODRIVE unit, sterile draping of the system, and alignment of the electrode insertion trajectory toward the round window. At our institution, setup of the OTODRIVE system typically requires ~10 minutes and is performed concurrently while trainees drill the receiver-stimulator bed. As a result, incorporation of the robotic system did not alter operative workflow. There were no intraoperative complications related to robotic electrode insertion. Full insertion of all 12 electrode contacts was achieved in 87.2% of implanted ears using robotic-assisted insertion. In 1 case, the final portion of insertion was completed manually due to trajectory limitations encountered with the forceps. In the remaining cases with incomplete insertion (10-11 contacts), the electrode array was released from the OTODRIVE and gentle manual advancement was attempted; however, further advancement was not achievable. These cases demonstrated shallower AIDs and reduced cochlear coverage. No cases of robotic system malfunction, electrode tip fold-over, or aborted robotic insertion occurred. No patients required revision surgery during the study period.

Audiologic Outcomes

Audiologic outcomes were available for the majority of implanted ears, with a mean audiologic follow-up duration of 3.8 ± 1.8 months postimplantation. All patients had preoperative unaided air-conduction thresholds demonstrating severe-to-profound HL across all tested frequencies (250-8000 Hz). The unaided PTA worsened modestly following implantation, increasing from 92.4 ± 16.7 dB HL preoperatively to 104.6 ± 11.9 dB HL postoperatively, corresponding to a mean threshold shift of +12.2 dB HL (paired t-test, p < .001). In contrast, aided thresholds improved significantly following cochlear implantation. The best aided PTA improved from 48.6 ± 13.2 dB HL preoperatively to 31.6 ± 6.3 dB HL in the postoperative CI-aided condition (paired t-test, p < .001). Improvements were observed consistently across the speech-relevant frequency range.

Speech Perception

Paired speech perception data were available for a subset of patients. AzBio sentence recognition in quiet (paired n = 28) improved from a preoperative mean of 18.1% to 75.7% postoperatively, representing a mean improvement of +57.6 percentage points (Wilcoxon signed-rank test, p < .001). CNC monosyllabic word recognition (paired n = 14) also demonstrated significant improvement, increasing from 37.7% preoperatively to 53.9% postoperatively, corresponding to a mean increase of +16.1 percentage points (Wilcoxon signed-rank test, p = .01). Collectively, postoperative aided thresholds and speech perception outcomes fell within expected performance ranges for contemporary adult CI recipients.

Postoperative CT Electrode Positioning and Insertion Metrics

Postoperative CBCT was available for all included robotic-assisted cochlear implantations and was used to evaluate electrode position, insertion depth, and cochlear coverage. Preoperative CBCT imaging suitable for quantitative analysis was available in 32 of 39 patients, while postoperative CBCT datasets suitable for analysis were available in 34 of 39 patients. Differences in sample size reflect exclusions due to incomplete electrode insertion, motion artifact, or suboptimal image quality precluding reliable measurement. Across the cohort, robotic-assisted insertion achieved consistent and reproducible AIDs. The mean postoperative C1 AID was 582.7° ± 35.2°. Predicted AID generated during preoperative planning averaged 569.0° ± 38.9°, with a mean absolute difference between predicted and achieved AID of 31.3°, indicating close concordance between planned and executed insertion depth (Figure 2).

Figure 2.

Planned versus achieved C1 AID. Boxplot comparison between predicted and achieved C1 AID. Boxes represent the interquartile range with median indicated; whiskers denote the range. AID, angular insertion depth.

Cochlear coverage was high overall, with a mean C1 cochlear coverage of 78.3% ± 3.8%, reflecting deep insertion while avoiding excessive apical advancement. Importantly, no radiographic evidence of electrode tip fold-over or gross scalar deviation was identified on postoperative CT imaging. Insertion metrics were consistent across surgeons, with no apparent surgeon-dependent variability in AID or cochlear coverage. All robotic insertions were performed at a standardized insertion speed of 0.1 mm/s, contributing to the observed reproducibility of electrode placement.

Discussion

This study represents one of the earliest real-world clinical evaluations of robot-assisted CI electrode insertion using the OTODRIVE^® system, integrating surgical safety metrics, imaging-based electrode placement, and early audiologic outcomes within a single adult cohort. The findings demonstrate that robotic electrode insertion can be implemented safely in routine clinical practice, with a low complication profile, reproducible insertion performance, and early speech perception outcomes consistent with expected CI benefit.

Surgical Safety and Feasibility

From a surgical standpoint, robot-assisted insertion was highly feasible across a heterogeneous adult population with predominantly acquired severe-to-profound sensorineural HL. No intraoperative complications were attributable to robotic insertion. Importantly, no cases of electrode tip fold-over, device malfunction, or aborted robotic insertion were observed. These findings support the procedural safety of OTODRIVE integration into standard cochlear implantation workflows. In the present cohort, the mean surgical time was 122.2 ± 49.4 minutes, which is comparable to operative durations reported for conventional cochlear implantation. In a large retrospective analysis of 455 unilateral CI procedures, mean procedural time was 147.1 ± 56.2 minutes. Despite the additional steps required for robotic system setup, operative times in our cohort remained within the range reported for conventional implantation. ²¹ These workflow characteristics may suggest that integration of the robotic system can be achieved without a meaningful increase in operative time once incorporated into routine practice, aligning with prior clinical experiences using robotic-assisted CI systems. ²²

Electrode Placement Accuracy and Imaging Correlates

Postoperative imaging demonstrated consistent electrode insertion depths and cochlear coverage across cases, with close agreement between preoperative planning and achieved insertion metrics. These findings are clinically meaningful when interpreted alongside prior cadaveric synchrotron-based analyses, which demonstrated low rates of intracochlear trauma and absence of severe structural injury using OTODRIVE-assisted insertion.^17,23 While direct trauma grading is not possible with clinical CBCT, the absence of radiographically-detectable scalar deviation or tip fold-over in this cohort is reassuring and consistent with the atraumatic insertion profile observed under high-resolution synchrotron imaging. ⁸

Together, these findings suggest that the controlled insertion speed and stable trajectory afforded by robotic assistance translate from experimental models into reproducible clinical electrode positioning. This consistency may be particularly valuable as cochlear implantation expands toward greater emphasis on individualized electrode selection, tonotopic alignment, and anatomy-informed programming strategies.

Audiologic Outcomes and Functional Benefit

Despite modest deterioration in unaided pure-tone thresholds postoperatively—an expected finding in a cohort largely implanted for non-hearing-preservation indications—patients demonstrated substantial improvements in aided thresholds and speech perception. Postoperative CI-aided PTAs reached levels consistent with effective CI performance, and speech recognition improved markedly on both sentence-level (AzBio) and word-level (CNC) testing.^12,24

In this cohort, postoperative audiologic testing demonstrated lower aided thresholds and increased speech perception scores following cochlear implantation and suggest that robotic insertion does not compromise functional hearing outcomes. While hearing preservation was not the primary objective in this cohort, the observed speech perception gains support the premise that atraumatic, controlled electrode insertion can achieve robust auditory benefit even in patients without residual low-frequency hearing.

Clinical Implications of Robotic Assistance

Robotic-assisted electrode insertion addresses several intrinsic limitations of manual insertion, including variability in insertion speed, micro-tremor, and force application—factors that have been shown to influence intracochlear pressure transients and structural trauma. ²⁵ By standardizing insertion parameters, robotic systems may reduce inter-surgeon variability and offer a reproducible platform for atraumatic implantation, particularly as candidacy expands to patients with greater residual hearing and more complex cochlear anatomy. ¹⁶

Furthermore, robotic assistance may serve as an enabling technology for future standardized integration of real-time intraoperative feedback, such as electrophysiologic monitoring or force-based safety thresholds, potentially allowing adaptive insertion strategies tailored to individual cochlear mechanics.

Limitations

Several limitations warrant consideration. First, the retrospective design and absence of a contemporaneous manual control group preclude direct comparative conclusions regarding superiority over conventional insertion. Second, audiologic follow-up remains relatively short, and longer term outcomes—including stability of speech perception and delayed scalar migration—require further evaluation. Third, interpretation of postoperative unaided thresholds is limited by audiometric ceiling effects, as many patients in this cohort demonstrated profound HL approaching the maximum output limits of the clinical audiometer. Fourth, while postoperative imaging confirms accurate electrode placement, clinical imaging lacks the resolution to detect subtle intracochlear trauma observable only with synchrotron-based methods. ⁸ Finally, this represents a single-center experience, and outcomes may reflect institutional expertise and workflow optimization. Multicenter studies will be important to confirm generalizability across diverse surgical environments. Future work should also explore integration of robotic systems with intraoperative monitoring modalities, such as electrophysiologic or force-based feedback, to further mitigate the risk of insertion-related trauma.

Conclusion

In this early clinical experience, robot-assisted CI electrode insertion using the OTODRIVE^® system demonstrated procedural safety, reproducible electrode placement, and early postoperative audiologic outcomes. Robotic insertion was successfully integrated into routine clinical practice without increased complication rates or operative inefficiency. While longer term and comparative data are needed, these findings support the feasibility and clinical effectiveness of robotic-assisted cochlear implantation and provide a foundation for continued prospective evaluation as CI surgery moves toward greater precision, standardization, and individualized care.