Evaluation of manual and robot-assisted electrode placement in malformed cochleae using imaging-based patient specific 3D models

Abstract Purpose Cochlear implantation in congenital inner ear malformations can be technically challenging due to the atypical anatomy which can limit electrode selection, trajectory, and insertion depth. Although high-resolution CT/MRI and planning tools (e.g., OTOPLAN ® ) support preoperative assessment, severe malformations may require additional validation. We evaluated whether robot-assisted insertion improves insertion performance in malformed cochleae using patient-specific 3D models. Methods High-resolution CT and/or MRI from 8 paediatric candidates were used to reconstruct and 3D print 16 cochleae (malformed cochleae plus one anatomically normal control with eighth nerve aplasia, as applicable). Four electrode arrays (MED-EL Medium, MED-EL Compressed, MED-EL Flex24, and Cochlear Straight) were inserted by an expert surgeon (P1), a novice surgeon (P2), and a robotic system (RobOtol ® , R1). Outcomes were insertion completeness (complete/incomplete) and ease (easy/hard). Data were analysed using Firth’s penalized logistic regression and classification tree analysis. Preoperative prediction by P1 alone versus P1 + OTOPLAN ® was compared using standard classification metrics. Results Electrode type and cochlear anatomy were the primary determinants of insertion outcomes. Electrode arrays with shorter active lengths (MED-EL Compressed) demonstrated higher insertion completeness in anatomically restricted malformed cochleae, whereas Flex24 and Cochlear Straight showed reduced completeness in restrictive hypoplastic cochleae. Operator identity, including robotic assistance, had minimal impact; ease was generally high. OTOPLAN ® assisted planning improved prediction performance and reduced false-negative assessments. Conclusion In malformed cochleae, insertion completeness was influenced primarily by cochlear morphology and electrode characteristics within the outcome measures assessed. Patient-specific 3D models provide a useful platform for electrode–anatomy compatibility assessment and preoperative planning; however, clinical validation incorporating tonotopic coverage, audiological outcomes, and objective robotic performance metrics is warranted.