Abstract
OBJECTIVE: To investigate the accuracy of rigid cochlear models in measuring intra-cochlear positions of cochlear implant (CI) electrodes. PATIENTS: Ninety three adults who had undergone CI and pre- and postoperative computed tomographic (CT) imaging. MAIN OUTCOME MEASURES: Seven rigid models of cochlear anatomy were constructed using micro-CTs of cochlear specimens. Using each of the seven models, the position of each electrode in each of the 98 ears in our dataset was measured as its depth along the length of the cochlea, its distance to the basilar membrane, and its distance to the modiolus. Cochlear duct length was also measured using each model. RESULTS: Standard deviation (SD) across rigid cochlear models in measures of electrode depth, distance to basilar membrane, distance to modiolus, and length of the cochlear duct at two turns were 0.68, 0.11, 0.15, and 1.54 mm. Comparing the estimated position of the electrodes with respect to the basilar membrane, i.e., deciding whether an electrode was located within the scala tympani (ST) or the scala vestibuli (SV), there was not a unanimous agreement between the models for 19% of all the electrodes. With respect to the modiolus, each electrode was classified into one of the three groups depending on its modiolar distance: close, medium, and far. Rigid models did not unanimously agree on modiolar distance for approximately 50% of the electrodes tested. CONCLUSIONS: Inter-model variance of rigid cochlear models exists, demonstrating that measurements made using rigid cochlear models are limited in terms of accuracy because of non-rigid inter-subject variations in cochlear anatomy.