Cochlear helix and duct length identification - Evaluation of different curve fitting techniques

Within the field of cochlear implantation (CIs), the role of utilizing patient-specific cochlear anatomy for choosing the optimal implant electrode is becoming increasingly important. Unfortunately, performing detailed anatomical measurements of a cochlea using clinical imaging data is rather time consuming and hence difficult to implement into the clinical routine. In order to accelerate clinical cochlear anatomy evaluations, previously developed mathematical models can be adjusted to the patient-specific anatomy by measuring just a few overall cochlear dimensions. However, the accuracy of model-based cochlear anatomy estimations is unclear, and incorrect evaluations may lead to false conclusions regarding the suitability of specific implant electrodes.

Based on 10 cochleae, an error evaluation of various commonly used curve fitting approaches for cochlear shape and duct length approximation was conducted. Spline tracings of the cochlear contours were used as reference values for the various approximations.

Parameterized average cochlear helix models and two of five analytical approaches were found to be suitable for reconstructing the cochlear helical shape and estimating its length.

Spline curve reconstructions are the most accurate and reliable method for assessing patient-specific cochlear geometry, especially in the case of anatomical irregularities. The most accurate results within the group of model-based evaluations still resulted in mean overall cochlear length deviations of approximately 5%.

Spline curve reconstructions appear to be the best option for anatomical diagnostics in clinical practice. Retrospective studies can be performed to further evaluate model-based evaluations.