Automatic segmentation of intra-cochlear anatomy in post-implantation CT of unilateral cochlear implant recipients

Abstract

A cochlear implant (CI) is a neural prosthetic device that restores hearing by directly stimulating the auditory nerve using an electrode array that is implanted in the cochlea. In CI surgery, the surgeon accesses the cochlea and makes an opening where he/she inserts the electrode array blind to internal structures of the cochlea. Because of this, the final position of the electrode array relative to intra-cochlear anatomy is generally unknown. We have recently developed an approach for determining electrode array position relative to intra-cochlear anatomy using a pre- and a post-implantation CT. The approach is to segment the intra-cochlear anatomy in the pre-implantation CT, localize the electrodes in the post-implantation CT, and register the two CTs to determine relative electrode array position information. Currently, we are using this approach to develop a CI programming technique that uses patient-specific spatial information to create patient-customized sound processing strategies. However, this technique cannot be used for many CI users because it requires a pre-implantation CT that is not always acquired prior to implantation. In this study, we propose a method for automatic segmentation of intra-cochlear anatomy in post-implantation CT of unilateral recipients, thus eliminating the need for pre-implantation CTs in this population. The method is to segment the intra-cochlear anatomy in the implanted ear using information extracted from the normal contralateral ear and to exploit the intra-subject symmetry in cochlear anatomy across ears. To validate our method, we performed experiments on 30 ears for which both a pre- and a post-implantation CT are available. The mean and the maximum segmentation errors are 0.224 and 0.734mm, respectively. These results indicate that our automatic segmentation method is accurate enough for developing patient-customized CI sound processing strategies for unilateral CI recipients using a post-implantation CT alone.

Keywords: Cochlear implant; Cochlear implant programming; Image registration; Statistical shape model; Surface-to-image registration.

Fig. 1

Shown in (a) and (b) are surfaces of ST (red), SV (blue), and SG (green). In (b), a surface model of a CI electrode array inserted into ST is shown. In (c), contours of ST (red), SG (green) and the electrodes (purple) in the coronal view of a pre-implantation CT and a corresponding post-implantation CT, and in (d) contours of the SV (blue) in the coronal view of a pre-implantation CT and a corresponding post-implantation CT. The bright structure in the post-implantation CTs is the artifact cause by the CI electrode array. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

Surfaces of the labyrinth (shown in transparent gold) and intra-cochlear anatomy (shown in transparent red (ST), transparent blue (SV), and transparent green (SG)) viewed in three orientations (a), (b), and (c). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Image registration process.

Fig. 4

Active shape model generation process.

Fig. 5

Active shape segmentation process.

Fig. 6

Iterative intra-cochlear anatomy segmentation process.

Fig. 7

Inter-ear registration process.

Fig. 8

Points shown in blue are the points we use for computing R, the main parameter in our weight function. The remaining points of the labyrinth surface are shown in yellow. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 9

Mean error in the SOIs versus selection of R as a function of image intensity.

Fig. 10

Subject one’s ST, SV, and SG surfaces viewed in two different orientations. The color at each point encodes the distance in mm to the corresponding point on the registered contralateral surfaces. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 11

Automatically generated surfaces colormapped with errors in mm for subject 16 (top row) and subject 2 (bottom row). Left, surface of the labyrinth generated by the ASM-based method; right surface of the labyrinth generated by the atlas-based method. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 12

Quantitative results for the proposed segmentation method. The green squares on the box plots are quantitative results for the subject with the smallest maximum error, and the red squares are quantitative results for the subject with the largest maximum error. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 13

Qualitative segmentation results for the case with the smallest maximum segmentation error (shown in green box on Fig. 12). The contours shown are the ST (left panel), SV (middle panel), SG (right panel). Structure contours for gold-standard ST (red), gold-standard SV (blue), gold-standard SG (green), and automatic contours for all structures (yellow) are shown in a slice of a post-implantation image (top row) and a corresponding pre-implantation image (middle row). On the bottom panels the structure surfaces colormapped with segmentation errors are shown. (b) Shows similar information for the subject with the largest maximum segmentation error (shown in red box on Fig. 12). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 14

Various quantitative segmentation results for all 30 testing post-implantation CTs. (a) Mean and maximum error box plots for the SOIs segmented using the initialization method (left), using the proposed segmentation method (middle). On the right are the mean and maximum error box plots for the best possible SOIs segmentation results. (b) Shows the same information for the labyrinth.