Cochlear implants (CIs) are surgically implanted neural prosthetic devices that are used to treat severe-to-profound hearing loss. These devices are programmed post implantation and precise knowledge of the implant position with respect to the intra cochlear anatomy (ICA) can help the programming audiologists. Over the years, we have developed algorithms that permit determining the position of implanted electrodes relative to the ICA using pre- and post-implantation CT image pairs. However, these do not extend to CI recipients for whom pre-implantation CT (Pre-CT) images are not available. This is so because post-operative images are affected by strong artifacts introduced by the metallic implant. To overcome this issue, we have proposed two methods to segment the ICA in post-implantation CT (Post-CT) images, but they lead to segmentation errors that are substantially larger than errors obtained with Pre-CT images. Recently, we have proposed an approach that uses 2D-conditional generative adversarial nets (cGANs) to synthesize pre-operative images from post-operative images. This permits to use segmentation algorithms designed to operate on Pre-CT images even when these are not available. We have shown that it substantially and significantly improves the results obtained with methods designed to operate directly on post-CT images. In this article, we expand on our earlier work by moving from a 2D architecture to a 3D architecture. We perform a large validation and comparative study that shows that the 3D architecture improves significantly the quality of the synthetic images measured by the commonly used MSSIM (Mean Structural SIMilarity index). We also show that the segmentation results obtained with the 3D architecture are better than those obtained with the 2D architecture although differences have not reached statistical significance.
Keywords: 3D-conditional generative adversarial networks; Cochlear implants; Segmentation, Metal artifact reduction.
Copyright © 2019. Published by Elsevier B.V.