Atlas-Based Segmentation of Intracochlear Anatomy in Metal Artifact Affected CT Images of the Ear with Co-trained Deep Neural Networks

TL;DR

This paper introduces a novel atlas-based deep learning method for segmenting intracochlear anatomy in metal artifact-affected CT images of cochlear implant patients, achieving comparable accuracy to state-of-the-art techniques with significantly reduced processing time.

Contribution

The authors develop a co-trained deep neural network approach that directly segments intracochlear structures in artifact-laden CT images without needing artifact removal or synthetic image generation.

Findings

Comparable segmentation accuracy to state-of-the-art methods

Significantly faster processing time

Effective learning of geometric features despite artifacts

Abstract

We propose an atlas-based method to segment the intracochlear anatomy (ICA) in the post-implantation CT (Post-CT) images of cochlear implant (CI) recipients that preserves the point-to-point correspondence between the meshes in the atlas and the segmented volumes. To solve this problem, which is challenging because of the strong artifacts produced by the implant, we use a pair of co-trained deep networks that generate dense deformation fields (DDFs) in opposite directions. One network is tasked with registering an atlas image to the Post-CT images and the other network is tasked with registering the Post-CT images to the atlas image. The networks are trained using loss functions based on voxel-wise labels, image content, fiducial registration error, and cycle-consistency constraint. The segmentation of the ICA in the Post-CT images is subsequently obtained by transferring the predefined segmentation meshes of the ICA in the atlas image to the Post-CT images using the corresponding DDFs generated by the trained registration networks. Our model can learn the underlying geometric features of the ICA even though they are obscured by the metal artifacts. We show that our end-to-end network produces results that are comparable to the current state of the art (SOTA) that relies on a two-steps approach that first uses conditional generative adversarial networks to synthesize artifact-free images from the Post-CT images and then uses an active shape model-based method to segment the ICA in the synthetic images. Our method requires a fraction of the time needed by the SOTA, which is important for end-user acceptance.