Cochlear implants (CIs) are neural prosthetics used to treat patients with severe-to-profound hearing loss. Patient-specific modeling of CI stimulation of the auditory nerve fiber (ANFs) can help audiologists improve the CI programming. These models require localization of the ANFs relative to surrounding anatomy and the CI. Localization is challenging because the ANFs are so small they are not directly visible in clinical imaging. In this work, we hypothesize the position of the ANFs can be accurately inferred from the location of the internal auditory canal (IAC), which has high contrast in CT, since the ANFs pass through this canal between the cochlea and the brain. Inspired by VoxelMorph, in this paper we propose a deep atlas-based IAC segmentation network. We create a single atlas in which the IAC and ANFs are pre-localized. Our network is trained to produce deformation fields (DFs) mapping coordinates from the atlas to new target volumes and that accurately segment the IAC. We hypothesize that DFs that accurately segment the IAC in target images will also facilitate accurate atlas-based localization of the ANFs. As opposed to VoxelMorph, which aims to produce DFs that accurately register the entire volume, our novel contribution is an entirely self-supervised training scheme that aims to produce DFs that accurately segment the target structure. This self-supervision is facilitated using a regional level set (LS) inspired loss function. We call our method Deep Atlas Based Segmentation using Level Sets (DABS-LS). Results show that DABS-LS outperforms VoxelMorph for IAC segmentation. Tests with publicly available datasets for trachea and kidney segmentation also show significant improvement in segmentation accuracy, demonstrating the generalizability of the method.

翻译：人工耳蜗（CI）是一种用于治疗重度至极重度听力损失的神经假体。针对CI刺激听神经纤维（ANFs）的患者特异性建模，有助于听力师优化CI编程参数。此类模型需定位ANF相对于周围解剖结构和CI的空间位置。由于ANF尺寸极小，在临床影像中无法直接显影，导致定位极具挑战性。本研究假设：由于ANF在耳蜗与大脑之间穿过内耳道（IAC），而IAC在CT影像中具有高对比度，因此可通过IAC位置准确推断ANF位置。受VoxelMorph启发，本文提出一种基于深度图谱的IAC分割网络。我们构建了一个单图谱，其中预先标注了IAC和ANF的位置。该网络经训练后生成形变场（DF），将图谱坐标映射至新目标体积，并精确分割IAC。我们假设：能精确分割目标图像中IAC的形变场，同样适用于实现基于图谱的ANF精确定位。与VoxelMorph追求生成精确配准全脑体积的形变场不同，本文的创新在于提出一种完全自监督的训练方案，旨在生成精确分割目标结构的形变场。该自监督机制通过引入基于区域水平集（LS）的损失函数实现。我们将该方法命名为基于水平集的深度图谱分割（DABS-LS）。实验结果表明，DABS-LS在IAC分割任务上优于VoxelMorph。在公开数据集上的气管和肾脏分割测试亦显示出显著的分割精度提升，验证了该方法的泛化能力。