X-ray virtual histology for the investigation of temporal lobe epilepsy in a mouse model (v1)

This study evaluates virtual histology with synchrotron radiation-based X-ray microtomography to track the histopathological progression of mesial temporal lobe epilepsy (mTLE) in a mouse model. The kainate (KA) model of mTLE is used, where stereotactic KA injection into the hippocampus induces a pathological phenotype resembling that in humans. Label-free X-ray virtual histology with 1.6 μm-wide voxels was used to image paraffin-embedded brain hemispheres from mice sacrificed 1, 7, 14, and 21 days after KA injection. The onset of hippocampal sclerosis, namely granule cell dispersion and pyramidal neuron loss, could be identified with virtual histology. The volumetric virtual histology data were benchmarked against corresponding sections in conventional histology. Finally, automatic segmentation of the molecular, granular, and polymorph layers of the dentate gyrus with a deep convolutional neural network enabled three-dimensional quantification of electron density and morphology changes during epileptogenesis.

本研究采用基于同步辐射的X射线显微计算机断层扫描（synchrotron radiation-based X-ray microtomography）的虚拟组织学方法，对小鼠模型中内侧颞叶癫痫（mesial temporal lobe epilepsy, mTLE）的组织病理学进展进行追踪评估。本研究使用内侧颞叶癫痫的红藻氨酸（kainate, KA）模型，通过向海马体进行立体定位红藻氨酸注射，可诱导出与人类患者相似的病理表型。研究采用体素宽度为1.6 μm的无标记X射线虚拟组织学技术，对红藻氨酸注射后1、7、14及21天处死的小鼠的石蜡包埋大脑半球进行成像。借助该技术可识别出海马硬化的发生进程，具体表现为颗粒细胞弥散与锥体神经元丢失现象。研究将三维虚拟组织学数据与常规组织学的对应切片进行了基准对标验证。最终，本研究通过深度卷积神经网络（deep convolutional neural network）对齿状回的分子层、颗粒层及多形层实现自动分割，完成了癫痫发生过程中电子密度与形态变化的三维定量分析。