Nowinski Wieslaw L
Biomedical Imaging Lab, Agency for Science Technology and Research, Singapore.
Acta Neurochir (Wien). 2008 Jan;150(1):1-13; discussion 13-4. doi: 10.1007/s00701-007-1270-6. Epub 2007 Nov 21.
The role of the brain atlas is changing in many aspects with the advancements in stereotactic and functional neurosurgery. Therefore, there is a critical need to construct a new atlas. This paper addresses the definition and construction of an atlas, ideal (in our opinion) for stereotactic and functional neurosurgery. The essence of the new atlas is not only its population-based structural and functional content, but also its continuous "self-updatability" with the new clinical results obtained.
The ideal atlas defined here contains four major components: brain models, knowledge database, tools, and clinical results. Towards its creation, a multi-atlas is proposed. The construction of the initial version of the multi-atlas is detailed with the probabilistic functional atlas (PFA), interpolated Talairach-Tournoux atlas, and enhanced Schaltenbrand-Wahren atlas. These atlases are put in a spatial register by matching their AC-PC distances and heights of the thalamus; the Schaltenbrand coronal and sagittal microseries are scaled laterally to match the target structure centroids with the locations of the best targets of the PFA.
Construction of an initial version of the ideal stereotactic atlas is feasible at present from the available resources. To achieve that, our three atlases (PFA, Talairach and Schaltenbrand) are enhanced and combined together. A single lateral scaling factor per target structure is feasible to co-register the Schaltenbrand atlas with PFA in four situations (compensated against the third ventricle, non-compensated, bilateral, and non-bilateral). The STN has to be stretched by 18% more than the VIM on the Schaltenbrand coronal microseries, and the VIM has to be compressed by 13% less than the STN on the Schaltenbrand sagittal microseries.
The new multi-atlas can potentially be more useful than the currently employed atlases and will facilitate further development of the ideal atlas for stereotactic and functional neurosurgery.
随着立体定向和功能神经外科的发展，脑图谱在许多方面的作用正在发生变化。因此，迫切需要构建一个新的图谱。本文阐述了一个对于立体定向和功能神经外科而言理想的（我们认为）图谱的定义和构建。新图谱的精髓不仅在于其基于人群的结构和功能内容，还在于其能根据获得的新临床结果持续“自我更新”。
这里定义的理想图谱包含四个主要组成部分：脑模型、知识库、工具和临床结果。为创建该图谱，提出了一种多图谱方法。详细介绍了多图谱初始版本的构建，包括概率功能图谱（PFA）、插值Talairach - Tournoux图谱和增强型Schaltenbrand - Wahren图谱。通过匹配它们的前连合 - 后连合（AC - PC）距离和丘脑高度，将这些图谱置于空间配准中；对Schaltenbrand冠状和矢状微系列进行横向缩放，以使目标结构质心与PFA最佳靶点的位置相匹配。
从现有资源来看，目前构建理想立体定向图谱的初始版本是可行的。为此，我们对三个图谱（PFA、Talairach和Schaltenbrand）进行了增强并组合在一起。在四种情况下（针对第三脑室进行补偿、未补偿、双侧和非双侧），为每个目标结构确定一个单一的横向缩放因子，可将Schaltenbrand图谱与PFA进行共同配准。在Schaltenbrand冠状微系列上，丘脑底核（STN）必须比腹中间核（VIM）多拉伸18%，而在Schaltenbrand矢状微系列上，VIM必须比STN少压缩13%。
新的多图谱可能比目前使用的图谱更有用，并将促进立体定向和功能神经外科理想图谱的进一步发展。
