Multimodal Registration of the Brain's Cortical Surface

大脑皮质表面的多模态配准

• 批准号: 6924475
• 负责人: Michael Ian Miga
• 金额: $ 30.99万
• 依托单位: VANDERBILT UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2005
• 资助国家: 美国
• 起止时间: 2005-04-01 至 2008-01-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/6924475
• 关键词: biomedical automation; brain imaging /visualization /scanning; brain mapping; brain morphology; brain neoplasms; clinical research; computed axial tomography; cues; digital imaging; human subject; image enhancement; image guided surgery /therapy; lasers; magnetic resonance imaging; method development; nervous system disorder therapy; neurosurgery; patient oriented research; preoperative state; visual fields

DESCRIPTION (provided by applicant): Within the backdrop of a cost-conscious healthcare system, the need for providing quantitative relationships between the abundance of preoperative imaging information and the physical patient is clear. This is particularly important within the neurosurgical context where there is a distinct lack of anatomical landmarks. Developing a cost-effective method to improve a neurosurgeon's recognition of cortical brain structures during surgery as well as dynamically tracking these regions intraoperatively would be an important contribution to image-guided procedures. Laser range scanning (LRS) is a candidate technology that has the advantage of providing a wealth of intraoperative data with minimal space requirements. The work proposed herein involves a clinically translatable use of LRS technology that could have direct impact on surgical therapy for patients afflicted with brain tumors. With current technology, neurosurgeons often develop a surgical treatment plan by studying the patient's MR tomogram as a segmented reconstructed gray-scale encoded volume rendering. By visualizing the segmented brain in its three-dimensional state with the MR gray-scale providing anatomical landmarks, the surgeon can identify sensitive regions which may be important in the delivery of therapy. For example, gyri that are associated with primary, motor, somesthetic, auditory and visual functions can be identified usually on these renderings or with the assistance of cortical stimulation and/or functional magnetic resonance (MR) imaging. Unfortunately, these regions can often be difficult to recognize intraoperatively due to the lack of landmark recognition within the surgeon's field of view (FOV), i.e. the intraoperative presentation of the cortical surface. One fundamental hypothesis encompassing this proposed research is that textured (texture is from a digital image of the FOV) laser-range scan (tLRS) data can be used to correlate preoperative image data to the surgeon's field of view using a multi-modal registration approach. Specifically, the aims within this proposal are focused at: (1) enhancing surgical guidance by registering textured point clouds to the MR tomogram, (2) providing new anatomical cues to the surgeon, and (3) measuring cortical brain shift during surgical procedures. Although these enhancements are critical for correlating the cortical surface to its MR counterpart, the need to characterize subsurface tissue motion is equally important. As a result, this proposal outlines a strategy to directly address the use of tLRS data as a means to predict subsurface shift within patients undergoing a neurosurgical intervention. In summary, if image-guided procedures are to advance surgical therapy, the next evolution of these systems will require methods to account for soft tissue deformation. The tLRS studies proposed herein represent a clinically translatable strategy for documenting brain shift which could be pivotal in developing affordable image-guided platforms capable of compensating for soft tissue deformation.

描述（由申请人提供）： 在成本意识的医疗保健系统的背景下，需要提供术前成像信息和物理患者之间的定量关系是明确的。这在明显缺乏解剖标志的神经外科环境中尤其重要。开发一种具有成本效益的方法来提高神经外科医生在手术过程中对大脑皮层结构的识别以及术中动态跟踪这些区域将是对图像引导手术的重要贡献。激光范围扫描（LRS）是一种候选技术，其优势在于以最小的空间要求提供大量术中数据。本文提出的工作涉及LRS技术的临床可转化使用，其可能对患有脑肿瘤的患者的手术治疗产生直接影响。利用当前技术，神经外科医生通常通过将患者的MR断层图像研究为分段重建的灰度编码体绘制来制定手术治疗计划。通过在其三维状态下可视化分割的大脑，其中MR灰度提供解剖标志，外科医生可以识别在治疗递送中可能重要的敏感区域。例如，与初级、运动、躯体感觉、听觉和视觉功能相关的脑回通常可以在这些渲染上或在皮质刺激和/或功能性磁共振（MR）成像的帮助下识别。不幸的是，由于外科医生的视野（FOV）内缺乏标志识别，即，皮质表面的术中呈现，这些区域通常难以在术中识别。一个基本的假设，包括这项研究提出的纹理（纹理是从FOV的数字图像）激光范围扫描（tLRS）数据可以用来关联术前图像数据的外科医生的视野，使用多模式注册的方法。具体而言，该提案的目标集中在：（1）通过将纹理点云配准到MR断层图像来增强手术引导，（2）为外科医生提供新的解剖学线索，以及（3）在手术过程中测量大脑皮层移位。虽然这些增强对于将皮质表面与其MR对应物相关联至关重要，但表征次表面组织运动的需求同样重要。因此，该提案概述了一种直接解决使用tLRS数据作为预测接受神经外科干预的患者体内亚表面移位的手段的策略。总之，如果图像引导程序要推进外科治疗，这些系统的下一个发展将需要考虑软组织变形的方法。本文提出的tLRS研究代表了一种用于记录脑移位的临床可翻译策略，这对于开发能够补偿软组织变形的负担得起的图像引导平台至关重要。