基于静息态fMRI和独立成分分析的神经胶质瘤术前多功能系统综合定位技术研究

Glioma is the most common type of intracranial tumors. Tumor resection is the conventional clinical treatment. The principles of surgery is to remove as much of the tumor as possible while minimizing damage essential brain functions. As a non-invasive technique, functional Magnetic Resonance Imaging (fMRI) is playing more and more important role in presurgical function mapping, with conventional task-based fMRI being widely used in clinic. However, some patients with intracranial tumors are difficult to cooperate, or even can't accomplish task experience during scanning, causing failure of presurgical mapping. Moreover, previous clinic applications were mostly restricted only to sensorimotor and language areas. Compared with task fMRI, resting-state fMRI (R-fMRI) measures brain spontaneous activity with better feasibility and higher signal-to-noise ratio (SNR). And more importantly, using R-fMRI one can simultaneously detect many different functional systems with only once scan. Based on this, we will apply R-fMRI to glioma presurgical mapping, and will not only pay attention to sensorimotor system, but also to other functional systems. We hereby propose a new method, named "Presurgical Multi-Functional System Mapping (MFS-Mapping)", based on the R-fMRI and our newly-proposed "consistent" independent component analysis (ICA) method. Using MFS-Mapping, one can extract, identify and localize multiple functional systems with a single R-fMRI dataset. Complex multiple task experiments and multiple fMRI sessions will thus be unnecessary. We will also perform accuracy, validity and consistency assessment and automate MFS-Mapping, etc. This new technique, while being realized, can be used as a complementary method to task imaging, leading to more complete and accurate presurgical evaluation.

神经胶质瘤是颅内肿瘤最常见的一种，最主要的治疗手段是手术切除。手术原则强调最大化肿瘤切除，同时最大化功能区保留。无创的功能磁共振成像（fMRI）技术在胶质瘤术前功能区定位中已发挥重要作用。传统的基于任务fMRI进行的功能定位虽已应用于临床，但仍存在以下问题：部分病人难以配合甚至无法完成任务，导致功能定位失准；仅能定位少数功能区如运动、语言区。相比任务fMRI，静息态fMRI能简便地观察大脑功能活动，且具更高信噪比；更重要的是拥有同时研究多个功能系统的可能。基于此，本项目拟将静息态fMRI应用于胶质瘤术前功能区定位，不同于现有仅仅关注运动功能区的研究，我们提出"多功能系统全面定位"的方法，拟采用我们最新提出的稳定的独立成分分析算法，同时提取和定位多个重要脑功能网络，并进行准确性、可靠性和稳定性评价。研究成果将和传统的任务fMRI等定位技术互为补充，实现胶质瘤更全面的术前评估。

脑胶质瘤是常见恶性度极高的肿瘤。临床常规治疗包括手术和放化疗。其中手术为最主要手段。手术的目的是：尽可能全切或者扩大化切除，避免肿瘤复发；同时尽可能少破坏重要神经功能。因此，脑胶质瘤术前计划成为研究的热点和最有希望临床转化的技术之一。项目申请人利用在神经影像计算领域特别是独立成分分析（ICA）方法上的优势和特长，利用该项目的研究对临床转化提供了技术支持。在本项目中，申请人和其团队主要对1）基于静息态功能磁共振（rs-fMRI）的个体化肿瘤常用重要功能区定位的可行性；2）基于rs-fMRI和ICA进行个体化多功能区定位的可行性；3）多模态脑功能成像综合术前评估的临床价值和临床/科研合作；4）临床医生适用的术前定位软件研发；以及5）术前定位涉及的其他神经科学问题，进行了探索和研究。截至结题时，共有3项计算机软件著作权获批，9篇SCI文章接收，1项专利申请均受理并正在公开阶段；累积培养研究生3名和科研助理1名；获得横向课题和开放课题共计70余万；带动合作医院开展相关临床研究共计10余家医院；和合作者申请并获批其他省部级科研项目3个，国家级重大项目1个（作为合作单位唯一成员参与国家十二五科技支撑计划“头部疾患的防治研究”子课题“多模态脑功能区定位技术研究”）。值得说明的是，申请人指导的研究生撰写的2篇SCI论文（未算在已发表论文中），以及申请人的2篇SCI文章均已初步得到杂志社的修改（大修和小修）意见，并可在将来数月内发表。在这些系列性的工作中，有几个工作具有重要意义：一个为国际上首次唤醒麻醉术中功能磁共振定位研究，受到同行好评：Should it be used in clinic, it will have benefit for both patients and doctors. Although this is just a preliminary study, it has promising future. I think it is suitable for publishing（一旦应用在临床，将对病人和医生都有益处，虽然只是研究的初步阶段，但是能够看出非常广泛的应用前景）；一篇以长篇编辑部评论的形式发表在脑科学研究顶级期刊Brain上，国内某医疗公司正在向将我们的算法集成在它们公司的术前/术中定位仪器中。临床转化和产业化已经开展。