基于前循环脑梗死神经血管介入专病库的临床科研一体化研究模式应用探讨  

Application of Integrated Clinical Research Model Based on Specialized Disease Database

作　　者：林琳[1] 孙瑄[2] 王韬[1] 缪中荣[2] 甘伟 牛明芳 LIN Lin;SUN Xuan;WANG Tao;MIAO Zhong-Rong;GAN Wei;NIU Ming-Fang(Information Center,Beijing Tian Tan Hospital,Capital Medical University,Beijing 100070,China;Neurointervention Center,Beijing Tian Tan Hospital,Capital Medical University,Beijing 100070,China;Goodwill Hessian Health Technology Co,Ltd.,Beijing 100085,China)

机构地区：[1]首都医科大学附属北京天坛医院信息中心,北京100070 [2]首都医科大学附属北京天坛医院神经介入中心 [3]北京嘉和海森健康科技有限公司

出　　处：《中国卒中杂志》2020年第3期332-336,共5页Chinese Journal of Stroke

摘　　要：神经介入是治疗脑血管病变的重要手段,首都医科大学附属北京天坛医院建立了以神经介入专病库为核心的临床科研一体化研究模式。该模式以自然语言处理、机器学习、深度学习等人工智能技术为依托,通过对神经介入中心在2012年5月-2019年6月收治的379例前循环脑梗死急诊取栓患者数据进行自动采集、处理和建模,构建了前循环脑梗死急诊取栓专病库。一方面形成了极细颗粒度的多维数据关联关系展现视图,为临床研究提供更多可能方向;另一方面,利用全过程诊疗时间轴及智能预测引擎,实现了对取栓后患者颅内出血的发生风险预测,对临床治疗决策的调整、优化具有重要意义。由此可见,基于专病库的科研临床一体化研究模式,将有效带动科研成果的产出及转化,实现以真实诊疗数据为基础、以研究成果辅助临床决策的目标,具有广阔的发展前景。Neurointervention is an important method for the treatment of cerebrovascular disease.In order to improve the level of clinical and scientific research more effectively,Beijing Tian Tan Hospital Affiliated to Capital Medical University established an integrated research model of clinical and research with the specialized disease database of neurointervention as the core.Based on artificial intelligence technologies such as natural language processing,machine learning and deep learning,through automatic collection,processing and modeling of the data of 379 patients with anterior circulation cerebral infarction who admitted to the intervention center for thrombectomy from May 2012 to June 2019,neurointervention center built a specialized disease database of emergency thrombectomy for anterior circulation cerebral infarction.After establishing the integrated research model,on the one hand,it formed a display view of multidimensional data relationship with extremely fine granularity,which provides more possible directions for clinical research;on the other hand,the whole-process diagnosis and treatment timeline data and intelligent prediction engine can be used to predict the risk of intracranial hemorrhage after thrombectomy,which is of great significance for the adjustment and optimization of clinical treatment decisions.Therefore,the integrated research model based on specialized disease database will effectively drive the output and transformation of scientific research achievements,and realize the goal of taking reality diagnosis and treatment data as the basis and research results as the aid to clinical decision-making,which has a broad development prospect.

关 键 词：临床科研一体化 专病库 神经介入 病历内涵质控 风险预测 

分 类 号：R743.33[医药卫生—神经病学与精神病学]