第一部分 中国社区人群基线脑小血管病影像标志物与人口危险因素的相关性分析 背景和目的：脑小血管病(cerebral small vessel disease,CSVD)是指由各种病因导致的脑小血管，包括脑小动脉、小静脉和毛细血管病变所引起的一组临床、影像学及病理表现的综合征，其影像学标志物包括：脑白质高信号(WMH)、腔隙(lacunes)、脑微出血(CMBs)，血管周围间隙扩大(ePVS)。本部分将探究上海静安社区队列基线影像标志物的特征，及与人口危险因素的相关性。 方法：上海静安社区的350名无痴呆和卒中的60-86岁的老人，于2009.09-2011.04期间进行了基线的检查，收集的资料包括临床资料、影像学资料、认知测试资料。影像标志物的评估分为定量和半定量评估，通过训练U-net网络深度学习的方法对WMH进行分割，以获得定量的WMH体积数据，腔隙灶的评估采用半定量计数，应用MARS量表对脑叶、深部、幕下区域的CMBs进行计数，应用半定量评分评估ePVS。统计分析应用stata14进行广义线性回归和顺序逻辑回归，分析基线影像标志物与危险因素的相关性。 结果：350例WMH体积分布从0.39ml~90.54ml不等，均值为10.14ml(标准差12.66ml)；腔隙灶的例数为56例(16%)；CMBs的例数为44例(13.8%)；高度ePVS (≥3分)的例数为49例(14%)。WMH体积与年龄、心源性疾病呈正相关(P<0.001， P=0.016)；腔隙灶与年龄、高血压呈正相关，IRR值及95%置信区间分别为1.36(1.10，1.69)，2.27(1.40，3.68)；CMBs与年龄、男性、BMI、高血压、心源性疾病呈正相关，IRR值及95%置信区间分别1.63(1.43，1.87)，2.80(2.04，3.83)， 1.27(1.12，1.44)，17.58(9.24，33.42)，5.09(3.77，6.89)；ePVS与年龄、男性、高血压呈正相关，OR值及95%置信区间分别为1.36(1.09，1.70)， 2.03(1.24，3.34)，1.70(1.07，2.70)。 结论：WMH几乎可见于所有60岁以上社区老人，其轻重程度分布差异较大，其余影像标志物在社区人群中也有一定的发生率。年龄、高血压与影像标志物的相关性最为显著。 第二部分 影像标志物的进展与基线影像标志物的相关性分析 背景和目的：国外的多个纵向社区队列描述了CSVD影像标志物随时间的进展情况，以及基线影像标志物与其进展的相关性，然而国内尚无对影像标志物进展的纵向研究。因此，本部分将探究CSVD影像标志物在7年内的进展情况，及基线影像标志物与各个影像标志物的进展的相关性。 方法：基线和随访影像标志物的评估方法同第一部分，定义WMH体积增加为随访期体积改变超过0.25ml，WMH体积减少为体积改变<-0.25ml，WMH体积不变为体积改变在[-0.25ml，0.25ml]；定义新发腔隙灶为随访期腔隙灶数目增加；定义新发CMBs为CMBs数目增加；定义ePVS进展为ePVS评分增加。统计分析应用stata14进行多元线性回归和多元逻辑回归，分析基线影像标志物与各个影像标志物的进展的相关性。 结果：191人在7年内出现WMH体积减少的有20例(10.5%)，WMH体积不变有12例(6.3%)，WMH体积增加有159例(83.2%)；19例(9.9%)出现新发腔隙灶；46例(28.9%)出现新发CMBs，65例(34.0%)出现ePVS进展。基线WMH体积与WMH体积的进展、新发腔隙灶、新发CMBs、ePVS进展显著正相关(P<0.001，P=0.008， P=0.010，P=0.017)。基线腔隙灶与WMH体积的进展、新发腔隙灶显著正相关(P=0.030，P=0.004)。校正了年龄、性别和危险因素后，以上相关性仍显著。 结论：社区人群CSVD影像标志物在7年内出现了明显的进展，且基线WMH体积与所有影像标志物的进展有显著相关性，基线WMH负荷越重，CSVD影像标志物进展越严重。 第三部分 影像标志物(基线＆进展)与认知功能改变的相关性分析 背景和目的：认知功能的减退是脑小血管病主要的临床结局之一，国外多个纵向社区队列探讨了CSVD影像标志物对认知功能改变的影响，然而国内缺乏影像标志物对认知功能改变影响的纵向研究数据。因此，本部分将探究基线影像标志物及影像标志物的进展与随访期认知功能的改变的相关性。 方法：基线和随访影像标志物的评估方法同第一、第二部分，认知功能的评估应用多个国际认可的汉化版的神经心理学测试，涵盖了总体认知功能，及记忆、语言、空间结构功能、注意、执行功能五大认知域。总体认知功能采用简明智能状态检查(MMSE)评估，记忆采用听觉词语学习测试(AVLT)或华山物品记忆测试(HOMT)评估，语言采用通用物品分类测试(COST)评估，空间结构功能采用搭火柴测试评估，注意采用连线测试(TMT)或人民币测试评估，执行功能采用冲突指令评估。认知测试的原始评分转化为代表各认知域相对水平的z评分。 结果：基线WMH体积与MMSE的改变显著相关(P<0.001)，且WMH体积每增加SD， MMSE的改变负向增加0.48；基线ePVS与记忆改变显著相关(P=0.042)，且ePVS每增加1分，记忆的改变负向增加0.25；基线WMH体积与执行功能的改变显著相关(P<0.001)，且WMH体积每增加SD，执行功能的改变负向增加0.46。ePVS进展与MMSE的改变显著相关(P=0.003)，且ePVS进展每增加1分，MMSE的改变负向增加0.87；WMH体积进展与语言的改变显著相关(P=0.028)，且WMH体积进展每增加SD，语言的改变负向增加0.17；WMH体积进展与执行功能的改变显著相关(P=0.019)，且WMH体积进展每增加SD，执行的改变负向增加0.22。 结论：基线WMH体积越大，总体认知功能和执行功能的减退越严重；基线ePVS越严重，记忆功能的减退越严重。WMH体积进展越多，语言及执行功能的减退越严重；ePVS进展越多，总体认知功能减退越严重。 关键词：脑小血管病、脑白质高信号、腔隙、脑微出血，血管周围间隙扩大 中图分类号：R743.9

 PartⅠ:Relationships between CSVD markers and demographic and risk factors in Chinese community-dwelling population Backgrounds and Objectives- Cerebral small vessel disease (CSVD) describes a syndrome of clinical, neuroimaging and neuropathological findings which is characterized by damage to cerebral arterioles, venules, and capillary lesions. MRI features include white matter hyperintensities (WMH), lacunes, cerebral microbleeds (CMBs) and enlarged perivascular spaces (ePVS). In this part, we explored the prevalence of these MRI markers in the residents of the Jing'an community in Shanghai, and their relationships with baseline demographic and risk factors. Methods- Three hundred and fifty people aged 60 to 86 years without dementia and stroke in the Jing'an community in Shanghai underwent baseline examinations in September of 2009 to April of 2011. Baseline examinations comprised a structured interview, a MRI scan of the brain and neuropsychological tests. Quantitative and semi-quantitative methods were applied in the evaluation of MRI markers. WMH was segmented on MRI by training U-net, and WMH volume was obtained automatically from U-net. Lacunes were evaluated by counting numbers. CMBs were counted according to MARS scale in lobar, deep, and infratentorial regions. ePVS were evaluated by semi-quantitative scales. We used generalized linear regression and ordinal logistic regression in Stata14 to assess the relationships between MRI markers and demographic and risk factors. Results- WMH volumes of 350 participants ranged from 0.39 ml to 90.54 ml, with a mean volume of 10.14 ml (standard deviation: 12.66 ml); 56(16%) had lacunes; 44(13.8 %) had CMBs; 49 (14%) had high degree ePVS (≥3 points). WMH volumes were positively correlated with age and cardiogenic disease (p<0.001, p=0.016, respectively); lacunes were positively correlated with age and hypertension, with IRR and 95% confidence intervals: 1.36 (1.10,1.69), 2.27 (1.40,3.68), respectively; CMBs were positively correlated with age, male, BMI, hypertension and cardiogenic disease, with IRR and 95% confidence intervals:1.63 (1.43,1.87),2.80 (2.04,3.83),1.27(1.12, 1.44),17.58(9.24,33.42),5.09(3.77,6.89),respectively;ePVS were positively correlated with age, male, and hypertension, with OR and 95% confidence intervals: 1.36 (1.09, 1.70), 2.03 (1.24, 3.34), 1.70 (1.07, 2.70), respectively. Conclusions- WMH was prevalent in almost all the residents over 60 with a wide range of severity. Other MRI markers were also prevalent in the community-dwelling population. Age and hypertension were most relevant to MRI markers. Part Ⅱ: Progression of CSVD markers and their relationships with baseline markers Backgrounds and Objectives- Longitudinal community studies in other countries described the progression of CSVD markers over time and examined the relationships between baseline markers and their progression. However, there was no longitudinal study on the progression of those markers in China. Therefore, in this part, we explored the progression of CSVD markers over 7 years and their relationships with baseline markers. Methods- MRI markers at baseline and follow-up were evaluated by the methods discussed in PartⅠ. Additionally, we defined WMH progression as volume change of more than 0.25 ml from baseline, WMH regression as change <-0.25 ml, WMH stability as change from -0.25ml to 0.25ml. Incident lacunes was defined as increase in the number of lacunes from baseline, incident CMBs as increase in the number of CMBs, and ePVS progression as increase in ePVS score. Statistical analysis was performed using stata14 for multiple linear regression and multiple logistic regression to analyze the relationships between baseline markers and progression of each markers. Results- Among 191 participants, 20(10.5%) developed WMH regression, 12 (6.3%) developed WMH stability, and 159 (83.2%) developed WMH progression over 7 years. 19 (9.9%) had incident lacunes; 46(28.9%) had incident CMBs; 65 (34.0%) had ePVS progression. Baseline WMH volumes were positively related with change of WMH volumes, incident lacunes, incident CMBs, and ePVS progression (P<0.001, P=0.008, P=0.010, P=0.017, respectively). Baseline lacunes were positively related with change of WMH volumes and incident lacunes (P=0.030, P=0.004, respectively). The relations remained significant after adjusting for age, gender and risk factors. Conclusions- CSVD markers displayed significant progression over 7 years, and baseline WMH volumes were significantly related with progression of all markers. Participants with severe WMH burden at baseline developed more progression of CSVD over follow-up period. Part Ⅲ: Relationships between MRI markers (baseline&progression), and change of cognitive function Backgrounds and Objectives- Decline of cognitive function is one of the main clinical outcomes of cerebral small vessel disease. Longitudinal community studies in other countries explored the impacts of CSVD markers on change of cognitive function. However, there was no longitudinal study on CSVD markers and change of cognitive function in China. Therefore, in this part, we explored the relationships between CSVD markers (baseline&progression) and change of cognitive function. Methods- MRI markers at baseline and follow-up were evaluated by the methods discussed in PartⅠ and Ⅱ. Multiple neuropsychological tests of Chinese version recognized internationally were applied to assess cognitive function, covering global cognitive function and five cognitive domains including memory, language, spatial construction function, attention, and execution function. Global cognitive function was assessed by Mini-Mental Status Examination (MMSE); memory was assessed by Auditory Verbal Learning Test (AVLT) or Huashan Objects Memory Test (HOMT); language was assessed by Common Objects Sorting Test (COST); attention was assessed by Trail making test (TMT) or Renminbi Test; execution function was assessed by conflicting instructions tasks. Raw scores of each cognitive domain were converted to z score for further analysis. Results- Baseline WMH volumes were significantly associated with change of MMSE (P<0.001), which negatively increased by 0.48 with each SD increase of WMH volumes. Baseline ePVS were significantly associated with change of memory (P=0.042), which negatively increased by 0.25 with each point increase of ePVS. Baseline WMH volumes were significantly associated with change of executive function (P<0.001), which negatively increased by 0.46 with each SD increase of WMH volumes. ePVS progression was significantly associated with change of MMSE (P=0.003), which negatively increased by 0.87 with each point increase of ePVS progression. WMH volume progression was significantly associated with change of language (P=0.028), which negatively increased by 0.17 with each SD increase of WMH volume progression. WMH volume progression was significantly associated with change of executive function (P=0.019), which negatively increased by 0.22 with each SD increase of WMH volume progression. Conclusions- Participants with more baseline WMH volumes had greater decline in global cognitive function and executive function. Participants with more baseline ePVS had greater decline in memory function. Participants with more WMH volume progression had greater decline in language and executive function. Participants with more ePVS progression had greater decline in global cognitive function. Key words: cerebral small vessel disease, white matter hyperintensities, lacunes, cerebral microbleeds, enlarged perivascular spaces Chinese Library Classification: R743.9