研究背景和目的：腔隙和血管周围间隙是人群中较常见的核磁共振影像表现脑脊液信号的颅内空洞样结构。既往病理研究和近年来的头颅核磁影像研究提示，多数血管周围间隙小于腔隙；但临床实践中发现很多直径在3mm以上的液化病灶并非腔隙，而是大的血管周围间隙。目前，对大血管周围间隙的空间分布特征及其发病机制的研究有限；本研究假设两者发病机制不同，在临床预后上可能存在差异。本研究对腔隙和大血管周围间隙的人群罹患率，心脑血管疾病危险因素及临床预后进行分析；同时对比腔隙和大血管周围间隙的空间分布状况，并分析两者与多种影像学改变（白质高信号、微出血、脑萎缩、脑萎缩及颅内动脉狭窄）之间的相关性。

研究方法：本研究基于顺义地区社区人群研究。研究纳入大孙各庄镇5个自然村村民1844人，完整记录受试者基线一般情况、主要心血管危险因素及影像学资料；通过面对面及电话随访的形式完整记录评估患者并记录脑卒中和冠心病新发事件情况。由专门经培训人员在头核磁共振成像上评估腔隙、脑白质高信号体积、微出血、血管周围间隙严重程度及大血管周围间隙、脑实质分数和颅内动脉狭窄状况；应用MRICron软件对腔隙和大血管周围间隙的病变进行标记勾绘，将每个被试者的病变区用T1序列配准到脑标准空间，叠加后分别构建腔隙和大血管周围间隙的病变概率热区图并进行对比。采用多因素logistic回归模型评估腔隙、大血管周围间隙与年龄、性别、传统血管病危险因素相关性、卒中和冠心病新发事件风险及其与脑影像学改变的相关性。

研究结果：本研究最终分析1232例受试者，平均年龄55.1岁（标准差 9.7岁），男性比例37.3%。主要结果：1. 该人群中14.0%（173例）可观察到腔隙，19.0%（234例）可观察到大血管周围间隙；病变概率热区图显示，大血管周围间隙在基底节区下1/3部呈优势分布；腔隙在皮层下白质、基底节区及脑干部均较多见。2. 在调整年龄和性别后，腔隙与年龄、男性性别、高血压病及糖尿病显著相关；而大血管周围间隙与年龄和含有等位基因ε4的ApoE基因型显著相关，与传统心血管疾病危险因素无相关性。3.平均随访时间2.3年，新发脑卒中23例；腔隙与更高的新发脑卒中风险相关（HR 8.69，95%CI 3.13, 24.14）；而与不含大血管周围间隙者相比，含大血管周围间隙的受试者并未表现出更高的新发脑卒中风险。4. 腔隙与白质高信号、脑萎缩、微出血、血管周围间隙及颅内动脉狭窄均相关；大血管周围间隙除与腔隙相关外，与其他影像学改变之间均未发现相关性。

结论：腔隙与大血管周围间隙在人群中常见，两者空间分布有差异，危险因素不同，新发脑卒中风险迥异，与其他影像改变相关性亦不同。腔隙与大血管周围间隙可能存在不同的发病机制和临床预后意义。

Background and Purpose  Lacunes and perivascular spaces, common seen in general population with a similar signal as CSF, are both small fluid-filled cavities on brain MRI. According to histopathological and neuroimaging studies, most of the perivascular spaces are less than lacunes; but in clinical practice, many of the cavity lesions larger than 3mm in diameter are not lacunes, but large perivascular spaces. Up to now, there is a dearth of literature on spatial distribution and clinical profiles of large perivascular spaces. We assume that the pathogenesis of lacunes and large perivascular spaces is different. Through assessing the number and distribution, we investigated the prevalence, risk factors, correlation with other imaging changes (white matter hyperintensity, cerebral microbleed, brain atrophy, and intracranial vascular stenosis) and clinical prognosis of lacunes and large perivascular spaces in a community-based population.

Methods  There are 1844 participants from 5 villages in Shunyi District, Beijing enrolled. Baseline characteristics and major cardiovascular risk factor were recorded. Finally, 1232 participants with good quality of brain MRI image were collected. Patients were assessed by the form of face-to-face and telephone follow-up for new events in stroke and coronary heart disease. The lucune, white matter hyperintensity volume, cerebral microbleed, perivascular space severity and large perivascular space, brain volume and intracranial artery stenosis were assessed by trained raters on brain MRI. The lacunes and large perivascular spaces were traced and masked with MRICron software, and then registered to the standard space to construct the lesion probability map for comparision. Finally, multivariate logistic regression model was applied to assess the risk factors of lacunes and large perivascular spaces and their correlation with brain imaging changes.

Results  This study finally analyzed 1232 subjects, with an average age of 55.1 years (SD 9.7 years) and a male ratio of 37.3%. The main results were as follows:

1). The prevalence of lacunes in non-stroke cummunity population was 14.0% (173/1232), while the prevalence of large perivascular spaces was 19.0% (234/1232); As was shown in lesion probability map, large perivascular spaces were predominant in the lower third of basal ganglia; while lacunes were more common seen in the subcortical white matter, basal ganglia and brain stem.

2). After adjusting for age and sex, the risk factors of age, male gender, hypertension and diabetes were independently associated with lacune; while age and ApoE genotype with allele ε4 were associated with L-PVS.

3). During 2.3 years’ follow-up, there were 23 stroke events were observed in subjects over 35 years old. Subejects with lacune had a higher risk of incident stroke compared with those without lacunes on MRI, with HR 8.69 (95% CI 3.13, 24.14), whereas subjects with large perivascular spaces had no prominent higher risk of incident stroke.

4). The appearance of lacunes was associated with white matter hyperintensity, brain atrophy, cerebral microbleed, and stenosis of intracranial vessels; while the appearance of large perivascular spaces was not related to any other radiological changes other than lacunes.

Conclusion  Lacunes and large perivascular spaces are both common in the general population. Besides the disagreement in spatial distribution, there are differences in risk factors and incident risk for stroke, as well as the correlation with other imaging changes. Lacunes and large perivascular spaces may bear different pathogenesis and clinical prognostic significance.