See related article, p 202

After excluding rare causes, the diagnostic question in a patient with primary intracerebral hemorrhage (ICH) is whether it was caused by arteriolosclerosis due to aging, hypertension and other vascular risk factors (sometimes simply called hypertensive ICH), or cerebral amyloid angiopathy (CAA). These 2 small vessel diseases have different risk factors and prognoses. The risk for recurrence is >6-fold higher for CAA-related ICH compared with hypertensive ICH.¹

For many years, the Boston criteria have served as a useful research and clinical tool for assigning the probability that a primary ICH was caused by CAA.² Taking advantage of the fact that CAA does not affect the vessels of the basal ganglia and brain stem to the same extent as non-CAA–related arteriolosclerosis does, a diagnosis of probable or possible CAA can be made based on the pattern of hemorrhages. When 2 or more hemorrhages, including microbleeds, are restricted to the cerebral cortical surface, cerebral cortex, or subcortical white matter, without affecting the basal ganglia or brain stem, then CAA is very likely to be present. The criteria were modified in 2010 to include cortical superficial siderosis as another sign of CAA.³

Until recently, the cerebellum was a neutral territory in CAA diagnosis. It has been assumed that most cerebellar hemorrhages are caused by hypertension and not CAA. However, a case series of cerebellar hematoma resections reported that CAA was the cause in a minority (14%).⁴ In light of the fact that CAA can cause cerebellar hemorrhages, the Boston criteria state that “cerebellar hemorrhage [is] allowed” when making a diagnosis of probable CAA. The intent was that cerebellar hemorrhages should not disqualify a patient from a diagnosis of probable CAA if there are lobar hemorrhages or cortical superficial siderosis without deep hemorrhages (personal communication, Dr Steven Greenberg). For example, a patient with an occipital lobar ICH, one or more lobar microbleeds (or a microbleed and cortical superficial siderosis), and additionally one or more cerebellar microbleeds would be classified as probable CAA. A patient with an occipital ICH and one or more cerebellar microbleeds, without lobar microbleeds or cortical superficial siderosis, would instead be classified as possible CAA. A patient with a primary cerebellar ICH and two or more microbleeds or one microbleed and cortical superficial siderosis would be classified as probable CAA.

Now, emerging evidence is showing that cerebellar hemorrhages and microbleeds can be classified as lobar (also termed superficial; ie, centered in the cerebellar cortex or underlying white matter) versus deep (ie, affecting the subcortical dentate nucleus), analogous to the classification of supratentorial microbleeds, and that the superficial and deep patterns are potentially related to the presence of CAA.^5–8 However, validation against neuropathologic evidence of β-amyloid, or validated in vivo surrogates, is still lacking. This gap in knowledge is addressed by an article in this issue of Stroke, in which the authors used positron emission tomography with Pittsburgh compound B—a validated biomarker of β-amyloid—to investigate its association with deep and lobar cerebellar microbleeds.⁹

The authors retrospectively analyzed 257 patients with primary ICH who were categorized by modified Boston criteria as CAA-ICH (36), hypertensive ICH (100), or mixed ICH (with both deep and lobar hemorrhages or microbleeds; 121). Cerebellar microbleeds were common, being present in 85 patients (33.1%) of whom 37 harbored superficial microbleeds while 48 had either deep or mixed cerebellar microbleeds. Patients with superficial cerebellar microbleeds were more likely to have CAA-related ICH by the modified Boston criteria (assigned independently of the cerebellar microbleeds) and to exhibit other CAA-related neuroimaging markers including cortical superficial siderosis and a higher number of visible centrum semiovale perivascular spaces. Conversely, deep cerebellar microbleeds were associated with a history of hypertension and supratentorial deep microbleeds. Pittsburgh compound B positron emission tomography amyloid imaging was done in a subgroup of 33 patients with cerebellar microbleeds including 16 patients with strictly superficial cerebellar microbleeds. Cerebellar amyloid signal was significantly higher in patients with CAA-related ICH by modified Boston criteria compared with non-CAA ICH and in patients with superficial cerebellar microbleeds compared with deep or mixed cerebellar microbleeds.

This study, although limited by relatively small sample size, suggests that superficial cerebellar microbleeds are associated with a highly specific pathologically validated biomarker of β-amyloid. This makes it likely that the presence or absence of superficial and deep cerebellar microbleeds will be included in future iterations of the Boston criteria. However, several important aspects of validation remain for future research. Direct pathological correlation, in hematoma resection specimens or full autopsies, is needed. The positron emission tomography ligand Pittsburgh compound B binds nonspecifically to parenchymal β-amyloid, as well as vascular β-amyloid¹⁰; so data are needed on the extent to which the positive Pittsburgh compound B signal is related to β-amyloid in either compartment. Studies with pathological confirmation will also be needed to determine how much cerebellar microbleeds add to the accuracy of CAA classification. In the present study, 1 of 36 (2.7%) patients with CAA-related ICH would have been reclassified as mixed ICH based on the presence of deep or mixed cerebellar microbleeds. The extent to which the presence or number of superficial cerebellar microbleeds increases the certainty that CAA is present is not clear. Many patients with deep or mixed ICH also had superficial cerebellar microbleeds (25/221, 11.3%). So it appears that CAA is not the exclusive cause of cerebellar superficial hemorrhages, just as it is not the exclusive cause of supratentorial lobar hemorrhages either. Whether cerebellar microbleed location, superficial versus deep, can be classified as reliably in routine radiological practice as in research studies is uncertain and warrants investigation. More studies are needed on patients with cerebellar ICH, who were excluded from the present study. Finally, prospective longitudinal studies are needed to determine whether ICH recurrence rates differ for superficial compared with deep cerebellar hemorrhages, as they do for supratentorial superficial and deep ICHs.

Cerebellar microbleed location is only one of a plethora of new markers of CAA that have been discovered since the original Boston criteria were derived. Other new markers include cortical superficial siderosis, centrum semiovale lacunes and perivascular spaces, acute diffusion-weighted imaging lesions, microinfarcts, abnormal white matter diffusion, cerebral atrophy, positive β-amyloid imaging, and low cerebrospinal fluid β-amyloid.¹¹ An international multicenter consensus effort is underway to correlate these new markers with CAA pathology, to produce a revised and recalibrated Boston criteria, version 2.0.¹² This effort should provide the pathological validation of cerebellar hemorrhage location with CAA and the added discriminative value compared with other CAA markers.

Dr Smith reports consulting fees from Alnylam Pharmaceuticals and Portola Pharmaceuticals and royalties from UpToDate. Dr Wollenweber reports receiving speakers fees and travel compensation from Bayer, Boehringer-Ingelheim, Bristol-Meyers Squibb, and Portola.

The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.

Guest Editor for this article was Jean-Claude Baron, MD, ScD.

参见相关文章，第 202 页

排除罕见病因后，原发性脑出血 (ICH) 患者的诊断问题是其是由衰老、高血压和其他血管危险因素引起的小动脉硬化（有时简称为高血压性 ICH）还是脑淀粉样血管病 (CAA) 引起的. 这两种小血管疾病具有不同的危险因素和预后。与高血压 ICH 相比，CAA 相关 ICH 的复发风险高出 6 倍以上。¹

多年来，波士顿标准一直是一种有用的研究和临床工具，用于确定原发性 ICH 是由 CAA 引起的概率。²利用 CAA 对基底节和脑干血管的影响程度与非 CAA 相关小动脉硬化相同的事实，可以根据出血模式诊断可能或可能的 CAA。当 2 处或更多处出血（包括微出血）局限于大脑皮质表面、大脑皮质或皮质下白质，而不影响基底节或脑干时，很可能存在 CAA。该标准在 2010 年进行了修改，将皮质浅表铁质沉着症作为 CAA 的另一个迹象。³

直到最近，小脑还是 CAA 诊断的中立区域。据推测，大多数小脑出血是由高血压而不是 CAA 引起的。然而，一系列小脑血肿切除术的病例报告称，CAA 是少数（14%）的病因。⁴鉴于 CAA 可导致小脑出血，波士顿标准规定在诊断可能的 CAA 时“允许小脑出血”。其目的是，如果有大叶出血或没有深部出血的皮质浅表铁质沉着症，小脑出血不应使患者失去对可能 CAA 的诊断资格（个人交流，Steven Greenberg 博士）。例如，患有枕叶 ICH、一种或多种脑叶微出血（或微出血和皮质浅表铁质沉着症）以及另外一种或多种小脑微出血的患者将被归类为可能的 CAA。枕部 ICH 和一个或多个小脑微出血，没有脑叶微出血或皮质浅表铁质沉着症的患者，将被归类为可能的 CAA。

现在，新出现的证据表明，小脑出血和微出血可分为脑叶（也称为浅表；即以小脑皮质或下方白质为中心）与深部（即影响皮质下齿状核），类似于以下分类：幕上微出血，表浅和深层模式可能与 CAA 的存在有关。^5-8然而，仍然缺乏针对 β-淀粉样蛋白的神经病理学证据的验证，或在体内验证的替代物。本期Stroke中的一篇文章解决了这种知识差距，其中作者使用匹兹堡化合物 B（一种经过验证的 β-淀粉样蛋白生物标志物）的正电子发射断层扫描来研究其与深部和小脑叶微出血的关联。⁹

作者回顾性分析了 257 名原发性 ICH 患者，这些患者根据改良的波士顿标准分为 CAA-ICH (36)、高血压 ICH (100) 或混合 ICH（伴有深部和大叶出血或微出血；121）。小脑微出血很常见，85 名患者 (33.1%) 存在，其中 37 名患有浅表微出血，48 名患有深部或混合性小脑微出血。根据改良的波士顿标准（独立于小脑微出血分配），小脑浅表微出血患者更可能患有 CAA 相关 ICH，并表现出其他 CAA 相关神经影像学标志物，包括皮质浅表铁质沉着症和更多可见的半卵圆中心血管周围间隙. 反过来，小脑深部微出血与高血压病史和幕上深部微出血有关。Pittsburgh 复合 B 正电子发射断层扫描淀粉样蛋白成像在 33 名小脑微出血患者的亚组中进行，其中 16 名患有严格的浅表小脑微出血患者。与非 CAA ICH 相比，CAA 相关 ICH 患者的小脑淀粉样蛋白信号显着高于非 CAA ICH，浅表小脑微出血患者与深部或混合小脑微出血患者相比。

这项研究虽然受限于相对较小的样本量，但表明小脑表面微出血与经过病理学验证的高度特异性的 β-淀粉样蛋白生物标志物有关。这使得小脑浅表和深部微出血的存在与否很可能被纳入波士顿标准的未来迭代中。然而，验证的几个重要方面仍有待未来研究。在血肿切除标本或完整尸检中，需要直接的病理相关性。正电子发射断层扫描配体匹兹堡化合物 B 与实质 β-淀粉样蛋白以及血管 β-淀粉样蛋白^{10非特异性结合}; 因此需要有关匹兹堡化合物 B 阳性信号与任一隔室中的 β-淀粉样蛋白相关程度的数据。还需要进行病理学证实的研究，以确定小脑微出血对 CAA 分类准确性的影响。在本研究中，36 名 CAA 相关 ICH 患者中有 1 名（2.7%）根据是否存在深部或混合性小脑微出血被重新分类为混合性 ICH。小脑浅表微出血的存在或数量在多大程度上增加了 CAA 存在的确定性尚不清楚。许多深部或混合性 ICH 患者也有浅表小脑微出血 (25/221, 11.3%)。因此，CAA 似乎不是小脑浅表性出血的唯一原因，正如它也不是幕上叶出血的唯一原因一样。小脑微出血的位置，浅表还是深部，在常规放射实践中是否可以像在研究中那样可靠地分类是不确定的，值得调查。需要对被排除在本研究之外的小脑 ICH 患者进行更多研究。最后，需要前瞻性纵向研究来确定浅表性脑出血与深部小脑出血的 ICH 复发率是否不同，就像幕上浅表和深部 ICH 一样。

小脑微出血位置只是自最初的波士顿标准得出以来已发现的大量 CAA 新标志物之一。其他新的标志物包括皮质浅表铁质沉着症、中央半卵圆形腔隙和血管周围间隙、急性弥散加权成像病变、微梗死、异常白质弥散、脑萎缩、β-淀粉样蛋白成像阳性和脑脊液 β-淀粉样蛋白含量低。¹¹一项国际多中心共识努力正在进行，以将这些新标志物与 CAA 病理学相关联，以产生修订和重新校准的波士顿标准 2.0 版。¹²这项工作应提供 CAA 小脑出血位置的病理学验证，并与其他 CAA 标志物相比增加鉴别价值。

Smith 博士报告了 Alnylam Pharmaceuticals 和 Portola Pharmaceuticals 的咨询费以及 UpToDate 的版税。Wollenweber 博士报告说，他们从拜耳、勃林格殷格翰、百时美施贵宝和波托拉获得演讲者费用和差旅补偿。

本文所表达的观点不一定是编辑或美国心脏协会的观点。

本文的客座编辑是医学博士、理学博士 Jean-Claude Baron。