Most intracerebral haemorrhages are due to cerebral small vessel diseases hypertensive arteriopathy (arteriolosclerosis), which affects deep perforating vessels, and cerebral amyloid angiopathy (CAA), which affects superficial cortical and leptomeningeal vessels. Diagnosis of CAA is important because it has a high recurrence risk (7·4% per year in a pooled analysis of cohort studies) and might require specific prevention strategies (eg, modifying the use of antithrombotic drugs) or, in the future, disease-modifying treatments targeting vascular β-amyloid. MRI with blood-sensitive sequences can detect haemorrhagic small vessel disease biomarkers, including cerebral microbleeds and cortical superficial siderosis, making it the best in-vivo diagnostic modality for CAA. By contrast, although widely available as the first diagnostic test for acute intracerebral haemorrhage, the value of CT for diagnosing CAA is uncertain. In The Lancet Neurology, Mark A Rodrigues and colleagues did a diagnostic accuracy study of acute non-contrast CT brain scans in a prospective populationbased inception cohort study of adults with first-ever intracerebral haemorrhage who died and underwent research autopsy. Of the 62 participants with lobar intracerebral haemorrhage, 36 (58%) were associated with moderate or severe CAA compared with 26 (42%) that were associated with absent or mild CAA, and were independently associated with subarachnoid blood extension (32 [89%] of 36; p=0·014), finger-like projections associated with the haematoma (14 [39%] of 36; p=0·043), and APOE ε4 possession (18 [50%] of 36; p=0·0020). The diagnostic criteria for CAA-associated lobar intracerebral haemorrhage had excellent discrimination (c statistic 0·92, 95% CI 0·86–0·98), confirmed by internal validation. For the rule-out criteria, neither subarachnoid haemorrhage nor APOE ε4 possession had 100% sensitivity (95% CI 88–100), whereas for the rule-in criteria, subarachnoid haemorrhage and either APOE ε4 possession or finger-like projections had 96% specificity (95% CI 78–100). A simplified CT-only model found subarachnoid haemorrhage alone had 89% sensitivity (95% CI 73–96), whereas subarachnoid haemorrhage and finger-like projections had 100% specificity (95% CI 84–100). The study was of high methodological quality, including population-based recruitment, blinded assessment of the index diagnostic test, and clear presentation of diagnostic accuracy and statistical uncertainty. Autopsy with standardised neuropathological evaluation is the best available reference standard to decide whether CAA is likely to have caused a particular intracerebral haemorrhage, although, as the authors note, other non-CAA small vessel disease (eg, arterioloscerosis or lipohyalinosis) frequently co-existed with CAA and so might have also contributed to some lobar intracerebral haemorrhage. The Edinburgh criteria could open a new chapter in CAA diagnosis because CT scanning is well tolerated by almost all acute patients with intracerebral haemorrhage and is more widely available than MRI, especially in countries with poor brain imaging resources. If CAA could be accurately diagnosed on the basis of acute CT findings, this diagnosis would rapidly provide clinicians, patients, and families with important information on prognosis in the acute stage of intracerebral haemorrhage that otherwise might be available only later or not at all. However, substantial limitations need to be addressed before the Edinburgh criteria might be useful for routine clinical practice. Most importantly, external validation must ensure that the excellent model performance does not relate to specific circumstances of patient population, recruitment, CT scan rating, or analysis. The included population was biased towards older patients with more severe intracerebral haemorrhage, early death, and other adverse prognostic features, so might not be applicable to patients with mild intracerebral haemorrhage in whom long-term management plans are needed. Furthermore, progressive cognitive impairment and transient focal neurological episodes are increasingly recognised as presenting syndromes of CAA, which were not included in this study. An important next step is to compare the Edinburgh CT-based criteria to the established and widely used MRI-based Boston diagnostic criteria, which, using strictly lobar bleeding and cortical superficial siderosis, have excellent sensitivity and good specificity for CAA, ideally using a brain tissue diagnosis as a common reference standard. Unfortunately, the authors could only obtain MRI in seven patients, so could not investigate this. Future studies should also determine whether CT-defined Lancet Neurol 2018

中文翻译：

---

CT扫描诊断CAA：回到未来？

大多数脑内出血是由影响深部穿支血管的脑小血管疾病高血压动脉病（动脉硬化）和影响浅表皮层和软脑膜血管的脑淀粉样血管病（CAA）引起的。CAA 的诊断很重要，因为它具有很高的复发风险（队列研究的汇总分析中为每年 7·4%），并且可能需要特定的预防策略（例如，改变抗血栓药物的使用），或者在未来的疾病- 针对血管β-淀粉样蛋白的改良疗法。具有血液敏感​​序列的 MRI 可以检测出血性小血管疾病生物标志物，包括脑微出血和皮质浅表铁沉积，使其成为 CAA 的最佳体内诊断方式。相比之下，尽管广泛用作急性脑出血的首选诊断测试，但 CT 诊断 CAA 的价值尚不确定。在《柳叶刀神经病学》中，Mark A Rodrigues 及其同事在一项基于人群的前瞻性初始队列研究中对急性非对比 CT 脑部扫描进行了诊断准确性研究，该队列研究对象为有史以来第一次脑内出血且死亡并接受研究尸检的成年人。在 62 名脑叶出血的参与者中，36 名 (58%) 与中度或重度 CAA 相关，而 26 名 (42%) 与无或轻度 CAA 相关，并且与蛛网膜下腔血肿独立相关 (32 [89%) ]，共 36 个；p=0·014)，与血肿相关的指状突起（36 个中的 14 个 [39%]；p=0·043）和 APOE ε4 占有（36 个中的 18 个 [50%]；p= 0·0020)。CAA 相关性脑叶出血的诊断标准具有极好的辨别力（c 统计量 0·92，95% CI 0·86–0·98），经内部验证证实。对于排除标准，蛛网膜下腔出血和 APOE ε4 占有率均不具有 100% 的敏感性（95% CI 88-100），而对于纳入标准，蛛网膜下腔出血和 APOE ε4 占有率或指状投影的敏感性为 96%特异性（95% CI 78–100）。一个简化的仅 CT 模型发现单独的蛛网膜下腔出血具有 89% 的敏感性（95% CI 73-96），而蛛网膜下腔出血和指状突起具有 100% 的特异性（95% CI 84-100）。该研究具有高方法学质量，包括基于人群的招募、指数诊断测试的盲法评估以及诊断准确性和统计不确定性的清晰呈现。具有标准化神经病理学评估的尸检是确定 CAA 是否可能导致特定脑内出血的最佳参考标准，尽管正如作者指出的那样，其他非 CAA 小血管疾病（例如，动脉硬化或脂蛋白变性）经常共存CAA 等也可能导致了一些脑叶出血。爱丁堡标准可以打开 CAA 诊断的新篇章，因为 CT 扫描几乎被所有急性脑出血患者耐受，并且比 MRI 更广泛可用，尤其是在脑成像资源较差的国家。如果可以根据急性 CT 结果准确诊断 CAA，该诊断将迅速为临床医生、患者、以及拥有有关脑出血急性期预后的重要信息的家庭，否则这些信息可能只能在稍后或根本无法获得。然而，在爱丁堡标准可能对常规临床实践有用之前，需要解决实质性的局限性。最重要的是，外部验证必须确保出色的模型性能与患者群体、招募、CT 扫描评级或分析的特定情况无关。纳入人群偏向于具有更严重脑出血、早期死亡和其他不良预后特征的老年患者，因此可能不适用于需要长期管理计划的轻度脑出血患者。此外，进行性认知障碍和短暂的局灶性神经系统发作越来越多地被认为是 CAA 的表现，但本研究未包括这些症状。下一步重要的工作是将基于爱丁堡 CT 的标准与已建立并广泛使用的基于 MRI 的波士顿诊断标准进行比较，后者严格使用脑叶出血和皮质浅表铁沉积，对 CAA 具有极好的敏感性和特异性，最好使用脑组织诊断作为通用参考标准。不幸的是，作者只能获得 7 名患者的 MRI，因此无法对此进行调查。未来的研究还应确定 CT 定义的 Lancet Neurol 2018 下一步重要的工作是将基于爱丁堡 CT 的标准与已建立并广泛使用的基于 MRI 的波士顿诊断标准进行比较，后者严格使用脑叶出血和皮质浅表铁沉积，对 CAA 具有极好的敏感性和特异性，最好使用脑组织诊断作为通用参考标准。不幸的是，作者只能获得 7 名患者的 MRI，因此无法对此进行调查。未来的研究还应确定 CT 定义的 Lancet Neurol 2018 下一步重要的工作是将基于爱丁堡 CT 的标准与已建立并广泛使用的基于 MRI 的波士顿诊断标准进行比较，后者严格使用脑叶出血和皮质浅表铁沉积，对 CAA 具有极好的敏感性和特异性，最好使用脑组织诊断作为通用参考标准。不幸的是，作者只能获得 7 名患者的 MRI，因此无法对此进行调查。未来的研究还应确定 CT 定义的 Lancet Neurol 2018 所以无法对此进行调查。未来的研究还应确定 CT 定义的 Lancet Neurol 2018 所以无法对此进行调查。未来的研究还应确定 CT 定义的 Lancet Neurol 2018