PDF(Pubmed)
Abstract：
White matter hyperintensities (WMH) are considered hallmark features of cerebral small vessel disease and have recently been linked to Alzheimer\'s disease (AD) pathology. Their distinct spatial distributions, namely periventricular versus deep WMH, may differ by underlying age-related and pathobiological processes contributing to cognitive decline. We aimed to identify the spatial patterns of WMH using the 4-scale Fazekas visual assessment and explore their differential association with age, vascular health, AD imaging markers, namely amyloid and tau burden, and cognition. Because our study consisted of scans from GE and Siemens scanners with different resolutions, we also investigated inter-scanner reproducibility and combinability of WMH measurements on imaging.
We identified 1144 participants from the Mayo Clinic Study of Aging consisting of a population-based sample from Olmsted County, Minnesota with available structural magnetic resonance imaging (MRI), amyloid, and tau positron emission tomography (PET). WMH distribution patterns were assessed on FLAIR-MRI, both 2D axial and 3D, using Fazekas ratings of periventricular and deep WMH severity. We compared the association of periventricular and deep WMH scales with vascular risk factors, amyloid-PET, and tau-PET standardized uptake value ratio, automated WMH volume, and cognition using Pearson partial correlation after adjusting for age. We also evaluated vendor compatibility and reproducibility of the Fazekas scales using intraclass correlations (ICC).
Periventricular and deep WMH measurements showed similar correlations with age, cardiometabolic conditions score (vascular risk), and cognition, (p < 0.001). Both periventricular WMH and deep WMH showed weak associations with amyloidosis (R = 0.07, p = < 0.001), and none with tau burden. We found substantial agreement between data from the two scanners for Fazekas measurements (ICC = 0.82 and 0.74). The automated WMH volume had high discriminating power for identifying participants with Fazekas ≥ 2 (area under curve = 0.97) and showed poor correlation with amyloid and tau PET markers similar to the visual grading.
Our study investigated risk factors underlying WMH spatial patterns and their impact on global cognition, with no discernible differences between periventricular and deep WMH. We observed minimal impact of amyloidosis on WMH severity. These findings, coupled with enhanced inter-scanner reproducibility of WMH data, suggest the combinability of inter-scanner data assessed by harmonized protocols in the context of vascular contributions to cognitive impairment and dementia biomarker research.
We identified 1144 participants from the Mayo Clinic Study of Aging consisting of a population-based sample from Olmsted County, Minnesota with available structural magnetic resonance imaging (MRI), amyloid, and tau positron emission tomography (PET). WMH distribution patterns were assessed on FLAIR-MRI, both 2D axial and 3D, using Fazekas ratings of periventricular and deep WMH severity. We compared the association of periventricular and deep WMH scales with vascular risk factors, amyloid-PET, and tau-PET standardized uptake value ratio, automated WMH volume, and cognition using Pearson partial correlation after adjusting for age. We also evaluated vendor compatibility and reproducibility of the Fazekas scales using intraclass correlations (ICC).
Periventricular and deep WMH measurements showed similar correlations with age, cardiometabolic conditions score (vascular risk), and cognition, (p < 0.001). Both periventricular WMH and deep WMH showed weak associations with amyloidosis (R = 0.07, p = < 0.001), and none with tau burden. We found substantial agreement between data from the two scanners for Fazekas measurements (ICC = 0.82 and 0.74). The automated WMH volume had high discriminating power for identifying participants with Fazekas ≥ 2 (area under curve = 0.97) and showed poor correlation with amyloid and tau PET markers similar to the visual grading.
Our study investigated risk factors underlying WMH spatial patterns and their impact on global cognition, with no discernible differences between periventricular and deep WMH. We observed minimal impact of amyloidosis on WMH severity. These findings, coupled with enhanced inter-scanner reproducibility of WMH data, suggest the combinability of inter-scanner data assessed by harmonized protocols in the context of vascular contributions to cognitive impairment and dementia biomarker research.
摘要：
背景：白质高强度（WMH）被认为是脑小血管疾病的标志特征，并且最近与阿尔茨海默病（AD）病理有关。它们独特的空间分布，即脑室周围与深WMH，可能因导致认知能力下降的潜在年龄相关和病理生物学过程而有所不同。我们旨在使用4尺度Fazekas视觉评估来识别WMH的空间模式，并探索它们与年龄的差异关联，血管健康,AD成像标记,即淀粉样蛋白和tau负担，和认知。因为我们的研究包括来自不同分辨率的GE和西门子扫描仪的扫描，我们还研究了WMH成像测量的扫描仪间可重复性和可组合性。
方法：我们从梅奥诊所老龄化研究中确定了1144名参与者，这些参与者包括来自奥姆斯特德县的基于人群的样本。明尼苏达州有可用的结构磁共振成像(MRI)，淀粉样蛋白,和tau正电子发射断层扫描（PET）。在FLAIR-MRI上评估WMH分布模式，二维轴向和三维，使用Fazekas对脑室周围和深WMH严重程度进行评级。我们比较了脑室周围和深WMH量表与血管危险因素的关系，淀粉样蛋白PET,和tau-PET标准化摄取值比率，自动WMH卷，和认知使用皮尔逊偏相关后调整年龄。我们还使用组内相关性（ICC）评估了Fazekas量表的供应商兼容性和可重复性。
结果：脑室周围和深度WMH测量显示出与年龄相似的相关性，心脏代谢状况评分（血管风险），和认知，（p<0.001）。脑室周围WMH和深部WMH与淀粉样变性弱相关(R=0.07,p=<0.001),也没有tau负担。我们发现Fazekas测量的两台扫描仪的数据之间存在实质性的一致性（ICC=0.82和0.74）。自动WMH体积对于识别Fazekas≥2（曲线下面积=0.97）的参与者具有很高的辨别能力，并且与视觉分级相似，与淀粉样蛋白和tauPET标记物的相关性较差。
结论：我们的研究调查了WMH空间格局潜在的风险因素及其对全球认知的影响，脑室周围和深部WMH之间没有明显差异。我们观察到淀粉样变性对WMH严重程度的影响最小。这些发现,结合增强的WMH数据的扫描仪间再现性，建议在血管对认知障碍的贡献和痴呆生物标志物研究的背景下，通过协调协议评估的扫描仪间数据的可组合性。
方法：我们从梅奥诊所老龄化研究中确定了1144名参与者，这些参与者包括来自奥姆斯特德县的基于人群的样本。明尼苏达州有可用的结构磁共振成像(MRI)，淀粉样蛋白,和tau正电子发射断层扫描（PET）。在FLAIR-MRI上评估WMH分布模式，二维轴向和三维，使用Fazekas对脑室周围和深WMH严重程度进行评级。我们比较了脑室周围和深WMH量表与血管危险因素的关系，淀粉样蛋白PET,和tau-PET标准化摄取值比率，自动WMH卷，和认知使用皮尔逊偏相关后调整年龄。我们还使用组内相关性（ICC）评估了Fazekas量表的供应商兼容性和可重复性。
结果：脑室周围和深度WMH测量显示出与年龄相似的相关性，心脏代谢状况评分（血管风险），和认知，（p<0.001）。脑室周围WMH和深部WMH与淀粉样变性弱相关(R=0.07,p=<0.001),也没有tau负担。我们发现Fazekas测量的两台扫描仪的数据之间存在实质性的一致性（ICC=0.82和0.74）。自动WMH体积对于识别Fazekas≥2（曲线下面积=0.97）的参与者具有很高的辨别能力，并且与视觉分级相似，与淀粉样蛋白和tauPET标记物的相关性较差。
结论：我们的研究调查了WMH空间格局潜在的风险因素及其对全球认知的影响，脑室周围和深部WMH之间没有明显差异。我们观察到淀粉样变性对WMH严重程度的影响最小。这些发现,结合增强的WMH数据的扫描仪间再现性，建议在血管对认知障碍的贡献和痴呆生物标志物研究的背景下，通过协调协议评估的扫描仪间数据的可组合性。
