Department of Neurology, Xiangya Hospital, Central South University, Changsha, China.
Health Management Center, The Third Xiangya Hospital, Central South University, Changsha, China.
J Alzheimers Dis. 2022;90(1):139-149. doi: 10.3233/JAD-220482.
Some previous studies showed abnormal pathological and vascular changes in the retina of patients with Alzheimer's disease (AD). However, whether retinal microvascular density is a diagnostic indicator for AD remains unclear.
This study evaluated the macular vessel density (m-VD) in the superficial capillary plexus and fovea avascular zone (FAZ) area in AD, explored their correlations with clinical parameters, and finally confirmed an optimal machine learning model for AD diagnosis.
77 patients with AD and 145 healthy controls (HCs) were enrolled. The m-VD and the FAZ area were measured using optical coherence tomography angiography (OCTA) in all participants. Additionally, AD underwent neuropsychological assessment, brain magnetic resonance imaging scan, cerebrospinal fluid (CSF) biomarker detection, and APOE ɛ4 genotyping. Finally, the performance of machine learning algorithms based on the OCTA measurements was evaluated by Python programming language.
The m-VD was noticeably decreased in AD compared with HCs. Moreover, m-VD in the fovea, superior inner, inferior inner, nasal inner subfields, and the whole inner ring declined significantly in mild AD, while it was more serious in moderate/severe AD. However, no significant difference in the FAZ was noted between AD and HCs. Furthermore, we found that m-VD exhibited a significant correlation with cognitive function, medial temporal atrophy and Fazekas scores, and APOE ɛ4 genotypes. No significant correlations were observed between m-VD and CSF biomarkers. Furthermore, results revealed the Adaptive boosting algorithm exhibited the best diagnostic performance for AD.
Macular vascular density could serve as a diagnostic biomarker for AD.
一些先前的研究表明，阿尔茨海默病（AD）患者的视网膜存在异常的病理和血管变化。然而，视网膜微血管密度是否是 AD 的诊断指标尚不清楚。
本研究评估 AD 患者的视网膜浅层毛细血管丛和黄斑无血管区（FAZ）的黄斑血管密度（m-VD），探讨其与临床参数的相关性，最终确定用于 AD 诊断的最佳机器学习模型。
纳入 77 例 AD 患者和 145 例健康对照者（HCs）。所有参与者均采用光学相干断层扫描血管造影术（OCTA）测量 m-VD 和 FAZ 面积。此外，AD 患者进行神经心理学评估、脑磁共振成像扫描、脑脊液（CSF）生物标志物检测和 APOE ɛ4 基因分型。最后，通过 Python 编程语言评估基于 OCTA 测量的机器学习算法的性能。
AD 患者的 m-VD 明显低于 HCs。此外，轻度 AD 患者的黄斑中心凹、上内、下内、鼻内和整个内环的 m-VD 明显下降，而中度/重度 AD 患者的 m-VD 下降更为严重。然而，AD 和 HCs 之间的 FAZ 没有显著差异。此外，我们发现 m-VD 与认知功能、内侧颞叶萎缩和 Fazekas 评分以及 APOE ɛ4 基因型显著相关。m-VD 与 CSF 生物标志物之间没有显著相关性。此外，结果表明自适应增强算法对 AD 具有最佳的诊断性能。
黄斑血管密度可作为 AD 的诊断生物标志物。
