一种神经影像学测量方法，用于捕捉帕金森病和路易体痴呆中萎缩的异质模式。

A neuroimaging measure to capture heterogeneous patterns of atrophy in Parkinson's disease and dementia with Lewy bodies.

机构信息

Dementia Research Centre, University College London, 8-11 Queen Square, London WC1N 3AR, United Kingdom; UCL Centre for Medical Image Computing, Department of Computer Science, University College London, 90 High Holborn, London WC1V 6LJ, United Kingdom.

出版信息

Neuroimage Clin. 2024;42:103596. doi: 10.1016/j.nicl.2024.103596. Epub 2024 Mar 21.

DOI:10.1016/j.nicl.2024.103596

PMID:38554485

原文链接:https://pmc.ncbi.nlm.nih.gov/articles/PMC10995913/

Abstract

INTRODUCTION

Parkinson's disease (PD) and Dementia with Lewy bodies (DLB) show heterogeneous brain atrophy patterns which group-average analyses fail to capture. Neuroanatomical normative modelling overcomes this by comparing individuals to a large reference cohort. Patient-specific atrophy patterns are measured objectively and summarised to index overall neurodegeneration (the 'total outlier count'). We aimed to quantify patterns of neurodegenerative dissimilarity in participants with PD and DLB and evaluate the potential clinical relevance of total outlier count by testing its association with key clinical measures in PD and DLB.

MATERIALS AND METHODS

We included 108 participants with PD and 61 with DLB. PD participants were subclassified into high and low visual performers as this has previously been shown to stratify those at increased dementia risk. We generated z-scores from T1w-MRI scans for each participant relative to normative regional cortical thickness and subcortical volumes, modelled in a reference cohort (n = 58,836). Outliers (z < -1.96) were aggregated across 169 brain regions per participant. To measure dissimilarity, individuals' Hamming distance scores were calculated. We also examined total outlier counts between high versus low visual performance in PD; and PD versus DLB; and tested associations between these and cognition.

RESULTS

There was significantly greater inter-individual dissimilarity in brain-outlier patterns in PD poor compared to high visual performers (W = 522.5; p < 0.01) and in DLB compared to PD (W = 5649; p < 0.01). PD poor visual performers had significantly greater total outlier counts compared to high (β = -4.73 (SE = 1.30); t = -3.64; p < 0.01) whereas a conventional group-level GLM failed to identify differences. Higher total outlier counts were associated with poorer MoCA (β = -0.55 (SE = 0.27), t = -2.04, p = 0.05) and composite cognitive scores (β = -2.01 (SE = 0.79); t = -2.54; p = 0.02) in DLB, and visuoperception (β = -0.67 (SE = 0.19); t = -3.59; p < 0.01), in PD.

CONCLUSIONS

Neuroanatomical normative modelling shows promise as a clinically informative technique in PD and DLB, where patterns of atrophy are variable.

摘要

简介

帕金森病（PD）和路易体痴呆症（DLB）表现出异质的脑萎缩模式，而群组平均分析无法捕捉到这些模式。神经解剖学的规范建模通过将个体与大型参考队列进行比较来克服这一问题。个体的特定萎缩模式被客观地测量并总结为总体神经退行性变的指标（“总离群计数”）。我们旨在量化 PD 和 DLB 参与者的神经退行性差异模式，并通过测试其与 PD 和 DLB 中关键临床指标的关联来评估总离群计数的潜在临床相关性。

材料和方法

我们纳入了 108 名 PD 患者和 61 名 DLB 患者。PD 患者被分为高和低视觉表现者亚组，因为先前的研究表明这可以将那些痴呆风险增加的患者分层。我们为每个参与者生成了 T1w-MRI 扫描的 z 分数，相对于参考队列（n=58836）中的皮质厚度和皮质下体积的区域模型。每个参与者的 169 个脑区的离群值（z < -1.96）被聚合。为了测量差异，计算了个体的汉明距离分数。我们还检查了 PD 中的高视觉表现者与低视觉表现者之间的总离群计数；以及 PD 与 DLB 之间的总离群计数；并测试了这些与认知之间的关联。

结果

PD 中视觉表现差的个体的脑离群模式的个体间差异明显大于视觉表现好的个体（W=522.5；p<0.01），DLB 中的差异也明显大于 PD（W=5649；p<0.01）。PD 中视觉表现差的个体的总离群计数明显大于视觉表现好的个体（β=-4.73（SE=1.30）；t=-3.64；p<0.01），而常规的群组水平 GLM 未能识别出差异。DLB 中较高的总离群计数与 MoCA 评分较差（β=-0.55（SE=0.27），t=-2.04，p=0.05）和综合认知评分（β=-2.01（SE=0.79）；t=-2.54；p=0.02）以及 PD 中的视知觉（β=-0.67（SE=0.19）；t=-3.59；p<0.01）呈负相关。