Carregant...
Tipus de document
TesiVersió
Versió publicadaData de publicació
Llicència de publicació
Si us plau utilitzeu sempre aquest identificador per citar o enllaçar aquest document: https://hdl.handle.net/2445/193164
Subtypes in Parkinson's disease and Dementia with Lewy bodies: MRI and neuropsychological profiles
Títol de la revista
Autors
Director/Tutor
ISSN de la revista
Títol del volum
Recurs relacionat
Resum
[eng] INTRODUCTION: Parkinson’s disease (PD) and Dementia with Lewy bodies (DLB) appear as neurodegenerative disorders with a wide range of symptomatology that differs among patients, among which are different levels of cognitive impairment. For instance, mild cognitive impairment (MCI) in PD contributes to a specific clinical profile with a higher risk of developing dementia. Looking at PD and DLB together provides evidence of the existence of different subtypes within both diseases. In recent years, complex imaging techniques, such as magnetic resonance imaging (MRI), have been used to study pathologies of the brain. MRI measures can be used to better characterize the brain basis of PD and DLB symptomatology, such as MCI. The reconstruction of the whole-brain connectome is a complex approach that can help to describe the heterogeneous symptomatology in neurodegenerative disorders. In addition, MRI in combination with new data-driven methods, such as cluster analysis, has been used to group patients according to their similarities, which allows subtypes to be identified. Until now, most studies in PD have described subtypes based on clinical and neuropsychological data, and just a few have used MRI measures to identify subtypes with different brain patterns. As DLB research is still in a relatively early stage, no cluster analyses have been yet performed based on MRI data.
OBJECTIVES AND HYPOTHESES: Given this context, the current Doctoral Thesis focuses on the heterogeneity that characterizes PD and DLB. The main objectives were to identify subtypes based on structural MRI measures in PD and DLB, as well as to characterize the structural brain connectivity of PD associated with MCI. We hypothesized that there would be PD subtypes with different patterns of grey and white matter alterations that would be associated with particular clinical and cognitive profiles. We also hypothesized that there would be DLB subtypes characterized by different grey matter (GM) patterns, and that these patterns would explain specific symptomatology of the disease and would be differentially associated to concomitant brain changes seen in cerebrovascular and Alzheimer’s diseases. Finally, we expected that PD-MCI would present a characteristic
pattern of impaired structural connectivity. In order to examine and clarify these issues, the current Doctoral Thesis is presented as a compendium of three studies.
METHODS: In Study 1, to identify subtypes in PD, we performed a hierarchical cluster analysis based on multimodal imaging using the Ward’s linkage method. We performed the analysis in a sample of 62 PD patients. GM volumes of cortical and subcortical brain regions as well as fractional anisotropy (FA) white matter (WM) measures were combined. Once the subtypes were identified, voxel-based morphometry (VBM) and tract-based spatial statistics (TBSS) analyses were carried out in order to compare the pattern of GM and WM of the PD subtypes to the 33 healthy control group. Demographical, clinical and neuropsychological data were used to characterize the subtypes.
In Study 2, the sample consisted of 27 PD-MCI and 35 PD without MCI, as well as 51 healthy controls. In this study we applied threshold-free network-based statistics (TFNBS), a novel technique based on whole-brain probabilistic tractography data useful to study structural connectivity. We complemented the analysis with TBSS and graph theory analyses (global and local measures).
Study 3 included 165 DLB subjects from the Mayo Clinic and 3 centres from the European DLB consortium (E-DLB). We performed a cluster analysis based on GM volumes using a random forest method, and characterized the subtypes based on GM volumes, clinical, demographical data as well as tau, β-amyloid and cerebrovascular biomarkers. Additionally, we characterized cognitive trajectories of the subtypes in a 3-year follow- up.
RESULTS: In Study 1, we identified 3 PD subtypes which mainly differed in GM patterns, while WM involvement appeared to be more limited. PD1 (24%) was characterized by temporo-parieto-occipital GM atrophy and subcortical atrophy, as well as FA reductions mainly affecting fronto-occipital WM tracts. This subtype was the oldest and had the worse neuropsychological profile. The second subtype, PD2 (34%), was characterized
by GM atrophy limited to frontal and temporal cortical regions, and a third subtype, PD3 (42%), with non-detectable GM atrophy or WM impairment, and preserved cognitive profile.
In Study 2, we found that PD patients had fewer streamlines (NOS) compared with healthy controls. Structural connectivity impairments were present in PD with and without MCI. However, the pattern and degree of connectivity impairment were different. PD-MCI showed a higher number of abnormal connections, primarily involving those between deep GM structures and cortical regions and posterior cortico-cortical connections, mainly in the temporal and occipital regions. PD without MCI showed fewer altered connections, and unlike PD-MCI, they were mainly located in the bilateral prefrontal cortex. What is more, the logistic regression and ROC curve analysis showed that the decreased NOS in the impaired connections characteristic of PD-MCI, were able to discriminate between both PD groups with high accuracy. The TBSS analysis revealed that only PD-MCI had reduced FA values compared to controls. The graph theory analysis showed PD groups differed in local graph measures.
In Study 3, three DLB subtypes with the same disease evolution were identified based on their GM volumes. The cortical predominant subtype (30%) was characterized by widespread reduced cortical GM, older age, worse cognition at baseline and faster cognitive decline over 3 years. The second subtype, the fronto-occipital subtype (46%), had lower GM volumes in frontal and occipital regions. Finally, the subcortical predominant subtype (24%) was characterized by the greatest GM volumes, and relatively low GM volumes in the basal ganglia, as they were the only brain regions where the 3 subtypes had equivalent GM volumes. The subcortical predominant subtype was also characterized by the highest frequency of cognitive fluctuations.
CONCLUSIONS: Our overall findings support the existence of different PD and DLB subtypes that can be identified by means of cluster analyses based on MRI data, which are in turn associated with specific cognitive profiles, and that cognitive impairment in PD is also associated to a specific pattern of structural connectivity impairment. These results contribute to clarifying the basis of heterogeneity in DLB and PD and give further information about which characteristics could be considered biomarkers of worse prognosis, with the final aim of approaching a more personalized medicine.
Matèries (anglès)
Citació
Col·leccions
Citació
INGUANZO PONS, Anna. Subtypes in Parkinson's disease and Dementia with Lewy bodies: MRI and neuropsychological profiles. [consulta: 13 de desembre de 2025]. [Disponible a: https://hdl.handle.net/2445/193164]