Ferreira, Daniel

AbstractBackgroundBiological heterogeneity in Alzheimer’s disease (AD) has been observed and described as subtypes in measures of neurodegeneration (e.g., hypometabolism and atrophy). However, the relationship between these AD subtypes have not been studied. Thus, we aimed to compare and identify pathways associated with hypometabolism‐based and atrophy‐based subtypes.MethodCross‐sectional FDG‐PET and MRI scans from 173 amyloid‐positive AD individuals and 120 amyloid‐negative cognitively normal control individuals from the ADNI were included. Subtypes were identified in each neuroimaging modality by random forest hierarchical clustering using FDG‐PET uptake (measuring hypometabolism) and MRI volumes (measuring atrophy) in cortical and subcortical regions. We performed cross‐modal comparisons of atrophy patterns in hypometabolism‐based subtypes and hypometabolism patterns in atrophy‐based subtypes. To identify neurodegenerative pathways corresponding to each subtype, we used event‐based modelling to predict sequential progression of abnormality in hypometabolism and atrophy across brain regions (hippocampus, AD signature, frontal, posterior areas).ResultClustering revealed four hypometabolism‐based (two relatively limbic‐predominant: 25.7%, 19.8% and two relatively cortical‐predominant: 25.7%, 28.6%) and three atrophy‐based (diffuse: 18.7%, relatively limbic‐predominant: 28.6%, minimal: 52.6%) subtypes of AD with distinct spatial patterns and demographic and clinical characteristics. Cross‐modally, we found comparable AD‐typical atrophy patterns (involving the hippocampus and temporal regions) across the four hypometabolism‐based subtypes and comparable AD‐typical hypometabolism patterns (involving the retrosplenial cortex) across the three atrophy‐based subtypes. Event‐based modelling showed that each hypometabolism‐based subtype was associated with a unique sequential progression of regional hypometabolism, but not regional atrophy (sequence always initiated with frontal atrophy across all subtypes). Similarly, each atrophy‐based subtype was associated with a unique sequential progression of regional atrophy, but not regional hypometabolism (sequence always initiated with hypometabolism in AD signature areas across all subtypes).ConclusionOur findings suggest that hypometabolism‐based and atrophy‐based AD subtypes are not interchangeable. We identified unique metabolic and atrophy pathways specific to each subtype, which supports the notion that neurodegeneration manifests and progresses differentially across subtypes of AD. Despite being measures of neurodegeneration, hypometabolism and atrophy may not share the same extent of heterogeneity, bearing a complex multimodal relationship. Disentangling such relationships and combining complementary biomarkers to investigate heterogeneity could aid personalized treatment of AD.

AbstractBackgroundSubjective cognitive decline (SCD) implies self‐reported cognitive complaints in absence of objective cognitive impairment. SCD has been postulated as an early marker of Alzheimer’s Disease (AD). However, longitudinal studies have linked SCD to future vascular dementia. Thus, we conducted a systematic review and meta‐analysis aiming to synthesise and analyse the available data on the association between SCD and the presence of underlying cerebrovascular pathology measured by MRI.MethodOur study was performed following the PRISMA statement. The search strategy included terms for “subjective cognitive decline” and “white matter signal abnormalities”. It was conducted in 3 databases (PubMed, Scopus and Web of Science) from origin by December 8th, 2021. Two authors screened the studies in a 2‐stepwise approach. Primary studies including cognitively unimpaired adults with subjective cognitive complaints meeting the SCD‐I criteria for SCD that evaluated cerebrovascular pathology based on white matter signal abnormalities (WMSA) on MR images were selected. Study quality was assessed using the FLC 3.0 tool. Meta‐analysis comparing case‐control mean difference of WMSA was performed.ResultOf 240 articles identified, 16 research articles were selected. Of those, 8 case‐control studies were selected for the meta‐analysis. The sample size of the SCD group across the studies ranged from 8 to 906 participants, with age ranged from 40 to 84 years. All eight studies comprised a total of 2.637 SCD individuals. Study quality was deemed moderate (n=6) to high (n=10). The meta‐analysis showed a significant overall effect‐size between the mean WMSA burden of SCD and the controls, that was not significantly heterogeneous (Q(gl=7)=14.20; p=0,04; I2=42,5%). SCD was significantly associated with higher CVP burden based on neuroimaging measures.ConclusionThis is the first study to synthesise and analyse the current evidence on the association between SCD and WMSA. Our findings show the usefulness of SCD as a marker of underlying cerebrovascular pathology. Thus, the potential role of SCD as a preclinical stage of vascular cognitive impairment needs further investigation. Changes in lifestyle (e.g. exercise, diet, smoking cessation), cognitive stimulation or pharmacotherapy may be key to avoid an evolution towards cognitive deterioration in SCD.

AbstractBackgroundSubjective cognitive decline (SCD) in cognitively normal individuals is a relevant predictor of future cognitive decline, dementia, and neurodegeneration. Previous studies showed that the cognitive decline cannot be captured by objective neuropsychological testing in SCD. However, whether sophisticated multivariate approaches such as graph theory can potentially capture subtle cognitive abnormalities in SCD individuals is unknown. This study aimed to assess the cognitive connectome in a relatively large cohort of SCD individuals, in comparison with healthy controls (HC), and assess whether such a connectome differs in SCD across age groups.MethodWe included a total of 390 individuals, of which 211 fulfilled the core SCD criteria proposed by the international SCD Initiative (SCD‐I), and 179 individuals were classified as healthy controls. Participants were stratified into early middle‐age (37–50 years, n = 125), late‐middle‐age (51–64 years, n = 119), and elderly (65–78 years, n = 146) groups. We built cognitive networks from 47 cognitive variables separately for each age and study group using graph theory and compared the groups using global measures of average strength, average global efficiency, average local efficiency, and transitivity.ResultSCD individuals had more segregated networks in all age groups, but more prominently in the late‐middle‐age group. We observed lower within‐module connectivity in SCD individuals, in verbal memory and executive—premotor functions compared with HC. The late middle‐age SCD group showed a pattern of connectivity similar to the elderly HC group. The late‐middle‐age SCD group also showed reduced transitivity compared with the HC. We did not observe significant differences in the rest of the graph measures and in the other groups.ConclusionUsing a sophisticated multivariate approach, we demonstrate that a segregated pattern of cognitive connectivity seems to underlie SCD, particularly in individuals in the age range 51 to 64 years. After the age of 65 years, this pattern tends to be less segregated, contrary to an efficient dedifferentiation process in aging. This, together with subtle signs of brain pathology and other risk factors identified in SCD may increase the likelihood of future cognitive impairment and dementia.

Abstract Dementia with Lewy bodies (DLB) is clinically defined by the presence of visual hallucinations, fluctuations, rapid eye movement (REM) sleep behavioral disorder, and parkinsonism. Neuropathologically, it is characterized by the presence of Lewy pathology. However, neuropathological studies have demonstrated the high prevalence of coexistent Alzheimer's disease, TAR DNA‐binding protein 43 (TDP‐43), and cerebrovascular pathologic cases. Due to their high prevalence and clinical impact on DLB individuals, clinical trials should account for these co‐pathologies in their design and selection and the interpretation of biomarkers values and outcomes. Here we discuss the frequency of the different co‐pathologies in DLB and their cross‐sectional and longitudinal clinical impact. We then evaluate the utility and possible applications of disease‐specific and disease‐nonspecific biomarkers and how co‐pathologies can impact these biomarkers. We propose a framework for integrating multi‐modal biomarker fingerprints and step‐wise selection and assessment of DLB individuals for clinical trials, monitoring target engagement, and interpreting outcomes in the setting of co‐pathologies.

Abstract Background. Dementia with Lewy bodies (DLB) is a neurodegenerative disorder with a wide heterogeneity of symptoms, which suggests the existence of different subtypes. Objectives. We used data-driven analysis of MRI data to investigate DLB subtypes. Methods. We included 165 probable DLB from the Mayo Clinic and 3 centers from the European DLB consortium and performed a hierarchical cluster analysis to identify subtypes based on gray matter (GM) volumes. To characterize the subtypes, we used demographic and clinical data, as well as β-amyloid and cerebrovascular biomarkers at baseline, and cognitive decline over three years. Results. We identified 3 subtypes: an older subtype with reduced cortical GM volumes, worse cognition, and faster cognitive decline (n=49, 30%); a subtype with low GM volumes in fronto-occipital regions (n=76, 46%); and a subtype of younger patients with the highest cortical GM volumes, proportionally lower GM volumes in basal ganglia and the highest frequency of cognitive fluctuations (n=40, 24%). Conclusions. This study shows the existence of MRI subtypes in DLB, which may have implications for clinical workout, research, and therapeutic decisions.

AbstractBackgroundThe field of biological subtypes of Alzheimer’s disease (AD) has recently gained attention as a strong driver of precision medicine and future clinical trials. Neuropathology and neuroimaging studies have consistently identified four subtypes: hippocampal sparing AD, typical AD, limbic predominant AD, and minimal atrophy or minimal tau NFT AD. An important and still unresolved question is whether a truly hippocampal sparing subtype of AD can really exist, since it would challenge the widely accepted model of NFT spread. To address this question, we conducted a systematic review of the literature and performed original analyses on tau PET data.MethodOur systematic review included a search on EMBASE, PubMed, and Web of Science databases up to October 2020. Original analyses included implementation of several methods for AD subtyping on tau PET, and utilization of five complementary cut points to identify hippocampal sparing AD cases both in published studies and in data from the Alzheimer's Disease Neuroimaging Initiative (ADNI).ResultIn the three reviewed neuropathologic studies, we did not find any individual case with tau NFT in the association cortex while completely sparing the hippocampus (i.e., truly hippocampal sparing cases). In the in vivo tau PET studies, more lenient cut points consistently identified AD patients with tau NFT in the association cortex while completely sparing the hippocampus.ConclusionOur current data suggests that the hippocampal sparing subtype of AD does exist ante‐mortem, at earlier stages of the disease, but more research is needed to understand its pathogenesis and progression during life. Based on the findings from both in vivo tau PET and neuropathologic studies, we suggest that there are two possible independent pathways of tau spread: (1) a canonical pathway with early involvement of entorhinal cortex with subsequent involvement of limbic regions before eventually affecting multimodal association cortices, and (2) a less common pathway that affects multimodal association cortices with limbic involvement observed at end stages of the disease. As emphasized in the Murray 2011 study, hippocampal sparing relative to greater cortical involvement defines the phenotype.

AbstractBackgroundDementia with Lewy bodies (DLB) is a neurodegenerative disorder that presents with a variety of clinical symptoms. Part of the clinical heterogeneity within DLB has been related to the atrophy/sparing of the hippocampus. We extended that previous research by investigating whether different atrophy patterns exist in DLB, beyond and above the atrophy/sparing of the hippocampus, and whether those patterns contribute to the clinical heterogeneity in DLB. We aimed to identify DLB atrophy subtypes using a hypothesis‐free data‐driven clustering approach based on regional gray matter (GM) volumes and characterize the resulting subtypes across key measures.MethodWe included data from three centers of the European DLB consortium. A high‐resolution 3D T1‐weigthed magnetization prepared rapid gradient echo (MPRAGE) sequence was acquired for 97 DLB patients. Eighty‐four cortical regions and 12 subcortical regions from Mayo Clinic Adult Lifespan Template (https://www.nitrc.org/projects/mcalt/) atlases were obtained. Clusters of DLB patients were then determined using the random forest method. Between‐group differences in key demographic, clinical, MRI volumes and cerebrospinal fluid biomarkers (CSF) (amyloid‐beta, phosphorylated‐tau) were performed for clusters characterization.ResultWe identified three subtypes of DLB patients, with distinct patterns of GM volumes and differences in several clinical and demographic measures. Cluster 1 (C1) included 32 DLB patients (33%), cluster 2 (C2) included 29 DLB patients (30%), and cluster 3 (C3) included 36 DLB patients (37%). The clusters differed in age, with C1 being significantly older (72±8.6) than C3 (66±9.5) (F=3.93, p=0.02). C1 was characterized by lower volumes in fronto‐temporal regions and had a higher frequency of visual hallucinations (68%, vs. C2 35%). C2 showed lower volumes in occipital regions. C3 had higher GM volumes overall compared with the other two clusters and presented with a higher frequency of cognitive fluctuations (91%, vs. C1 67%). There were no differences between groups in CSF biomarkers. These results were independent of cluster differences in age and global GM volume.ConclusionThere are distinct patterns of brain atrophy in patients with DLB. These patterns seem to differentially contribute to the frequencies of visual hallucinations and cognitive fluctuations, independently of differences in age and global GM volume across subtypes.