Abdelnour et al. Alzheimer’s Research & Therapy (2022) 14:14 https://doi.org/10.1186/s13195-021-00946-w RESEARCH Open Access Parsing heterogeneity within dementia with Lewy bodies using clustering of biological, clinical, and demographic data Carla Abdelnour1,2*† , Daniel Ferreira3†, Marleen van de Beek4, Nira Cedres3,5, Ketil Oppedal6,7,8, Lena Cavallin9,10, Frédéric Blanc11,12,13, Olivier Bousiges13,14, Lars‑Olof Wahlund3, Andrea Pilotto15, Alessandro Padovani15, Mercè Boada1, Javier Pagonabarraga16, Jaime Kulisevsky16, Dag Aarsland6,17, Afina W. Lemstra4† and Eric Westman3,18† Abstract Background: Dementia with Lewy bodies (DLB) includes various core clinical features that result in different pheno‑ types. In addition, Alzheimer’s disease (AD) and cerebrovascular pathologies are common in DLB. All this increases the heterogeneity within DLB and hampers clinical diagnosis. We addressed this heterogeneity by investigating sub‑ groups of patients with similar biological, clinical, and demographic features. Methods: We studied 107 extensively phenotyped DLB patients from the European DLB consortium. Factorial analy‑ sis of mixed data (FAMD) was used to identify dimensions in the data, based on sex, age, years of education, disease duration, Mini-Mental State Examination (MMSE), cerebrospinal fluid (CSF) levels of AD biomarkers, core features of DLB, and regional brain atrophy. Subsequently, hierarchical clustering analysis was used to subgroup individuals based on the FAMD dimensions. Results: We identified 3 dimensions using FAMD that explained 38% of the variance. Subsequent hierarchical cluster‑ ing identified 4 clusters. Cluster 1 was characterized by amyloid-β and cerebrovascular pathologies, medial temporal atrophy, and cognitive fluctuations. Cluster 2 had posterior atrophy and showed the lowest frequency of visual hal‑ lucinations and cognitive fluctuations and the worst cognitive performance. Cluster 3 had the highest frequency of tau pathology, showed posterior atrophy, and had a low frequency of parkinsonism. Cluster 4 had virtually normal AD biomarkers, the least regional brain atrophy and cerebrovascular pathology, and the highest MMSE scores. Conclusions: This study demonstrates that there are subgroups of DLB patients with different biological, clinical, and demographic characteristics. These findings may have implications in the diagnosis and prognosis of DLB, as well as in the treatment response in clinical trials. Keywords: Dementia with Lewy bodies, Alzheimer’s disease, Factorial analysis, Hierarchical clustering, Biomarkers, Heterogeneity *Correspondence: carla.abdelnour@gmail.com Background † Carla Abdelnour and Daniel Ferreira are shared first authors. The current diagnosis of probable dementia with † Afina W. Lemstra and Eric Westman are shared last authors. 1 Research Center and Memory Clinic, Ace Alzheimer Center Barcelona, Lewy bodies (DLB) is based on the presence of cog- Institut Català de Neurociències Aplicades, Universitat Internacional de nitive impairment, sufficient to impact patients’ abil- Catalunya-Barcelona, Centro de Investigación en Red-Enfermedades ity to perform activities of daily living. In addition, at Neurodegenerativas (CIBERNED), Barcelona, Spain Full list of author information is available at the end of the article least two of the following core clinical features must © The Author(s). 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Abdelnour et al. Alzheimer’s Research & Therapy (2022) 14:14 Page 2 of 13 be present: parkinsonism, recurrent visual hallucina- Methods tions, cognitive fluctuations, and/or rapid eye move- Participants ment (REM) sleep behavior disorder (RBD) [1]. These Participants were selected from the European DLB con- core clinical features often manifest in different com- sortium (E-DLB) [12]. The E-DLB consortium archives binations at the time of diagnosis or during the course data from 40 centers across Europe, including patients of the disease, increasing the clinical heterogeneity with probable DLB, Parkinson’s disease with dementia, within probable DLB. Previous studies have addressed or AD. For the current study, we included patients with part of this heterogeneity by investigating subgroups of probable DLB from the E-DLB centers that had MRI patients with certain core clinical features or different and CSF biomarkers available. Six centers satisfied these rates of clinical progression [2, 3]. However, extending criteria, including the Alzheimer Center Amsterdam, these analyses to biological features of the disease is Amsterdam UMC (Amsterdam, the Netherlands, n = warranted to elucidate the pathophysiology underlying 38); Day Hospital of Geriatrics, Memory Resource and the heterogeneity within probable DLB. Research Centre (Strasbourg, France, n = 38); Karolin- Although very few studies have directly addressed ska Institutet (Stockholm, Sweden, n = 17); University of the biological heterogeneity in DLB, there is an Brescia (Brescia, Italy, n = 6); Ace Alzheimer Center Bar- increasing interest in how Alzheimer’s disease (AD)- celona (Barcelona, Spain, n = 5); and Stavanger Univer- related pathology contributes to clinical presenta- sity Hospital (Stavanger, Norway, n = 3). A total of 107 tion in DLB. Part of the heterogeneity in DLB could probable DLB patients were included. be related to concomitant AD pathology, which is The diagnostic procedure and clinical examinations are present in more than 50% of DLB patients in neu- described elsewhere [13]. Briefly, the diagnosis was made ropathological studies [4, 5] and in around 30% in according to the 2005 International Consensus Crite- in vivo biomarker studies [6, 7]. A recent multi-center ria for probable DLB [14], based on detailed history and study in DLB patients showed that amyloid-β pathol- clinical assessment including physical, neurological, and ogy influences cognitive performance, whereas tau psychiatric examinations performed by a licensed neurol- affects clinical presentation through an association ogist. The criteria from 2005 were used because many of with lower frequency of parkinsonism and probable the patients were assessed prior to the 2017 International RBD [8]. In other cohorts, DLB patients with positive Consensus Criteria [15]. Exclusion criteria were patients AD biomarkers more frequently showed visual hal- with acute delirium, terminal illness, stroke, psychotic or lucinations [9]. Biological heterogeneity can also be bipolar disorder, craniocerebral trauma, or a major neu- studied through structural magnetic resonance imag- rological illness other than dementia. All centers recorded ing (MRI). A recent study investigated four atrophy whether patients fulfilled the criteria for parkinsonism, subtypes in DLB, and concluded that the pattern with visual hallucinations, cognitive fluctuations, and a clinical prominent cortical atrophy and sparing of the hip- history of probable RBD. Data about clinical core features pocampus was the most common subtype in probable were classified into present or absent in order to standard- DLB [10]. However, how all these dimensions of het- ize the information across centers, for statistical analyses. erogeneity inter-relate with each other is completely The Mini-Mental State Examination (MMSE) was scored unknown. Perhaps subgroups with distinct CSF pro- as a measure of global cognition [16]. files, atrophy patterns and clinical phenotypes are present. Multimodal subtyping studies are urgently needed to address this question, but such studies are Magnetic resonance imaging lacking so far [11]. MRI scanners and protocols used at each center are The goal of the current study was to parse the described in Supplementary Table 1 (Additional file 1). heterogeneity within probable DLB by using a mul- Due to variability in MRI scanners and protocols, we timodal subtyping method applied on the combina- favored visual rating scales by an experienced neurora- tion of CSF biomarkers, structural MRI, and clinical diologist (L.C.), rather than the application of automated and demographic measures. We gathered data from methods for regional brain atrophy. The neuroradiolo- a large multi-center cohort of patients with prob- gist was blind to any clinical information including diag- able DLB (N = 107). Firstly, we identified subgroups nosis. Regional atrophy was assessed with three visual of patients with factorial analysis and multimodal rating scales based on T1-weighted images as detailed clustering. Secondly, we characterized the resulting elsewhere [17]. Briefly, atrophy in the medial temporal subgroups across key CSF, MRI, clinical, and demo- lobe was assessed with the MTA scale [18]; atrophy in graphic measures. the posterior cortex was assessed with the PA scale [19]; Abdelnour et al. Alzheimer’s Research & Therapy (2022) 14:14 Page 3 of 13 and atrophy in the frontal lobe was assessed with the current discussions about the role and meaning of tau GCA-F scale [20]. In the three visual rating scales, a score pathology in DLB [26–28], and in consistence with our of zero denotes no atrophy, whereas scores from one to previous study [8], we described this non-AD pathologic three/four indicate an increasing degree of atrophy. MTA change as amyloid-independent tau-pathology in DLB analysis was based on coronal reconstructions, GCA-F patients. on axial reconstructions, and PA on reconstructions from all three planes. Our neuroradiologist (L.C.) has Statistical analysis previously demonstrated excellent intra-rater reliability The main aim of this study was to parse DLB heterogene- in 120 random cases: weighted kappa values of 0.94 and ity and identify different subgroups of patients based on 0.89 for MTA in left and right hemispheres, respectively; CSF biomarkers, regional brain atrophy, and key demo- 0.88 for posterior atrophy (PA); and 0.83 for global cor- graphic and clinical measures. This was done in two steps tical atrophy scale–frontal subscale (GCA-F) [17]. The as explained below and depicted in Fig. 1. same neuroradiologist assessed white matter hyperin- In the first step, we aimed to identify the latent dimen- tensities (WMHs) on axial FLAIR images, as a marker sions/components in the data that determine DLB het- of cerebrovascular disease, using the Fazekas scale [21]. erogeneity. Since our data included both continuous and Briefly, the Fazekas scale grades WMHs as 0 (i.e., absence categorical variables, we used a multivariate method of WMHs), 1 (i.e., punctate WMHs), 2 (i.e., early conflu- for data analysis called factorial analysis of mixed data ent WMHs), and 3 (i.e., WMHs in large confluent areas). (FAMD) [29]. The main strength of FAMD is that it Fazekas scores were classified into low (Fazekas scores 0 accommodates both quantitative and qualitative data or 1) and high (Fazekas scores 2 or 3) WMH burden, as in simultaneously. FAMD works as a principal compo- previous studies [22, 23]. nent analysis for quantitative data and as a multiple cor- respondence analysis for qualitative data [29]. In our Cerebrospinal fluid biomarkers FAMD model, age, years of education, MMSE scores, and Amyloid-β and tau neurofibrillary tangles were assessed disease duration were included as continuous variables, through CSF levels of Aβ42 and phosphorylated tau and sex (male vs. female), CSF Aβ42, p-tau and total tau (p-tau) at threonine 181. We also included total tau CSF levels, MTA, PA, and GCA-F scales (normal vs. abnor- levels as a marker of unspecific neurodegeneration. All mal), and parkinsonism, visual hallucinations, cognitive CSF analyses were performed locally following stand- fluctuations, and probable RBD (absent vs. present) were ard routines. Methods for CSF sampling, analysis, and included as categorical variables. Fazekas scores (low cut-off values for each center are described elsewhere [6, vs. high WMH burden) were not included in the FAMD 24] and detailed in Supplementary Table 2 (Additional model and subsequent cluster analysis due to missing file 1). Briefly, INNOTEST enzyme-linked immunosorb- data, but they were used to characterize the resulting ent assays (ELISAs) from Fujirebio, Ghent, Belgium, were subgroups, post hoc. used for total tau and p-tau biomarkers in all samples In the second step, we aimed to classify patients into and for Aβ42 in 101 samples. ELISA kits from Biosource subgroups using a cluster analysis based on the dimen- Europe S.A were used to analyze Aβ42 in the remain- sions provided by the FAMD model. Cluster analysis was ing 6 samples. To further standardize the information not applied directly on the original data because variables on CSF biomarkers across centers, CSF Aβ42, p-tau, and come in different scales and have a mixed nature (quanti- total tau values were classified as normal (-) or abnormal tative and qualitative). Instead, the output of the FAMD (+) using well-established center-specific cut-off points, model is a suitable input for cluster analysis because it is as described in previous E-DLB studies [6, 24]. The fre- scaled (all dimensions have the same scale) and continu- quency of abnormal CSF biomarker values was compared ous, and the high dimensionality of the original data is across DLB subgroups. In addition, subgroup characteri- reduced to a few latent dimensions (three in our study, zation was also done on the basis of a CSF AD profile, please see in the “Results” section). Furthermore, the following the current AT(N) classification framework to original variables are represented with different weights define AD biologically [25]. Briefly, abnormal levels of the in the dimensions, according to their contribution to Aβ42 biomarker alone were considered as indicative of the dimensions and the portion of variance explained by an AD pathological change (A+T-). Abnormal levels of each dimension. We applied an agglomerative hierarchi- the p-tau biomarker were considered as indicative of AD cal clustering algorithm with Ward’s linkage method [30]. pathology when in combination with abnormal levels of This clustering method starts by assigning every DLB the Aβ42 biomarker (A+T+) and considered as a non- patient to one cluster and sequentially combines pairs of AD pathologic change when in combination with normal clusters at each step while minimizing the sum of square levels of the Aβ42 biomarker (A-T+). In the context of errors from the cluster mean. The algorithm continues Abdelnour et al. Alzheimer’s Research & Therapy (2022) 14:14 Page 4 of 13 Fig. 1 Methodological design. Overview of factorial analysis of mixed data (FAMD) and hierarchical clustering analysis merging DLB patients into clusters until all the patients the patients were female. The average MMSE score was form a single group. We identified the optimal number of 25 ± 4. Parkinsonism and cognitive fluctuations were the clusters by using the Calinski-Harabasz criterion [31] and most frequently reported clinical features (81% and 84%, by visual inspection of the dendrogram from the agglom- respectively). Regarding the AD CSF biomarker profile, erative hierarchical clustering. 11% of the patients had AD pathology (A+T+), 18% We characterized the resulting subgroups using one- had an AD pathological change (A+T-), and 24% had way ANOVA for continuous variables, with a t-test for amyloid-independent tau pathology (A-T+). Thus, 29% post hoc pair-wise analysis, using Hochberg’s correction of patients can be categorized within the AD continuum for multiple testing [32]. The chi-square test was used for according to the AT(N) framework. Atrophy was more categorical data. We also used supervised random forest frequent in the parietal lobe (57%) than in the medial classification models to identify the measures that con- temporal (33%) and frontal (39%) lobes. tributed the most in the characterization of the clusters (discrimination of each cluster from all other clusters). Factorial analysis of mixed data (FAMD) In these random forest models, the cluster was a dichot- The FAMD model identified three dimensions that omous outcome (cluster k vs. all other clusters), and all together explained 38% of the variance in the data. the variables included in the FAMD were the predictors. Table 2 shows variables’ contribution to these dimen- Please see supplementary methods for more detail about sions. Figures 2, 3, and 4 display the three dimensions these random forest analyses (Additional file 2). pair-wise, and Fig. 5B displays all the three dimensions in All statistical analyses were conducted with the R sta- a 3D space. The first dimension accounted for 15.7% of tistical software (R Foundation for Statistical Computing, the variance and was mostly driven by atrophy in frontal Vienna, http://​www-R-​proje​ct.​org) [29]. A p-value ≤0.05 and parietal lobes, CSF p-tau levels, and age. In particu- was deemed statistically significant. lar, older patients had increased atrophy in frontal and parietal lobes and more often had abnormal CSF p-tau Results levels. In addition, CSF total tau levels, MMSE, years of Characteristics of the cohort education, CSF Aβ42 levels, sex, disease duration, and The key characteristics of the cohort are shown in parkinsonism also contributed statistically significantly Table 1. The average age was 68 ± 9 years and 28% of to the first dimension. Abdelnour et al. Alzheimer’s Research & Therapy (2022) 14:14 Page 5 of 13 Table 1 Characteristics of the whole cohort and DLB clusters Whole cohort Cluster 1 Cluster 2 Cluster 3 Cluster 4 Between- cluster ANOVA (N = 107) (n = 39) (n = 25) (n = 24) (n = 19) (p-value) Age 68 (± 8.7) 70 (± 7.2)b,d 64 (± 7.7)a,c 71 (± 10)a,d 64 (± 6.9)a,c 0.001 Sex, n men (%) 77 (72.0%) 28 (71.8%) 21 (84.0%)d 21 (87.5%)d 7 (36.8%)b,c 0.001 Education, years mean (SD) 11 (± 3.8) 11 (± 2.8)b,d 8.2 (± 2.4)a,c,d 12 (± 3.5)a,d 15 (± 3.3)a,b,c <0.001 Disease duration, years mean (SD) 4.3 (± 3.8) 4.2 (± 4.9)d 3.7 (± 2.7)d 3.5 (± 2.3)d 6.3 (± 3.8)a,b,c 0.013 d d d a,b,c MMSE score, mean (SD) 25 (± 4.0) 24 (± 3.9) 22 (± 3.9) 25 (± 3.8) 28 (± 2.0) <0.001 Core clinical features Parkinsonism, n present (%) 87 (81 %) 33 (85 %)c 25 (100%)c 11 (46 %)a,b,d 18 (95 %)c <0.001 b a Visual hallucinations, n present (%) 68 (64 %) 29 (74 %) 8 (32 %) 17 (71 %) 14 (74 %) 0.003 Cognitive fluctuations, n present (%) 90 (84 %) 39 (100 %)b 12 (48 %)a,d 20 (83 %) 19 (100 %)b <0.001 Probable RBD, n present (%) 68 (64 %) 23 (59 %) 20 (80 %) 15 (62 %) 10 (53 %) 0.234 CSF biomarkers Aβ42, n abnormal (%) 31 (29 %) 16 (41 %) 6 (24 %) 8 (33 %) 1 (5 %) 0.037* Total tau, n abnormal (%) 23 (21 %) 4 (10 %)c 0 (0 %)c 19 (79 %)a,b,d 0 (0 %)c <0.001 p-tau, n abnormal (%) 38 (36 %) 11 (28 %)c 4 (16 %)c 23 (96 %)a,b,d 0 (0 %)c <0.001 AD CSF profile, n abnormal (%) <0.001   AD pathology 12 (11 %) 3 (8 %) 1 (4 %) 8 (33 %) 0 (0 %)    AD pathological change 19 (18 %) 13 (33 %) 5 (20%) 0 (0 %) 1 (5%)    Amyloid independent tau-pathology 26 (24 %) 8 (21 %) 3 (12 %) 15 (63 %) 0 (0 %)   Normal 50 (47 %) 15 (38 %) 16 (64 %) 1 (4 %) 18 (95 %) Visual rating scales MTA, n abnormal (%) 35 (33 %) 23 (59 %)b,c 5 (20 %)a 3 (12 %)a 4 (21 %) <0.001 d d GCA-F, n abnormal (%) 42 (39 %) 20 (51 %) 11 (44 %) 11 (46 %)d 0 (0 %)a,b,c 0.002 PA, n abnormal (%) 61 (57 %) 19 (49 %) 19 (76 %)d 19 (79 %)d 4 (21 %)b,c <0.001 Fazekas, n high WMH burden (%) 29/92 (32%)§ 15/32 (47%)d 6/24 (25%) 7/18 (39%) 1/17 (6%)a 0.018 No missing data was recorded for the rest of the variables. p < 0.05 compared to cluster 1. p<0.05 compared to cluster 2. p<0.05 compared to cluster 3. dp<0.05 a b c compared to cluster 4. §Available data for Fazekas is n = 92. *Does not survive Hochberg’s correction in post hoc pair-wise comparisons. Abbreviations: ANOVA analysis of variance, MMSE Mini-Mental State Examination, Aβ42 amyloid-beta 1-42, p-tau phosphorylated tau, AD Alzheimer’s disease, MTA medial temporal lobe atrophy, GCA-F global cortical atrophy-frontal subscale, PA posterior brain atrophy, na not applicable The second dimension accounted for 12.5% of the Hierarchical clustering analysis variance and was mostly driven by parkinsonism, CSF Subsequently, we clustered the patients using agglom- total tau levels, years of education, and MMSE. Patients erative hierarchical clustering analysis on the three with higher education showed higher MMSE scores dimensions from the FAMD model as the input data. despite more frequently having abnormal CSF total tau Calinski-Harabasz (CH) values showed that four clus- levels, and they had a lower frequency of parkinson- ters (CH = 44.5) were more appropriate than two, ism. In addition, CSF p-tau levels, cognitive fluctua- three, or five clusters (CH < 42.0). Figure 5A shows the tions, visual hallucinations, sex, and CSF Aβ42 levels dendrogram from the cluster analysis, and Fig. 5B dis- also contributed statistically significantly to the second plays the distribution of the DLB patients colored by dimension. clusters 1 to 4. The third dimension explained 9.7% of the variance and Cluster 1 included 37% of the patients (n = 39), cluster was mostly driven by atrophy in medial temporal lobes, 2 included 23% (n = 25), cluster 3 included 22% (n = 24), cognitive fluctuations, and visual hallucinations. Patients and cluster 4 included 18% (n = 19) of the DLB patients. with atrophy in the medial temporal lobes more often Table 1 shows key demographic and clinical data, as had cognitive fluctuations and visual hallucinations. In well as CSF and MRI measures for all clusters. Briefly, addition, age, CSF Aβ42, p-tau, and total tau levels, as patients in cluster 1 were among the oldest and had inter- well as years of education, probable RBD, and atrophy in mediate levels of education, disease duration, and MMSE frontal lobes, also contributed statistically significantly to scores. Furthermore, all the patients in cluster 1 had the third dimension. cognitive fluctuations. Regarding AD CSF biomarkers, Abdelnour et al. Alzheimer’s Research & Therapy (2022) 14:14 Page 6 of 13 Table 2 Contribution of each variable to the dimensions of the FAMD Legend: Values represent the percentage of contribution of each variable to the total variation captured by each dimension. Gray-shadowed cells reflect the variables with highest contribution to each dimension. The asterisk (*) reflects the variables that contributed statistically significantly to each dimension. Abbreviations: MMSE mini-mental State examination; CSF cerebrospinal fluid; Aβ amyloid-beta; p-tau phosphorylated tau; MTA medial temporal lobe atrophy; GCA-F global cortical atrophy- frontal subscale; PA posterior brain atrophy; FAMD factorial analysis of mixed data Fig. 2 Dimension 1 vs. dimension 2. Continuous variables are depicted as arrows projecting lines (arrows represent the direction and degree of contributions). Categorical variables are depicted as triangles, which reflect variables’ centroids in the different levels of categorical variables. Abbreviations: MMSE Mini-Mental State Examination, CF cognitive fluctuations, PK parkinsonism, VH visual hallucinations, A+ abnormal CSF Aβ42, A- normal CSF Aβ42, T+ abnormal CSF p-tau, T- normal CSF p-tau, N+ abnormal CSF total tau, N- normal CSF total tau, GCA-F global cortical atrophy-frontal brain atrophy subscale, PA posterior brain atrophy Abdelnour et al. Alzheimer’s Research & Therapy (2022) 14:14 Page 7 of 13 Fig. 3 Dimension 1 vs. dimension 3. Continuous variables are depicted as arrows projecting lines (arrows represent the direction and degree of contributions). Categorical variables are depicted as triangles, which reflect variables’ centroids in the different levels of categorical variables. Abbreviations: MMSE Mini-Mental State Examination, CF cognitive fluctuations, PK parkinsonism, VH visual hallucinations, RBD REM sleep behavior disorder, A+ abnormal CSF Aβ42, A- normal CSF Aβ42, T+ abnormal CSF p-tau, T- normal CSF p-tau, N+ abnormal CSF total tau, N- normal CSF total tau, MTA medial temporal lobe atrophy, GCA-F global cortical atrophy-frontal brain atrophy subscale, PA posterior brain atrophy Fig. 4 Dimension 2 vs. dimension 3. Continuous variables are depicted as arrows projecting lines (arrows represent the direction and degree of contributions). Categorical variables are depicted as triangles, which reflect variables’ centroids in the different levels of categorical variables. Abbreviations: MMSE Mini-Mental State Examination, CF cognitive fluctuations, PK parkinsonism, VH visual hallucinations, RBD REM sleep behavior disorder, A+ abnormal CSF Aβ42, A- normal CSF Aβ42, T+ abnormal CSF p-tau, T- normal CSF p-tau, N+ abnormal CSF total tau, N- normal CSF total tau, MTA medial temporal lobe atrophy, GCA-F global cortical atrophy-frontal brain atrophy subscale Abdelnour et al. Alzheimer’s Research & Therapy (2022) 14:14 Page 8 of 13 Fig. 5 Dendrogram and clusters from the cluster analysis. A Dendrogram from the cluster analysis, with DLB patients depicted on the x-axis (each lower branch is a patient) and similarity depicted on the y axis (the shorter the distance along the axis, the greater the similarity). B Three-dimensional space generated by dimensions 1, 2, and 3 from the FAMD model. Dots represent the DLB patients colored by cluster [1 to 4] and distributed across the three-dimensional space cluster 1 had the highest frequency of an AD pathological in combination with a positive Aβ42 biomarker (AD change (A+T-). As for regional brain atrophy, patients in pathology, A+T+) or independently of Aβ42 (amyloid- cluster 1 had the highest frequency of medial temporal independent tau-pathology, A-T+). Additionally, cluster and frontal atrophy, showed intermediate levels of pari- 3 patients had a significantly higher frequency of abnor- etal atrophy, and had more often high WMH burden. The mal levels of total tau in CSF. The supervised random supervised random forest model showed that cognitive forest model showed that parkinsonism, and CSF total fluctuations, neurodegeneration markers (CSF total tau tau and p-tau were the measures that best characterize and regional atrophy), and age were the measures that this cluster (Supplementary Figure 1, Additional file 3). best characterize this cluster (Supplementary Figure 1, Patients in cluster 4 were among the youngest, had Additional file 3). the lowest frequency of men, had the highest levels of Patients in cluster 2 had the lowest levels of educa- education and MMSE scores, and had the longest dis- tion, MMSE scores, and frequency of visual hallucina- ease duration. All patients in cluster 4 had cognitive tions and cognitive fluctuations, and were among the fluctuations. All patients but one had a normal CSF AD clusters with younger age and shortest disease dura- biomarker profile (A-T-). Furthermore, patients in clus- tion. Moreover, patients in cluster 2 had the highest ter 4 had the lowest frequency of parietal atrophy and prevalence of parkinsonism and, together with clus- WMH burden, and none of them had frontal atrophy. ter 3, showed the highest frequency of parietal atro- The supervised random forest model showed that edu- phy. The supervised random forest model showed that cation, frontal atrophy, and CSF p-tau were the meas- visual hallucinations and cognitive fluctuations, as ures that best characterize this cluster (Supplementary well as education and age, were the measures that best Figure 1, Additional file 3). characterize this cluster (Supplementary Figure 1, Clusters did not significantly differ in the frequency Additional file 3). of probable RBD or abnormal levels of Aβ42 (irrespec- Patients in cluster 3 were the oldest, had intermediate tively of p-tau levels). Yet, the difference in abnormal levels of education and MMSE scores, had the shortest levels of amyloid-β emerged when considered together disease duration, and were the patients with the lowest with the tau biomarker (AD pathology (A+T+) or AD frequency of parkinsonism. Furthermore, patients in pathological change (A+T-)), likely due to the contri- cluster 3 had the highest levels of tau pathology, either bution of tau-pathology. Abdelnour et al. Alzheimer’s Research & Therapy (2022) 14:14 Page 9 of 13 Discussion older DLB patients. This combination of biological find- In this study, we expanded the current knowledge about ings suggests a subtype with concomitant amyloid-β and the biological heterogeneity within probable DLB by cerebrovascular pathologies. The association between studying a relatively large biomarker cohort. We applied amyloid-β and older age [8], atrophy in the medial tem- a method for multimodal subtyping on CSF biomarkers, poral lobe [7, 42–44], and cerebral amyloid angiopathy structural MRI, and clinical and demographic measures, [45, 46] has been reported in previous studies. The nov- all of them combined. We identified four DLB subgroups elty of our study is the identification of a subgroup that that ranged from a cluster with almost no concomitant encapsulates all those features. Similarly, the limbic pre- AD or cerebrovascular pathologies (cluster 4) to three dominant subtype of AD also includes older patients with clusters with various degrees of concomitant AD and/or prominent medial temporal atrophy and a high burden of cerebrovascular pathologies (clusters 1, 2, and 3), and as cerebrovascular pathology [11]. well different regional brain atrophy, clinical and demo- On the other hand, clusters 2 and 3 were characterized graphic features. by a low frequency of medial temporal atrophy. What did Cluster 4 was characterized by the presence of virtu- characterize clusters 2 and 3 was the high frequency of ally normal AD CSF biomarkers and a very low burden posterior brain atrophy in both subgroups. The combi- of cerebrovascular disease. Therefore, we suggest that nation of posterior brain atrophy and sparing of medial the underlying pathology in this subgroup very likely temporal lobes describes the signature pattern of brain is mainly α-synuclein-related. This subgroup included atrophy in probable DLB [10]. Cluster 3 included older younger DLB patients with longer disease duration and DLB patients with the highest frequency of tau pathol- better MMSE performance than the other 3 subgroups. ogy, either in combination with amyloid-β pathology Similarly, a previous study comparing DLB patients (hence reflecting AD pathology, A+T+) or in isolation with and without concomitant AD pathology found (hence reflecting a non-AD pathological change, in this that “pure” DLB subjects were younger and had higher case, amyloid-independent tau-pathology, A-T+). In MMSE scores [9]. Furthermore, our cluster 4 showed a contrast, patients in cluster 2 were younger and most of slight predominance of women, while the whole cohort them showed normal CSF AD biomarker levels (A–T–). was mostly constituted by men. This sex distribution Therefore, results from clusters 2 and 3 are again in could be influenced by the sample site characteris- agreement with recent studies showing that concomi- tics, since most of the patients in cluster 4 come from tant AD increases with age in probable DLB [8]. Further- the Strasbourg center. Researchers from the Strasbourg more, these results suggest that tau-related pathology center have recently a the predominance of women in can contribute to α-synuclein pathology either in iso- DLB patients in France [33]. However, previous stud- lation or in combination with amyloid-β, in DLB. One ies on sex differences in DLB have found mixed results. example of this is the high frequency of medial tempo- Some studies have demonstrated a predominance of ral atrophy in cluster 1, and the high frequency of pos- women [33, 34], while other studies have shown an asso- terior brain atrophy in clusters 2 and 3. Previous studies ciation between male sex and DLB [35, 36]. In relation showed that medial temporal atrophy in DLB is associ- to the core clinical features, all patients in cluster 4 had ated with amyloid-β pathology [7, 42–44], while poste- cognitive fluctuations, which is one of the most typi- rior brain atrophy is associated with the combined effect cal characteristics of DLB [37]. Additionally, cluster 4 of amyloid-β and tau-related pathologies [42], matching showed the least regional brain atrophy, and all patients the pattern of tau accumulation in the posterior cortex had normal total tau CSF levels. This implies no bio- in positron emission tomography studies in DLB [26, 47, marker evidence of neurodegeneration in cluster 4, 48]. probably due to the absence of concomitant AD pathol- In addition, our data suggest a possible association ogy and cerebrovascular disease [38–41]. Cluster 4 between higher tau pathology and shorter disease dura- might thus reflect the purest DLB subtype in our cohort. tion. Cluster 3, the subgroup with the highest tau pathol- In contrast, the other three DLB subgroups showed ogy, was among the clusters with the shortest disease varied degrees of concomitant AD or cerebrovascular duration. Previous studies showed that tau pathology is pathologies. Our biological data based on CSF biomark- associated with a worse prognosis in DLB patients [49]. ers and structural MRI suggest two different profiles. On Cluster 2 was the other cluster among those with the the one hand, cluster 1 showed the highest frequency of shortest disease duration. Although tau levels were not AD pathological change (A+T-) and was characterized high in cluster 2, patients in that subgroup had the low- by medial temporal atrophy, and a high burden of cer- est level of education (i.e., lower cognitive reserve). Hence, ebrovascular pathology. In addition, cluster 1 included lower levels of tau pathology may be enough to lead to Abdelnour et al. Alzheimer’s Research & Therapy (2022) 14:14 Page 10 of 13 low MMSE scores in a shorter time, at younger ages, all of cerebrovascular pathologies but not of amyloid-β were asso- these being findings that characterize cluster 2. Altogether, ciated with a lower frequency of probable RBD [8, 41]. disease duration was the shortest in both clusters 2 and 3, Interestingly, CSF amyloid-β was not one of the main the two clusters with greater posterior brain atrophy, sug- drivers in the dimensions of the FAMD model, but it was gesting a more aggressive presentation of the disease. The still a significant contributor to all three dimensions. Oppo- subtype of AD with greater posterior brain atrophy has site to having no contribution to the heterogeneity within been proposed as the most aggressive presentation of the DLB, this could indicate that CSF amyloid-β may be an disease, possibly due to a higher frequency of concomi- underlying factor in all dimensions, contributing to more tant AD and Lewy body pathology in that subtype of AD than one dimension at the same time. It is widely known [11]. Similarly, Poulakis et al. reported two subgroups with that the contribution of tau pathology to brain atrophy and posterior brain atrophy in AD, one with an older age (like cognitive impairment is stronger than that of amyloid-β cluster 3 in the current study) and one with a younger age pathology [58–60], which could explain the not so strong (like cluster 2 in the current study) [50]. contribution of CSF amyloid-β in our FAMD model. We observed statistically significant differences in the frequency of parkinsonism, visual hallucinations, and cog- Limitations nitive fluctuations across clusters that also had specific Our study has some limitations. Firstly, we used a retrospec- AD, cerebrovascular, and atrophy profiles. This result may tive and cross-sectional cohort, and longitudinal studies will have clinical implications. Clusters 1 and 4 showed the help to investigate the progression of these clusters over time. highest frequency of cognitive fluctuations, cluster 2 had Secondly, our approach was data-driven, and thus, our current the lowest frequency of visual hallucinations and cogni- findings should be considered hypothesis generating—replica- tive fluctuations, and cluster 3 had the lowest frequency tion in independent cohorts is warranted. Thirdly, although of parkinsonism. Hence, there seems to be an association our FAMD model included key factors that are known to between concomitant AD and cerebrovascular pathologies, explain DLB heterogeneity, other important biomarkers such and patterns of brain atrophy with clinical heterogeneity as DaTSCAN, MIBG, and EEG and supportive clinical fea- across subgroups of DLB patients. Although our current tures like postural instability, syncope, systematic delusions, study did not primarily aim to investigate associations autonomic dysfunction, and others may explain the additional of specific pathological features with particular clinical variance of DLB heterogeneity [1]. Those data were not availa- core features, the FAMD and hierarchical cluster analysis ble or were difficult to harmonize across centers in our current showed that patients with a low frequency of parkinson- study. To move the field forward, future studies should inves- ism usually have a higher frequency of abnormal CSF total tigate a wider range of variables in a prospective longitudinal tau and p-tau biomarkers. Our current data could help in cohort. Finally, given the multi-center nature of the current guiding future studies that target specific clinic-patholog- study, we cannot completely exclude that part of the heteroge- ical associations, perhaps using more detailed measures neity investigated is due to differences between centers. How- for both pathology and clinical features (as opposed to ever, it is difficult to separate between-center differences due the dichotomized variables in the current study). Previous to pure methodological reasons from between-center differ- studies have found that amyloid-β, tau, and cerebrovascular ences due to subpopulations with actual biological/phenotypi- pathologies are associated with a lower frequency of core cal differences, as recently discussed [61]. Our finding showing clinical features [51, 52] and a less typical presentation of that variables that are very well harmonized across centers DLB [53, 54]. We thus highlight the relevance of these find- (e.g., age, sex, education, MMSE) contributed to the dimen- ings, since they suggest that probable DLB patients with sions in a similar manner than variables that are traditionally concomitant AD or cerebrovascular pathologies may have more difficult to harmonize (e.g., biomarkers) is reassuring a higher risk to be misdiagnosed. The different presenta- and suggests that the influence of potential methodological tions of the disease with fast and slow progression rates differences across centers is likely not influencing our clusters. also signify the clinical relevance of these DLB subgroups. Indeed, combining data from several centers likely amplifies There were no differences in the frequency of probable the heterogeneity in clinically diagnosed DLB patients, which RBD across clusters. Few studies have investigated the influ- was the main focus in our study. Therefore, cross-collabora- ence of AD or cerebrovascular pathologies upon probable tion between specialized centers that used standard diagnos- RBD. Autopsy confirmed studies suggest that the burden tic procedures to reduce methodological differences across of concomitant AD in DLB patients is inverse to the fre- centers is a strength of our study, as well as the combination quency of RBD, meaning that patients with a clinical history of both clinical and biomarker data, reflecting current clinical of RBD have less AD-related pathology and a higher fre- practice in DLB. We leveraged these rich multimodal data by quency of diffuse Lewy body disease, and vice versa [55–57]. using a method for multimodal subtyping for the first time in Recent biomarker studies found that higher levels of tau and DLB, as far as we are aware. Abdelnour et al. Alzheimer’s Research & Therapy (2022) 14:14 Page 11 of 13 Conclusions Acknowledgements The authors are thankful to all members of the E-DLB consortium. This Our current study provides several insights on the contribu- research study was performed as part of the Medicine Doctoral Program of tors to the heterogeneity within probable DLB. The existence Carla Abdelnour at Universitat Autònoma de Barcelona (Barcelona, Spain). of subgroups of probable DLB has implications to clinical Authors’ contributions diagnosis. Furthermore, our preliminary data suggest the pos- CA, DF, and MvdB designed the study. CA, MvdB, KO, FB, OB, LOW, APadovani, sibility that these different subtypes may have their own dis- APilotto, AWL, MB, and DA participated in the acquisition of the data. LC made ease trajectories, and may need to be managed differently due the MRI assessments. DF, MvdB, and NC analyzed the data. MvdB and NC made the figures. CA, DF, MvdB, AWL, and EW interpreted the data. CA and DF to distinct combinations of core clinical features and concomi- drafted the manuscript. CA, DF, MvdB, NC, APilotto, JP, JK, DA, AWL, and EW tant AD and cerebrovascular pathologies. Therefore, future revised the manuscript. All authors read and approved the submitted version. studies investigating longitudinal data of DLB subgroups are Funding warranted. The recent development of the real-time quaking- This study was supported by The Swedish Foundation for Strategic Research induced conversion (RTQuIC), which accurately assesses (SSF), The Swedish Research Council (VR), Center forInnovative Medicine α-synuclein pathology in vivo [62], may also help in elucidat- (CIMED), The Foundation for Geriatric Diseases at Karolinska Institutet, Research Funding at Karolinska Institutet, Swedish Dementia Funding, The ing longitudinal associations between α-synuclein, AD, and Strategic Research Programme in Neuroscience at Karolinska Institutet cerebrovascular pathologies in the near future. We believe that (StratNeuro), The Åke Wiberg foundation, Swedish Brain Funding, Swedish consideration of this heterogeneity is a first step into imple- Alzheimer’s Funding, Stiftelsen Olle Engkvist Byggmästare, Birgitta och Sten Westerberg, Stiftelsen För Gamla Tjänarinnor, Gun och Bertil Stohnes Stiftelse, menting personalized medicine approaches in DLB. Likewise, and The Regional Agreement on Medical Training and Clinical Research (ALF) the presence of different subgroups of DLB may need to be between the Stockholm County Council and Karolinska Institutet. accommodated in the design of future clinical trials in DLB. Availability of data and materials The recent approval of aducanumab by the US Food and Drug The data that support the findings of this study are available from the E-DLB Administration posts the question about the potential benefit consortium but restrictions apply to the availability of these data, which were of anti-amyloid treatment in DLB patients with concomitant used under permission for the current study, and so are not publicly available. Anonymized data supporting the conclusions of the current study are avail‑ AD pathology. Our current findings suggest that the existence able to qualified researchers on reasonable request to the E-DLB consortium. of different DLB subgroups with possibly different responses to anti-amyloid treatment should be considered. Declarations Ethics approval and consent to participate Abbreviations An institutional ethics committee at each E-DLB center approved the study. AD: Alzheimer’s disease; ANOVA: Analysis of variance; Aβ: Beta-amyloid; CH: Informed consent on participation was obtained from all patients or an appro‑ Calinski-Harabasz; CIBERNED: Centro de Investigación en Red-Enfermedades priate surrogate according to the Declaration of Helsinki. Neurodegenerativas; CSF: Cerebrospinal fluid; DLB: Dementia with Lewy bod‑ ies; E-DLB: European dementia with Lewy bodies consortium; ELISA: Enzyme- Consent for publication linked immunosorbent assay; FAMD: Factorial analysis of mixed data; FLAIR: Not applicable. Fluid-attenuated inversion recovery; GCA-F: Frontal brain atrophy; IIB-Sant Pau: Biomedical Research Institute-Sant Pau; MMSE: Mini-Mental State Examination; Competing interests MRI: Magnetic resonance imaging; MTA: Medial temporal lobe atrophy; P-tau: C. Abdelnour has received fees for lectures from Zambon, KRKA, Roche, Nutricia, Phosphorylated tau; PA: Posterior brain atrophy; RBD: REM sleep behavior and Schwabe. M. Boada has consulted for Araclon, Avid, Grifols, Lilly, Nutricia, disorder; REM: Rapid eye movement; RT-QuIC: Real-time quaking-induced Roche, and Servier. She received fees for lectures and funds for research from conversion; SCI-lab: Sensory Cognitive Interaction Laboratory; WMH: White Araclon, Grifols, Nutricia, Roche, and Servier. D. Aarsland has received research matter hyperintensity. support and/or honoraria from Astra-Zeneca, H. Lundbeck, Novartis Pharma‑ ceuticals, and GE Health and served as a paid consultant for H. Lundbeck, Eisai, Heptares, and Mentis Cura. He is a Royal Society Wolfson Research Merit Award Supplementary Information Holder and would like to thank the Wolfson Foundation and the Royal Society The online version contains supplementary material available at https://​doi.​ for their support. Other authors declare that they have no competing interests. org/​10.​1186/​s13195-​021-​00946-w. Author details 1 Research Center and Memory Clinic, Ace Alzheimer Center Barcelona, Institut Additional file 1 Supplementary table 1: Overview of MRI parameters per Català de Neurociències Aplicades, Universitat Internacional de Catalunya- center. Supplementary table 2: Overview of CSF procedures per center. Barcelona, Centro de Investigación en Red-Enfermedades Neurodegenerativas These tables present detailed information about MRI parameters and CSF (CIBERNED), Barcelona, Spain. 2 Department of Medicine of the Universitat procedures per collaborating center. Autònoma de Barcelona, Barcelona, Spain. 3 Division of Clinical Geriatrics, Additional file 2 Supplementary methods. This file contains detail informa‑ Centre for Alzheimer Research, Department of Neurobiology, Care Sciences, tion of the supervised random forest classification model for the discrimi‑ and Society, Karolinska Institutet, Stockholm, Sweden. 4 Alzheimer Center nation of each cluster from all other clusters. Amsterdam, Department of Neurology, Amsterdam Neuroscience, Vrije Uni‑ versiteit Amsterdam, Amsterdam UMC, Amsterdam, The Netherlands. 5 Depart‑ Additional file 3 Supplementary Fig. 1. Random forest classification models ment of Psychology, Sensory Cognitive Interaction Laboratory (SCI‑lab), for the discrimination between a given cluster and all other clusters. This Stockholm University, Stockholm, Sweden. 6 Centre for Age‑Related Medicine, figure depicts the lollipops of importance to report the results of the Stavanger University Hospital, Stavanger, Norway. 7 Department of Radiology, random forest models for the discrimination between a given cluster and Stavanger University Hospital, Stavanger, Norway. 8 Department of Electrical all other clusters. Engineering and Computer Science, University of Stavanger, Stavanger, Nor‑ way. 9 Department of Neuroscience, Karolinska Institutet, Stockholm, Sweden. Abdelnour et al. Alzheimer’s Research & Therapy (2022) 14:14 Page 12 of 13 10 Department of Radiology Karolinska University Hospital, Stockholm, Swe‑ 14. McKeith IG, Dickson DW, Lowe J, Emre M, O’Brien JT, Feldman H, et al. den. 11 Service, Memory Resources and Research Centre, University Hospital Diagnosis and management of dementia with Lewy bodies: third report of Strasbourg, Strasbourg, France. 12 Team IMIS/Neurocrypto, French National of the DLB consortium. Neurology. 2005;65(12):1863–72. Center for Scientific Research, ICube Laboratory and Fédération de Médecine 15. McKeith I, Boeve B, Dickson D, Halliday G, Taylor J, Weintraub D, et al. Translationnelle de Strasbourg (FMTS), University of Strasbourg, Strasbourg, Diagnosis and management of dementia with Lewy bodies Fourth France. 13 Centre Mémoire, de Ressources et de Recherche d’Alsace (Stras‑ consensus report of the DLB Consortium. Neurology. 2017;89(1):88–100. bourg-Colmar), Strasbourg, France. 14 Laboratory of Biochemistry and Molecu‑ 16. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical lar Biology, CNRS, Laboratoire de Neurosciences Cognitives et Adaptatives, method for grading the cognitive state of patients for the clinician. J UMR7364, University Hospital of Strasbourg, Strasbourg, France. 15 Neurology Psychiatr Res. 1975;12(3):189–98. Unit, Department of Clinical and Experimental Sciences, University of Brescia, 17. Ferreira D, Verhagen C, Hernández-Cabrera JA, Cavallin L, Guo CJ, Ekman Brescia, Italy. 16 Movement Disorders Unit, Neurology Department, Hospital de U, et al. Distinct subtypes of Alzheimer’s disease based on patterns of la Santa Creu i Sant Pau. Biomedical Research Institute (IIB-Sant Pau), Centro brain atrophy: longitudinal trajectories and clinical applications. Sci Rep. de Investigación en Red-Enfermedades Neurodegenerativas (CIBERNED), 2017;7:46263. Barcelona, Spain. 17 Institute of Psychiatry, Psychology and Neuroscience, 18. Ph Scheltens D, Leys FB, Huglo D, Weinstein HC, Vermersch P, Kuiper M, King’s College London, London, UK. 18 Department of Neuroimaging, Centre et al. Atrophy of medial temporal lobes on MRI in “probable” Alzheimer’s for Neuroimaging Sciences, Institute of Psychiatry, Psychology and Neurosci‑ disease and normal ageing: diagnostic value and neuropsychological ence, King’s College London, London, UK. correlates. J Neurol Neurosurg Psychiatry. 1992;55:967–72. 19. Koedam ELGE, Lehmann M, Van Der Flier WM, Scheltens P, Pijnenburg Received: 20 September 2021 Accepted: 6 December 2021 YAL, Fox N, et al. Visual assessment of posterior atrophy development of a MRI rating scale. Eur Radiol. 2011;21(12):2618–25. 20. Ferreira D, Cavallin L, Granberg T, Lindberg O, Aguilar C, Mecocci P, et al. Quantitative validation of a visual rating scale for frontal atrophy: associa‑ tions with clinical status, APOE e4. CSF biomarkers and cognition. Eur References Radiol. 2016;26(8):2597–610. 1. McKeith IG, Boeve BF, Dickson DW, Halliday G, Taylor J-P, Weintraub D, 21. Fazekas F, Chawluk JB, Alavi A. MR signal abnormalities at 1.5 T in Alzhei‑ et al. Diagnosis and management of dementia with Lewy bodies: fourth mer’s dementia and normal aging. Am J Neuroradiol. 1987;8(3):421–6. consensus report of the DLB Consortium. Neurology. 2017;89(1):88–100. 22. Cedres N, Ferreira D, Machado A, Shams S, Sacuiu S, Waern M, Wahlund 2. Rongve A, Brønnick K, Ballard C, Aarsland D. Core and suggestive L-O, Zettergren A, Kern S, Ingmar Skoog EW. Predicting Fazekas scores symptoms of dementia with Lewy bodies cluster in persons with mild from automatic segmentations of white matter signal abnormalities. dementia. Dement Geriatr Cogn Disord. 2010;29(4):317–24. Aging (Albany NY). 2020;12(1):894–901. 3. Morenas-Rodríguez E, Sala I, Subirana A, Pascual-Goñi E, Sánchez- 23. Bos I, Verhey FR, Ramakers IHGB, Jacobs HIL, Soininen H, Freund-Levi Y, Saudinós MB, Alcolea D, et al. Clinical subtypes of dementia with Lewy et al. Cerebrovascular and amyloid pathology in predementia stages: the bodies based on the initial clinical presentation. J Alzheimer’s Dis. relationship with neurodegeneration and cognitive decline. Alzheimer’s 2018;64(2):505–13. Res Ther. 2017;9(1):101. 4. Jellinger KA, Wenning GK, Seppi K. Predictors of survival in demen‑ 24. Abdelnour C, van Steenoven I, Londos E, Blanc F, Auestad B, Kramberger tia with Lewy bodies and Parkinson dementia. Neurodegener Dis. MG, et al. Alzheimer’s disease cerebrospinal fluid biomarkers predict 2007;4(6):428–30. cognitive decline in lewy body dementia. Mov Disord. 2016;31(8):1203–8. 5. Halliday GM, Holton JL, Revesz T, Dickson DW. Neuropathology underly‑ 25. Jack CR Jr, Bennett DA, Blennow K, Carrillo MC, Dunn B, Budd Haeberlein ing clinical variability in patients with synucleinopathies. Acta Neuro‑ S, et al. NIA-AA research framework: toward a biological definition of pathol. 2011;122(2):187–204. Alzheimer’s disease. Alzheimers Dement. 2018;14(4):535–62. 6. Van Steenoven I, Aarsland D, Weintraub D, Londos E, Blanc F, Van Der 26. Kantarci K, Lowe VJ, Boeve BF, Senjem ML, Tosakulwong N, Lesnick TG, Flier WM, et al. Cerebrospinal fluid Alzheimer’s disease biomarkers across et al. AV-1451 tau and β-amyloid positron emission tomography imaging the spectrum of Lewy body diseases: results from a large multicenter in dementia with Lewy bodies. Ann Neurol. 2017;81(1):58–67. cohort on behalf of the European DLB consortium. J Alzheimers Dis. 27. Smith R, Schöll M, Londos E, Ohlsson T, Hansson O. F-AV-1451 in Parkin‑ 2016;54(1):287–95. son’s disease with and without dementia and in dementia with Lewy 7. van der Zande JJ, Steenwijk MD, ten Kate M, Wattjes MP, Scheltens P, bodies. Sci Rep. 2018;8(1):4717. Lemstra AW. Gray matter atrophy in dementia with Lewy bodies with and 28. Gomperts S, Locascio JJ, Makaretz SJ, Schultz A, Caso C, Vasdev N, et al. Tau without concomitant Alzheimer’s disease pathology. Neurobiol Aging. PET imaging in the Lewy body diseases. J AMA Neurol. 2016;73(11):1334–41. 2018;71:171–8. 29. Lê S, Josse J, Husson F. FactoMineR: an R package for multivariate analysis. 8. Ferreira D, Przybelski SA, Lesnick TG, Lemstra AW, Londos E, Blanc F, et al. J Stat Softw. 2008;25(1):1–18. β-Amyloid and tau biomarkers and clinical phenotype in dementia with 30. Ward JH. Hierarchical grouping to optimize an objective function. J Am Lewy bodies. Neurology. 2020;95(24):e3257–68. Stat Assoc. 1963;58(301):236–44. 9. Lemstra AW, De Beer MH, Teunissen CE, Schreuder C, Scheltens P, Van Der 31. Calinski T, Harabasz J. A dendrite method foe cluster analysis. Commun Flier WM, et al. Concomitant AD pathology affects clinical manifestation Stat. 1974;3(1):1–27. and survival in dementia with Lewy bodies. J Neurol Neurosurg Psychia‑ 32. Hochberg Y, Benjamini Y. More powerful procedures for multiple signifi‑ try. 2017;88(2):113–8. cance testing. Stat Med. 1990;9(7):811–8. 10. Oppedal K, Ferreira D, Cavallin L, Lemstra A, ten Kate M, Padovani A, et al. 33. Mouton A, Blanc F, Gros A, Manera V, Fabre R, Sauleau E, et al. Sex ratio in A signature pattern of cortical atrophy in dementia with Lewy bodies: a dementia with Lewy bodies balanced between Alzheimer’s disease and study on 333 patients from The European DLB Consortium. Alzheimer’s Parkinson’s disease dementia: a cross-sectional study. Alzheimer’s Res Dement. 2019;15(3):400–9. Ther. 2018;10(1):92. 11. Ferreira D, Nordberg A, Westman E. Biological subtypes of Alzheimer 34. Ferman TJ, Aoki N, Crook JE, Murray ME, Graff-Radford NR, van Gerpen disease. Neurology. 2020;94(10):436–48. JA, et al. The limbic and neocortical contribution of α-synuclein, tau, and 12. Oppedal K, Borda MG, Ferreira D, Westman E, Aarsland D. European DLB amyloid β to disease duration in dementia with Lewy bodies. Alzheimer’s consortium: diagnostic and prognostic biomarkers in dementia with Dement. 2018;14(3):330–9. Lewy bodies, a multicenter international initiative. Neurodegener Dis 35. Barnes LL, Lamar M, Schneider JA. Sex differences in mixed neuropathol‑ Manag. 2019;9(5):247–50. ogies in community-dwelling older adults. Brain Res. 2019;1719:11–6. 13. Kramberger MG, Auestad B, Garcia-Ptacek S, Abdelnour C, Olmo JG, Walker 36. Nelson PT, Schmitt FA, Jicha GA, Kryscio RJ, Abner EL, Smith CD, et al. Z, et al. Long-term cognitive decline in dementia with Lewy bodies in a Association between male gender and cortical Lewy body pathology in large multicenter, international cohort. J Alzheimer’s Dis. 2017;57(3):787–95. large autopsy series. J Neurol. 2010;257(11):1875–81. Abdelnour et al. Alzheimer’s Research & Therapy (2022) 14:14 Page 13 of 13 37. O’Dowd S, Schumacher J, Burn DJ, Bonanni L, Onofrj M, Thomas 59. Cho H, Choi JY, Hwang MS, Kim YJ, Lee HM, Lee HS, et al. In vivo cortical A, et al. Fluctuating cognition in the Lewy body dementias. Brain. spreading pattern of tau and amyloid in the Alzheimer disease spectrum. 2019;142(11):3338–50. Ann Neurol. 2016;80(2):247–58. 38. Nedelska Z, Ferman TJ, Boeve BF, Przybelski SA, Lesnick TG, Murray ME, 60. Bejanin A, Schonhaut DR, La Joie R, Kramer JH, Baker SL, Sosa N, et al. Tau et al. Pattern of brain atrophy rates in autopsy-confirmed dementia with pathology and neurodegeneration contribute to cognitive impairment in Lewy bodies. Neurobiol Aging. 2015;36(1):452–61. Alzheimer’s disease. Brain. 2017;140(12):3286–300. 39. Harper L, Bouwman F, Burton EJ, Barkhof F, Scheltens P, O’brien JT, et al. 61. Petrova T, Orellana C, Jelic V, Oeksengaard AR, Snaedal J, Høgh P, et al. Patterns of atrophy in pathologically confirmed dementias: a voxelwise Cholinergic dysfunction, neurodegeneration, and amyloid-beta pathol‑ analysis. J Neurol Neurosurg Psychiatry. 2017;88:908–16. ogy in neurodegenerative diseases. Psychiatry Res Neuroimaging. 40. Whitwell JL, Weigand SD, Shiung MM, Boeve BF, Ferman TJ, Smith GE, 2020;302:111099. et al. Focal atrophy in dementia with Lewy bodies on MRI: a distinct pat‑ 62. Fairfoul G, McGuire LI, Pal S, Ironside JW, Neumann J, Christie S, et al. tern from Alzheimer’s disease. Brain. 2007;130(3):708–19. Alpha-synuclein RT-QuIC in the CSF of patients with alpha-synucleinopa‑ 41. Ferreira D, Nedelska Z, Graff-Radford J, Przybelski SA, Lesnick TG, thies. Ann Clin Transl Neurol. 2016;3(10):812–8. Schwarz CG, et al. Cerebrovascular disease, neurodegeneration, and clinical phenotype in dementia with Lewy bodies. Neurobiol Aging. 2021;105:252–61. Publisher’s Note 42. Abdelnour C, Ferreira D, Oppedal K, Cavallin L, Bousiges O, Wahlund LO, Springer Nature remains neutral with regard to jurisdictional claims in pub‑ et al. The combined effect of amyloid-β and tau biomarkers on brain atro‑ lished maps and institutional affiliations. phy in dementia with Lewy bodies. NeuroImage Clin. 2020;27:102333. 43. Mak E, Donaghy PC, McKiernan E, Firbank MJ, Lloyd J, Petrides GS, et al. Beta amyloid deposition maps onto hippocampal and subiculum atro‑ phy in dementia with Lewy bodies. Neurobiol Aging. 2019;73:74–81. 44. Sarro L, Senjem ML, Lundt ES, Przybelski SA, Lesnick TG, Graff-Radford J, et al. Amyloid-b deposition and regional grey matter atrophy rates in dementia with Lewy bodies. Brain. 2016;139:2740–50. 45. Joki H, Higashiyama Y, Nakae Y, Kugimoto C, Doi H, Kimura K, et al. White matter hyperintensities on MRI in dementia with Lewy bodies, Parkinson’s disease with dementia, and Alzheimer’s disease. J Neurol Sci. 2018;385:99–104. 46. Jellinger KA, Attems J. Cerebral amyloid angiopathy in Lewy body dis‑ ease. J Neural Transm. 2008;115(3):473–82. 47. Smith R, Schöll M, Londos E, Ohlsson T, Hansson O. 18F-AV-1451 in Par‑ kinson’s disease with and without dementia and in dementia with Lewy bodies. Sci Rep. 2018;8(1):4717. 48. Gomperts SN, Locascio JJ, Makaretz SJ, Schultz A, Caso C, Vasdev N, et al. Tau positron emission tomographic imaging in the Lewy body diseases. JAMA Neurol. 2016;73(11):1334–41. 49. Irwin DJ, Grossman M, Weintraub D, Hurtig HI, Duda JE, Xie SX, et al. Neu‑ ropathological and genetic correlates of survival and dementia onset in synucleinopathies: a retrospective analysis. Lancet Neurol. 2017;16(1):55. 50. Poulakis K, Ferreira D, Pereira JB, Smedby Ö, Vemuri P, Westman E. Fully Bayesian longitudinal unsupervised learning for the assessment and visualization of AD heterogeneity and progression. Aging (Albany NY). 2020;12(13):12622–47. 51. Tiraboschi P, Attems J, Thomas A, Brown Evelyn Jaros A, Lett DJ, Maria Ossola B, et al. Clinicians’ ability to diagnose dementia with Lewy bodies is not affected by β-amyloid load. Neurology. 2015;84(5):496–9. 52. Di Censo R, Abdelnour C, Blanc F, Bousiges O, Lemstra AW, van Steenoven I, et al. CSF tau proteins correlate with an atypical clinical presenta‑ tion in dementia with Lewy bodies. J Neurol Neurosurg Psychiatry. 2020;91(1):109–10. 53. Del Ser T, Hachinski V, Merskey H, Munoz DG. Clinical and pathologic fea‑ tures of two groups of patients with dementia with Lewy bodies: effect of coexisting Alzheimer-type lesion load. Alzheimer Dis Assoc Disord. 2001;15(1):31–44. 54. Weisman D, Cho M, Taylor C, Adame A, Thal LJ, Hansen LA. In dementia with Lewy bodies, Braak stage determines phenotype, not Lewy body Ready to submit your research ? Choose BMC and benefit from: distribution. Neurology. 2007;69(4):356–9. 55. Murray ME, Ferman TJ, Boeve BF, Przybelski SA, Lesnick TG, Liesinger AM, • fast, convenient online submission et al. MRI and pathology of REM sleep behavior disorder in dementia with Lewy bodies. Neurology. 2013;81(19):1681–9. • thorough peer review by experienced researchers in your field 56. Ferman TJ, Aoki N, Boeve BF, Aakre JA, Kantarci K, Graff-Radford J, et al. • rapid publication on acceptance Subtypes of dementia with Lewy bodies are associated with α-synuclein • support for research data, including large and complex data types and tau distribution. Neurology. 2020;95(2):E155–65. 57. Pilotto A, Romagnolo A, Tuazon JA, Vizcarra JA, Marsili L, Zibetti M, et al. • gold Open Access which fosters wider collaboration and increased citations Orthostatic hypotension and REM sleep behaviour disorder: impact on • maximum visibility for your research: over 100M website views per year clinical outcomes in α-synucleinopathies. J Neurol Neurosurg Psychiatry. 2019;90(11):1257–63. At BMC, research is always in progress. 58. La JR, Visani AV, Baker SL, Brown JA, Bourakova V, Cha J, et al. Prospective longitudinal atrophy in Alzheimer’s disease correlates with the intensity Learn more biomedcentral.com/submissions and topography of baseline tau-PET. Sci Transl Med. 2020;12(524):5732.