Journals
2025 EN
Napoleon Darian A. · Fausto Bernadette A. · Piaszczynska Wiktoria
+17 more
Abstract Background Alzheimer's disease (AD) and type 2 diabetes (T2D) share common pathophysiological mechanisms and cognitive sequelae. A potential mechanism by which T2D raises AD risk among aging adults relates to alterations in the medial temporal and prefrontal cortices (Napoleon et al., 2024). However, the relationship of hyperglycemia with neuropathological biomarkers of AD is understudied. Here, we examined the effect of T2D status on blood‐based neuropathology biomarkers in a cohort of cognitively unimpaired older adults of African ancestry. Method Cognitively unimpaired older adults (N = 134; 71.87±7.06y; 78.4% female) were recruited from the Aging & Brain Health Alliance at Rutgers University–Newark. Individuals were categorized based on T2D status using hemoglobin A1c (≥6.5%, n=28) or non‐T2D (≤6.4%, n=106). Blood samples were analyzed for the following neuropathology biomarkers: plasma p ‐tau217 (an early marker of AD neuropathology, elevated in individuals with amyloid‐β [Aβ] deposition); neurofilament‐light chain (NfL, a neuronal cytoplasmic protein that accumulates proportionally to the degree of axonal damage in a variety of neurology disorders; Aβ42/40 ratio (decreased levels are reflective of higher AD risk); and glial fibrillary acidic protein (GFAP; provides structural stability to astrocytes; elevated levels serve as a non‐specific biomarker for neurological diseases). ANCOVAs were used to examine differences between groups on neuropathology markers with age, gender, waist‐to‐hip ratio, and diabetes prescription medication use (0=no, 1=yes) as covariates. Result Participants with T2D had significantly higher p ‐tau217 levels (η p 2 = 0.03, p = .043), independent of covariates. There were no significant group differences on NfL levels (η p 2 = 0.01, p = .530), Aβ42/40 ratio (η p 2 = 0.01, p = .96), nor GFAP (η p 2 = 0.01, p = .54). Conclusion T2D was associated with higher levels of plasma p ‐tau‐217 compared with those who do not have T2D. This suggests T2D may promote early pathological changes that raise AD risk in cognitively unimpaired older adults (Ashton et al., 2022). These results build on our earlier work demonstrating that T2D and associated hyperglycemia may lower medial temporal lobe cognitive function in cognitively unimpaired older adults, which may be mediated by increased AD neuropathology.
Journals
2025 EN
van Dyck Christopher H · Li David · Kanekiyo Michio
+5 more
Abstract Background Lecanemab is a humanized IgG1 monoclonal antibody that binds with high affinity to amyloid‐beta (Aβ) protofibrils. In the 18‐month, phase 3 Clarity AD study, lecanemab demonstrated amyloid reduction and cognition and function decline slowing in participants with early symptomatic Alzheimer’s disease (AD). An ongoing open‐label extension (OLE) of Clarity AD is evaluating long‐term safety and efficacy of lecanemab. Herein, we report the initial findings up to 48 months from the ongoing Clarity AD OLE study. Method Clarity AD is an 18‐month, randomized study (Core) followed by an OLE phase in individuals with early AD. Clinical (CDR‐SB, ADAS‐Cog14, and ADCS‐MCI‐ADL) outcomes and safety were evaluated from OLE data out to 48 months. Continued lecanemab treatment in the OLE beyond the 18 months from the Core study was compared to a matched control from Alzheimer's Disease Neuroimaging Initiative (ADNI) data. Subgroup analyses were conducted for participants with no/low baseline tau. Efficacy assessments were also summarized as the percentage of participants who had ‘no decline’ or had ‘improvement’ from Core baseline at each timepoint. Result Overall, 1734 participants were treated with lecanemab across the Core and OLE. Across clinical endpoints, lecanemab‐treated participants continued to benefit through 48 months. In the OLE, lecanemab treatment continued to delay progression through 48 months compared to ADNI matched control, with differences in CDR‐SB increasing over 18 months (0.52), 36 months (1.01), and 48 months (1.75). Consistent rates of clinical stability or improvements were observed across assessments regardless of baseline tau levels, with the highest rates of improvements observed for the no/low tau group at 48 months (CDR‐SB: no decline:69%, improvement:56%; ADAS‐Cog14: no decline:51%, improvement:51%; ADCS MCI‐ADL: no decline:64%, improvement:58%). No new safety signals were observed. ARIA rates were low and similar to ARIA rates on placebo after 6 months. Conclusion In Clarity AD, the observed increasing treatment difference with ongoing lecanemab treatment in participants treated through 48 months versus ADNI matched placebo data is consistent with a durable disease‐modifying effect. The long‐term safety profile of lecanemab was confirmed, with no new safety signals.
Journals
2025 EN
Horie Kanta · Koyama Akihiko · Sachdev Pallavi
+8 more
Abstract Background Soluble amyloid‐β protofibrils (Aβ‐PF) are synaptotoxic species and are in equilibrium with amyloid plaques. Lecanemab is a humanized IgG1 monoclonal antibody that binds with high affinity to Aβ protofibrils. We developed a sensitive immunoassay using lecanemab as the capture antibody that selectively quantifies Aβ‐PF in cerebrospinal fluid (CSF), and found CSF Aβ‐PF levels significantly increased in Alzheimer’s disease (AD) and correlated with neurodegeneration biomarkers (Shinohara‐Noguchi, Ann Neurol. 2025). In this report, we further characterized CSF Aβ‐PF in the Phase 3 study of lecanemab (Clarity AD): (1) at baseline, (2) longitudinally in placebo arm (natural history), and (3) demonstrated the target engagement ( in vivo protofibril binding) of lecanemab. Method CSF samples collected from the Clarity AD were analyzed using the developed immunoassay. The assay is tolerant to therapeutic lecanemab in the CSF, and measures total Aβ‐PF (bound and unbound). 410 subjects (Placebo=207, Lecanemab=203) had a CSF sample at baseline and 253 subjects (Placebo=127, Lecanemab=126) had a CSF sample at baseline and at least one post‐baseline sample. Other biomarker assessments including amyloid‐PET, and CSF biomarkers including neurogranin were described previously and used for correlation analyses. Result Baseline CSF Aβ‐PF showed significant correlation with neurogranin, a biomarker of synaptic dysfunction (r=0.153, p =0.0127). In the placebo arm, there was an increase from baseline in CSF Aβ‐PF levels which significantly correlated with the increase from baseline observed in CSF neurogranin levels. In the lecanemab‐treatment arm, the CSF total Aβ‐PF levels (free and bound) increased relative to placebo at each of the post‐baseline assessments ( p =0.0126 at 12 months and numerically higher at 18 month). In the subjects achieving amyloid‐negativity defined as Centiloid<30 at 18 months, lecanemab‐treatment group demonstrated a larger percent change from baseline compared with subjects who remained amyloid‐positive (Centiloid≥30) at 18 months. Conclusion The novel immunoassay for the CSF Aβ‐PF indicates that soluble Aβ‐PF is quantifiable in human CSF in vivo . Increase in total (free and bound) CSF Aβ‐PF with lecanemab‐treatment is consistent with binding of Aβ‐PF to lecanemab (target engagement), and mobilization of Aβ‐PF from amyloid plaques (pharmacodynamic effect), confirming the unique dual mechanism of lecanemab targeting toxic Aβ‐PF and plaques.
Journals
2025 EN
Torso Mario · Ridgway Gerard R · Khosropanah Pegah
+2 more
Abstract Background Atherosclerosis of the Circle of Willis (ACW) is a common comorbidity in Alzheimer’s disease (AD). Therefore, in the era of disease‐modifying treatments understanding how ACW may influence cortical neurodegeneration driven by AD neuropathological processes is valuable. By investigating the impact of ACW on cortical microstructural changes, we aim to identify potential cortical patterns to improve diagnostic precision, offering better‐targeted therapeutic approaches for individuals with combined vascular and neurodegenerative pathologies. Method Structural and diffusion ante mortem MRI scans of 58 participants with intermediate or severe Alzheimer's disease neuropathologic change (ADNC) were obtained from the National Alzheimer's Coordinating Center (NACC). Participants were grouped by ACW severity (37 with none/mild and 21 moderate/severe) (Table 1). Macrostructural MRI metrics (cortical volume and cortical thickness) and three minicolumn‐related diffusivity metrics were extracted: the angle between the radial minicolumnar direction and the principal diffusion direction (AngleR); the principal diffusion component parallel with the minicolumns (ParlPD), and the diffusion components perpendicular to the minicolumns (PerpPD + ) (Torso et al. 2022, PMID:36281682). The groups were compared to investigate potential differences in clinical, demographic vascular risk factors and whole‐brain MRI macrostructural data. Regional differences in macrostructural (Cortical volume and cortical thickness) and diffusion metrics were analyzed using a linear model, adjusted for the interval between the MRI scan and autopsy dates, acquisition protocol, age, and sex. The results were corrected for multiple comparisons using the false discovery rate (pFDR < 0.05). Result Regional analysis revealed that participants with higher ACW severity exhibited a significant lower ParlPD values in the bilateral temporal and occipital regions (Figure 1), as well as significantly lower PerpPD + values in the left occipital regions. No differences were found in demographic, clinical, vascular risk factor and whole‐brain MRI macrostructural data when comparing the two groups. Conclusion These findings highlight the potential of cortical diffusivity in detecting distinct patterns of microstructural changes in individuals with the same ADNC severity but varying levels of ACW severity. This underscores the utility of cortical diffusivity for patient stratification in clinical practice and trials.
Journals
2025 EN
Flaherty Ryn · Sui Yu Veronica · Masurkar Arjun V.
+5 more
Abstract Background Our prior work applying diffusion kurtosis imaging (DKI) to subjective cognitive decline (SCD) showed significantly decreased kurtosis fractional anisotropy (KFA) and significantly increased mean kurtosis (MK) for SCD in bilateral amygdala, an early site for tau tangle pathology 1 . However, the microstructural alterations driving these differences are unclear. Here, we assess associations of MK and KFA with neurite orientation dispersion and density imaging (NODDI) and magnetization transfer imaging (MTI) metrics, which provide a more specific characterization of tissue microstructure. Method 175 cognitively normal participants from Cam‐Can 2 (75 SCD) ages 55‐88 were included in the analysis. Participants were defined as SCD if they endorsed problems with their memory and in the control group otherwise. Diffusion images were processed to obtain MK, Radial Kurtosis (RK), Axial Kurtosis (AK) and KFA from DKI and Neurite Density (ND) and Orientation Dispersion (OD), a marker of neurite organization, from NODDI. MTI was used to calculate the Magnetization Transfer Ratio (MTR), a marker of myelin and potentially amyloid aggregation. Mean metric values were calculated for bilateral amygdala (Figure 1). Between‐group comparisons were conducted using Wilcoxon rank‐sum tests, corrected for multiple testing. Associations between DKI, NODDI, and MTR metrics were examined using linear models corrected for age and sex. Result SCD had lower KFA, higher MK, and higher RK in the right amygdala (Table 1). KFA had a weak negative correlation with ND, while MK and RK had strong positive correlations (Figure 2A‐C). Only MK had a weak positive correlation with OD (Figure 2D‐F). Neither KFA, MK, nor RK correlated with MTR (Figure 2G‐I). Conclusion DKI metrics are more sensitive to amygdala changes in SCD than NODDI metrics or MTR. Lower KFA, higher MK, and higher RK were associated with higher ND but not MTR, suggestive of dendritic or glial branching. Higher MK was additionally associated with higher OD, potentially indicating reduced neurite organization. Further analyses on the impact of these amygdala changes on SCD related neuropsychiatric symptoms are needed. 1. Flaherty R, Sui YV, Li M, et al. Alzheimers Dement. 2024;20(S9):e093982. 2. Shafto MA, Tyler LK, Dixon M, et al. BMC Neurol. 2014;14(1):1‐25.
Journals
2025 EN
Ridgway Gerard R · Nakajima Takashi · Torso Mario
+5 more
Abstract Background Measures of cortical microstructure from diffusion MRI (dMRI) have been shown to relate to amyloid pathology and to neuroinflammation, and can predict subsequent macrostructural atrophy (Torso et al., 2022, PMID:36281682). We have previously shown practicality of these methods using 1.5T MRI in a real‐world hospital setting (Ridgway, Nakajima, et al., 2023, AAIC; jRCT1032210367). Anti‐amyloid treatments such as Leqembi (Lecanemab) have been shown to reduce rates of clinical deterioration, but are associated with accelerated brain volume loss, which is presumed to be transient, but is not yet fully understood. We have commenced a study ( https://jrct.niph.go.jp/en‐latest‐detail/jRCT1031240123 ) to explore the utility of dMRI measures to support early diagnosis and to track treatment response. Information on tissue microstructure is expected to add insights regarding volumetric atrophy or pseudoatrophy. Method Patients are recruited at National Hospital Organization Niigata National Hospital, with inclusion and exclusion criteria based on those of the approved drug Lecanemab. Amyloid positivity is confirmed with either amyloid PET or CSF Aβ42/40 ratio < 0.067. T1‐weighted anatomical MRI (T1w) and dMRI (32‐directions at b=1000 s/mm 2 ) are acquired on a 1.5T Philips Ingenia MRI scanner and processed using FreeSurfer, FSL, and proprietary software to produce minicolumn‐associated microstructural measures (AngleR, ParlPD and PerpPD + ; McKavanagh et al., 2019, PMID:31355989). Result The study has enrolled 17 AD patients at the time of analysis, summarised in Table 1. Patients were scanned at baseline, and after 2, 3 and 6 months. 59 total scans were analysed, including 13 6‐month scans. Figure 1 shows baseline associations of cognition (MMSE; lower scores worse) and cortical disarray (AngleR, higher values worse) with p ‐tau (higher values worse) for cases with CSF. Figure 2 shows regional associations of change in PerpPD+ with an overall improvement score that averaged MMSE, MMSE‐DR, CDR‐GS, CDR‐SB and IADL. Conclusion We have introduced a promising real‐world observational study that could shed light on the microstructural nature of amyloid‐lowering treatment response and the debate about pseudo‐atrophy. Microstructural imaging measures are sensitive treatment biomarkers that correlate with clinical changes. Future work with a larger dataset will investigate the potential prediction of clinical change using baseline cortical disarray measurements.
Journals
2025 EN
Torso Mario · Ridgway Gerard R · Radhakrishnan Hamsi
+6 more
Abstract Background Frontotemporal dementia (FTD) can arise from frontotemporal lobar degeneration (FTLD) driven by distinct proteinopathies, such as tau (FTLD‐tau) or TDP‐43 (FTLD‐TDP), which can lead to remarkably similar clinical syndromes. Previous research (PMID:34997851) identified a predominance of tau pathology in the lower cortical layers and TDP‐43 pathology in the upper cortical layers. The aim of the present study was to investigate the effect of laminar distribution of pathology on cortical architecture using cortical diffusivity metrics. Method Forty cases with a primary bvFTD clinical phenotype and autopsy confirmation from Penn Frontotemporal Degeneration Center were included in the study. The patients were grouped based on the primary neuropathological diagnosis: 20 FTLD‐tau and 20 FTLD‐TDP (Table‐1). Structural and diffusion MRI (dMRI) were used to calculate whole‐brain and regional cortical diffusivity measures. For each cortical region and for four macroregions (idiotypic M1, paralimbic association, association and idiotypic V1) previously explored (PMID:34997851) (Figure 1), a minicolumn‐inspired cortical diffusivity measure that combines the components perpendicular to the radial minicolumns (PerpPD + ) was calculated (PMID:36281682). Previous findings have shown that this measure is sensitive to tau neuropathology (PMID:37794477). For a subgroup of 17 cases, ratios of layer pathology (RLP) in four cortical regions (Figure‐1) were generated. Differences in diffusion metrics at regional and macroregional level were tested with a linear model adjusting for interval between MRI scan date and autopsy date, acquisition protocol, disease duration, age and sex, with false discovery rate correction (pFDR<0.05). Partial Spearman’s rank correlation was conducted, including cases from both groups, to investigate the relationship between regional PerpPD + and RLP values. Result Regional analysis showed a significant pattern of higher PerpPD + values in FTLD‐tau group, involving mainly fronto‐temporal regions (Figure‐2). Macroregion comparisons revealed higher PerpPD + in bilateral paralimbic and right association regions (Figure‐1). Correlation analysis identified significant associations between ratio (RLP) values and PerpPD + values in anterior cingulate, superior/middle temporal and primary visual cortex (Figure‐1). Conclusion Regional differences suggest that PerpPD + can distinguish cortical microstructural changes in FTLD due to tau vs. TDP‐43. The correlation between PerpPD + values and laminar pathology ratios reinforces that PerpPD + is an MRI marker sensitive to tau pathology distribution.
Journals
2025 EN
WilliePermor Daniel · Lopez Oscar L · Reis Steven E.
+8 more
Abstract Background Cognitive impairment (CI) in Alzheimer’s disease (AD) is driven, among others in the AT(V)N framework, by tau and amyloid pathology, but individual responses to these pathologies vary widely. Some individuals maintain cognitive function despite significant pathology (resilience), while others remain pathology‐free and cognitively intact (resistance). Emerging evidence suggests that health behavior factors, such as physical activity, cognitive engagement, and social interaction, along with demographic and clinical variables, may play a protective role. Our aim was to identify predictors of resistance and resilience. Method This study included 152 participants from the Heart Strategies Concentrating on Risk Evaluation (Heart SCORE) study with longitudinal imaging data and cognitive assessments. Tau burden was measured using [18F]Flortaucipir (FTP) and amyloid burden using [11C]Pittsburgh Compound‐B (PiB) PET imaging. Cognition was categorized clinically as unimpaired or impaired (MCI or dementia) cognition. Resistance and resilience profiles were developed by combining tau, amyloid, and CI status (e.g., tau‐resilient: tau‐positive, CI‐negative; tau‐resistant: tau‐negative, CI‐negative). Stepwise logistic regression with a p ‐value cutoff of 0.2 was used to identify predictors of tau and amyloid resistance and resilience. Discrete‐time survival analysis (logistic regression) was used to examine transitions to MCI/dementia. Model performance was assessed using the Area Under the ROC Curve (AUC) for discrimination and calibration plots for the accuracy of predicted probabilities. Calibration plots were generated using local regression (loess) and evaluated for internal validity. Result The demographic, clinical and health behavior characteristics of participants stratified by cognition status at baseline visit are summarized in Table 1. White race, history of CVD, cognitive activity were positively associated with tau resilience, while leisure activity was inversely related. Hypercholesterolemia showed an inverse association with tau resistance. Social activity, leisure activity, household activity, and education were positively associated with amyloid resistance. Within every year of study follow‐up, amyloid positivity and tau‐positivity predicted higher odds of progression to MCI/dementia, while CVD history, White race and cognitive activity predicted reduced lower odds. (See Table 2 & Figure 1 for model discrimination/calibration). Conclusion Tau and amyloid status influence cognitive transitions, with health behavior factors, demographic and CVD history playing key roles in resistance and resilience.
Journals
2025 EN
Sanchez Ana Bea Solana · Fernandez Brice · Lythgoe David J
+13 more
Abstract Background The PREDICTOM (Prediction of Neurodegenerative Disease using an AI driven Screening Platform – www.predictom.eu ) project is an EU funded project aimed at developing new biomarkers to identify people at high risk of developing Alzheimer Disease (AD) (1). Here, we aim to describe the rationale behind the MR protocol that will be used at the seven different clinical centers across Europe in the study. Method The PREDICTOM study will collect MR data from 615 subjects (target ratio: 80/20% at higher/lower risk). The targeted population will be healthy and over 50 years old. Seven clinical centers across Europe equipped with 3T MR scanners from two different vendors and six different MR models will acquire the MR data. Learning from ADNI4 (2), the PREDICTOM MR protocol is designed to target only modern clinical MR scanners. It is intended to be a comprehensive protocol that allows extracting the most promising anatomical and functional features for AD. As PREDICTOM is a multicenter study, the protocol must be harmonized to reduce the bias across centers/scanners to a reasonable level. The total exam time will not exceed 1 hour including the subject positioning. Result PREDICTOM MRI protocol contains eight different datatypes. Figure 1 and Figure 2 contain the main parameters and exemplary images in one healthy volunteer scanned in one of the PREDICTOM clinical centers. The choice of QSM was driven by its performance to better detect microbleeds (3). The MRS acquisition has been included due to consistent evidence of alterations in N‐acetylaspartate and myo‐Inositol in early stages of AD (4). For scanner with lower gradient performance, the rsfMRI protocol might use slightly partial Fourier and the multi‐shell diffusion will be reduced to 2‐shell protocol, excluding b=2000. Conclusion The protocol has been tested on 4 scanners (2 vendors, 3 models) and the harmonisation is still in progress. The PREDICTOM MR protocol was designed to allow the exploration of existing and novel MRI biomarkers for early identification of people at high risk of developing AD. (1) Brem AK, AAIC2024, (2) Arani A, Alzheiemer´s & dementia2024, (3) Lee K, Neuroimaging2023, (4) Maul S, Frontiers in Psychiatry2020.
Journals
2025 EN
TristãoPereira Catarina · Niño David Fernando Aguillón · Baena Ana Y
+15 more
Abstract Background Cerebral glucose hypometabolism, measured with 18 F‐fluorodeoxyglucose (FDG) PET, is a hallmark of Alzheimer’s disease (AD). While FDG‐PET changes are highly effective at differentiating dementia types, the pathological bases for metabolic changes are complex. Hypometabolism is often assumed to reflect decreased neuronal activity, but glucose uptake is not limited to neurons. Astrocytes consume about half of the glucose‐derived brain energy and have been recognized as major contributors to the FDG‐PET signal. This study investigates the differential contribution of plasma glial fibrillary acidic protein (GFAP), a protein upregulated in reactive astroglia, and neurofilament light chain (NfL), a neuron‐specific cytoskeleton component, to FDG‐PET in autosomal‐dominant AD. Method We included 40 Presenilin‐1 E280A mutation carriers (37.9±7.6 years, 26 females, 5 cognitively impaired) and 37 cognitively unimpaired non‐carriers (39.5±6.9 years, 23 females) from the Colombia‐Boston (COLBOS) Biomarkers Study (Table 1). Plasma GFAP and NfL were quantified using Neurology 4‐Plex E Advantage Kits (Quanterix). FDG uptake was processed across Freesurfer‐derived regions of interest. Spearman correlation was used to examine the associations between plasma biomarkers and regional FDG uptake, correcting for multiple comparisons. Lasso regression and mediation analysis identified the plasma biomarker most strongly linked to global FDG‐PET uptake (Landau meta‐ROI). Result Mutation carriers had higher plasma GFAP and NfL levels than non‐carriers (ps<0.001). Among carriers, both plasma GFAP and NfL showed negative correlations with FDG uptake, following a similar regional pattern that included temporo‐parietal regions (uncorrected). Notably, the associations with FDG uptake in the hippocampus remained significant after multiple comparisons correction (GFAP: p adj =0.014; NfL: p adj =0.005). In the joint Lasso model, plasma GFAP (b=‐0.082 [‐0.156,‐0.002]), but not plasma NfL (b=‐0.063 [‐0.137,0.032]), remained associated with global FDG uptake. Plasma GFAP showed a direct effect on global FDG uptake, independently of plasma NfL (indirect effect: p = 0.153; direct effect: p = 0.021). Conclusion Our data suggest that the FDG‐PET signal may reflect both reactive astrogliosis and neuronal injury in AD, with the effect of reactive astrogliosis on hypometabolism being independent of neurodegeneration. With astrocytes interfacing neurovascular communication, underlying vascular or neuroinflammatory mechanisms could contribute to hypometabolism and early disease onset, calling for a more cautious interpretation of brain FDG‐PET imaging.