Gebre, Robel K

Abstract There is increasing interest in Alzheimer’s disease related plasma biomarkers due to their accessibility and scalability. We hypothesized that integrating plasma biomarkers with other commonly used and available participant data (MRI, cardiovascular factors, lifestyle, genetics) using machine learning (ML) models can improve individual prediction of cognitive outcomes. Further, our goal was to evaluate the heterogeneity of these predictors across different age strata. This longitudinal study included 1185 participants from the Mayo Clinic Study of Aging who had complete plasma analyte work-up at baseline. We used the Quanterix Simoa immunoassay to measure neurofilament light, Aβ1–42 and Aβ1–40 (used as Aβ42/Aβ40 ratio), glial fibrillary acidic protein, and phosphorylated tau 181 (p-tau181). Participants’ brain health was evaluated through gray and white matter structural MRIs. The study also considered cardiovascular factors (hyperlipidemia, hypertension, stroke, diabetes, chronic kidney disease), lifestyle factors (area deprivation index, body mass index, cognitive and physical activities), and genetic factors (APOE, single nucleotide polymorphisms, and polygenic risk scores). An ML model was developed to predict cognitive outcomes at baseline and decline (slope). Three models were created: a base model with groups of risk factors as predictors, an enhanced model included socio-demographics, and a final enhanced model by incorporating plasma and socio-demographics into the base models. Models were explained for three age strata: younger than 65 years, 65–80 years, and older than 80 years, and further divided based on amyloid positivity status. Regardless of amyloid status the plasma biomarkers showed comparable performance (R² = 0.15) to MRI (R² = 0.18) and cardiovascular measures (R² = 0.10) when predicting cognitive decline. Inclusion of cardiovascular or MRI measures with plasma in the presence of socio-demographic improved cognitive decline prediction (R² = 0.26 and 0.27). For amyloid positive individuals Aβ42/Aβ40, glial fibrillary acidic protein and p-tau181 were the top predictors of cognitive decline while Aβ42/Aβ40 was prominent for amyloid negative participants across all age groups. Socio-demographics explained a large portion of the variance in the amyloid negative individuals while the plasma biomarkers predominantly explained the variance in amyloid positive individuals (21% to 37% from the younger to the older age group). Plasma biomarkers performed similarly to MRI and cardiovascular measures when predicting cognitive outcomes and combining them with either measure resulted in better performance. Top predictors were heterogeneous between cross-sectional and longitudinal cognition models, across age groups, and amyloid status. Multimodal approaches will enhance the usefulness of plasma biomarkers through careful considerations of a study population’s socio-demographics, brain and cardiovascular health.

AbstractBackgroundThe mapping between risk factors and biomarkers with cognitive performance and decline in the older adults is complex and varies significantly across different age and sex groups. Our goals were to (i) develop a machine learning (ML) model to predict baseline cognition and cognitive decline over a 5‐year follow‐up using only baseline assessments, and (ii) identify important modifiable and non‐modifiable predictors of cognitive decline across different age strata.MethodWe utilized longitudinal data from 2219 participants (mean age (SD): 70 (12) years, 49% male) from Mayo Clinic Study of Aging who had 43 features belonging to cardiovascular health, plasma, MRI (diffusion and structural), and lifestyle measures. Cognition was measured using a global cognition z‐score. We built ML models to predict baseline cognition and rate of cognitive decline over 5‐years (linear regression slopes over all individual followup measurements for each participant) using baseline assessments as inputs. We then built a model for each category separately and as a pooled model with all measures included, incorporating education and sex as demographic features in all models (Fig. 1). We applied a multi‐layer stack ensemble ML technique that utilized repeated k‐fold bagging to maximize accuracy. Importance of the top predictors were assessed in three age strata (55‐65 years (31%), 65‐80 years (46%), 80‐95 years (23%)).ResultThe pooled model explained the largest variance for prediction of baseline cognition and cognitive decline (Fig. 2). In individuals ≤80 years, lower NfL was most predictive of higher baseline cognition. A combination of imaging and plasma features predicted the 5‐year cognitive decline. Additionally, in individuals ≤65 years, diabetes and hypertension were important predictors of faster cognitive decline and in those >80, lower BMI and midlife physical/cognitive activity was an important predictor of faster cognitive decline (Fig. 3).ConclusionThe plasma model was best for baseline cognition prediction and imaging model was best for cognitive decline prediction. There was significant heterogeneity in important features across the age strata. In individuals >65 years, plasma GFAP was the most important predictor of cognitive decline. Our model allowed identification of key risk factors that predicted cognitive decline.

AbstractBackgroundSmall vessel disease (SVD) is a common vascular contributor to cognitive impairment and dementia. In this work, we evaluated and compared non‐imaging measures that can predict variability in SVD related white matter (WM) damage in a population‐based sample.MethodWe identified 752 participants in the population‐based sample of Mayo Clinic Study of Aging (≥50 years of age), who had at least 2 diffusion tensor imaging (DTI) scans and white matter hyperintensity (WMH) assessments along with baseline demographics, systemic vascular health, lifestyle, motor impairment, behavioral symptoms, and plasma markers (Summarized in Figure 1). We computed the WM markers of SVD: fractional anisotropy of genu of the corpus callosum (Genu‐FA) representative of early WM changes, WMH representative of late WM changes, and Vascular WM score (combination of Genu‐FA and WMH). Using all available data, we fit linear mixed effect models and evaluated predictors of WM markers by considering each category of features independently (with demographics) as well as by combining all categories of featuresResultPlasma, gait, systemic vascular, and life‐style measures were significant predictors of all three WM markers of SVD (Figure 2). The predictors of early and late WM slightly differed such that there were no additional significant contributions from behavioral and plasma measures to Genu‐FA. In the final composite models shown in Table 1, higher gait speed, lower number of cardiovascular and metabolic conditions (CMC) and lower Unified Parkinson’s disease rating scale scores were associated with better Genu‐FA whereas higher CMC, Beck Anxiety Inventory scores (BAI), and glial acidic protein (GFAP) levels were associated with higher burden of WMH.ConclusionOur observations evaluated the non‐imaging markers of greater WM damage that are relevant to vascular contributions to cognitive impairment to dementia (VCID). We found that in addition to higher vascular risk, worse plasma GFAP, gait measures, and BAI scores were predictors of greater WM damage. Some non‐imaging markers were more associated with early WM changes and others with late WM changes.

AbstractBackgroundNeurite orientation dispersion and density imaging (NODDI) measures from diffusion MRI provide information about tissue microstructure. Our goal was to compare the sensitivity of NODDI to structural MRI (sMRI) in capturing neurodegenerative changes in the gray matter due to aging and dementia.MethodWe included 815 participants [644 cognitively unimpaired (CU), 103 MCI, 68 AD] from the Mayo Clinic Study of Aging (MCSA) and Mayo ADRC with sMRI and multi‐shell diffusion MRI. We assessed gray matter volumes scaled by head size and cortical thicknesses from sMRI and neurite density index (NDI) and isotropic volume fraction (ISOVF) from NODDI. Pairwise Spearman correlations were calculated between imaging measures and Tau‐PET SUVRs in Braak I‐II, III‐IV, and V‐VI regions of interest (ROI) in participants ≥ 60 years. Separate multivariable linear regression models were fit for the prediction of brain age using sMRI and NODDI in 515 CU participants. The predicted age gaps were calculated as chronological age minus predicted age in the remaining participants. We compared age gaps computed by sMRI and NODDI using disease surrogates (APOE4 genotype, AD pathologies via PET, WMH via FLAIR‐MRI, vascular risk, cognition via MMSE).ResultBraak ROIs, sMRI, and NODDI measures were correlated. The correlations between ISOVF and Tau‐PET SUVRs were stronger than correlations between NDI, volume, and thickness and Tau‐PET SUVR (Fig.1). MMSE was more associated with ISOVF than Tau‐PET and other imaging measures. The sMRI regression models had a higher mean absolute error (MAE) of 4.13 years than NODDI models with an MAE of 4.02 years. The NODDI model better predicted age gap separation between CU, MCI, and AD compared to the sMRI model (Fig.2). The NODDI predicted age gap showed stronger correlations with global amyloid‐PET SUVR, meta‐ROI Tau‐PET SUVR, CMC, APOE4, and MMSE (Fig. 3).ConclusionNODDI, especially ISOVF, had stronger correlations with MMSE and Tau‐PET SUVRs across all Braak ROIs. The brain age prediction model trained using NODDI was more sensitive in detecting neurodegeneration changes than sMRI based on associations with disease surrogates. Our work suggests that NODDI in gray matter may be a more sensitive surrogate of neurodegeneration compared to sMRI.

AbstractBackgroundNeurite orientation dispersion and density imaging (NODDI) measures from diffusion MRI provide information about tissue microstructure. Our goal was to compare the sensitivity of NODDI to structural MRI (sMRI) in capturing neurodegenerative changes in the gray matter due to aging and dementia.MethodWe included 815 participants [644 cognitively unimpaired (CU), 103 MCI, 68 AD] from the Mayo Clinic Study of Aging (MCSA) and Mayo ADRC with sMRI and multi‐shell diffusion MRI. We assessed gray matter volumes scaled by head size and cortical thicknesses from sMRI and neurite density index (NDI) and isotropic volume fraction (ISOVF) from NODDI. Pairwise Spearman correlations were calculated between imaging measures and Tau‐PET SUVRs in Braak I‐II, III‐IV, and V‐VI regions of interest (ROI) in participants = 60 years. Separate multivariable linear regression models were fit for the prediction of brain age using sMRI and NODDI in 515 CU participants. The predicted age gaps were calculated as chronological age minus predicted age in the remaining participants. We compared age gaps computed by sMRI and NODDI using disease surrogates (APOE4 genotype, AD pathologies via PET, WMH via FLAIR‐MRI, vascular risk, cognition via MMSE).ResultBraak ROIs, sMRI, and NODDI measures were correlated. The correlations between ISOVF and Tau‐PET SUVRs were stronger than correlations between NDI, volume, and thickness and Tau‐PET SUVR (Fig.1). MMSE was more associated with ISOVF than Tau‐PET and other imaging measures. The sMRI regression models had a higher mean absolute error (MAE) of 4.13 years than NODDI models with an MAE of 4.02 years. The NODDI model better predicted age gap separation between CU, MCI, and AD compared to the sMRI model (Fig.2). The NODDI predicted age gap showed stronger correlations with global amyloid‐PET SUVR, meta‐ROI Tau‐PET SUVR, CMC, APOE4, and MMSE (Fig. 3).ConclusionNODDI, especially ISOVF, had stronger correlations with MMSE and Tau‐PET SUVRs across all Braak ROIs. The brain age prediction model trained using NODDI was more sensitive in detecting neurodegeneration changes than sMRI based on associations with disease surrogates. Our work suggests that NODDI in gray matter may be a more sensitive surrogate of neurodegeneration compared to sMRI.

AbstractINTRODUCTIONCost‐effective screening tools for vascular contributions to cognitive impairment and dementia (VCID) has significant implications. We evaluated non‐imaging indicators of VCID using magnetic resonance imaging (MRI)‐measured white matter (WM) damage and hypothesized that these indicators differ based on age.METHODSIn 745 participants from the Mayo Clinic Study of Aging (≥50 years of age) with serial WM assessments from diffusion MRI and fluid‐attenuated inversion recovery (FLAIR)‐MRI, we examined associations between baseline non‐imaging indicators (demographics, vascular risk factors [VRFs], gait, behavioral, plasma glial fibrillary acidic protein [GFAP], and plasma neurofilament light chain [NfL]) and WM damage across three age tertiles.RESULTSVRFs and gait were associated with diffusion changes even in low age strata. All measures (VRFs, gait, behavioral, plasma GFAP, plasma NfL) were associated with white matter hyperintensities (WMHs) but mainly in intermediate and high age strata.DISCUSSIONNon‐imaging indicators of VCID were related to WM damage and may aid in screening participants and assessing outcomes for VCID.Highlights Non‐imaging indicators of VCID can aid in prediction of MRI‐measured WM damage but their importance differed by age. Vascular risk and gait measures were associated with early VCID changes measured using diffusion MRI. Plasma markers explained variability in WMH across age strata. Most non‐imaging measures explained variability in WMH and vascular WM scores in intermediate and older age groups. The framework developed here can be used to evaluate new non‐imaging VCID indicators proposed in the future.

Abstract Alzheimer's disease (AD) exhibits spatially heterogeneous 3R/4R tau pathology distributions across participants, making it a challenge to quantify extent of tau deposition. Utilizing Tau-PET from three independent cohorts, we trained and validated a machine learning model to identify visually positive Tau-PET scans from regional SUVR values and developed a novel summary measure, THETA, that accounts for heterogeneity in tau deposition. The model for identification of tau positivity achieved a balanced test accuracy of 95% and accuracy of ≥87% on the validation datasets. THETA captured heterogeneity of tau deposition, had better association with clinical measures, and corresponded better with visual assessments in comparison with the temporal meta-region-of-interest Tau-PET quantification methods. Our novel approach aids in identification of positive Tau-PET scans and provides a quantitative summary measure, THETA, that effectively captures the heterogeneous tau deposition seen in AD. The application of THETA for quantifying Tau-PET in AD exhibits great potential.

Abstract Blood-based biomarkers offer strong potential to revolutionize diagnosis, trial enrolment and treatment monitoring in Alzheimer’s disease (AD). However, further advances are needed before these biomarkers can achieve wider deployment beyond selective research studies and specialty memory clinics, including the development of frameworks for optimal interpretation of biomarker profiles. We hypothesized that integrating Alzheimer’s disease genetic risk score (AD-GRS) data would enhance the diagnostic value of plasma AD biomarkers by better capturing extant disease heterogeneity. Analysing 962 individuals from a population-based sample, we observed that an AD-GRS was independently associated with amyloid PET levels (an early marker of AD pathophysiology) over and above APOE ε4 or plasma p-tau181, amyloid-β42/40, glial fibrillary acidic protein or neurofilament light chain. Among individuals with a high or moderately high plasma p-tau181, integrating AD-GRS data significantly improved classification accuracy of amyloid PET positivity, including the finding that the combination of a high AD-GRS and high plasma p-tau181 outperformed p-tau181 alone in classifying amyloid PET positivity (88% versus 68%; P = 0.001). A machine learning approach incorporating plasma biomarkers, demographics and the AD-GRS was highly accurate in predicting amyloid PET levels (90% training set; 89% test set) and Shapley value analyses (an explainer method based in cooperative game theory) indicated that the AD-GRS and plasma biomarkers had differential importance in explaining amyloid deposition across individuals. Polygenic risk for AD dementia appears to account for a unique portion of disease heterogeneity, which could non-invasively enhance the interpretation of blood-based AD biomarker profiles in the population.