Metabolic Factors Are Related to Brain Amyloid Among Mexican Americans: A HABS-HD Study

The association of glucose metabolism measures and diabetes status with Alzheimer's disease biomarkers of amyloid and tau: A systematic review and meta-analysis

Conflicting evidence exists on the relationship between diabetes mellitus (DM) and Alzheimer's disease (AD) biomarkers. Therefore, we conducted a random-effects meta-analysis to evaluate the correlation of glucose metabolism measures (glycated hemoglobin, fasting blood glucose, insulin resistance indices) and DM status with AD biomarkers of amyloid-β and tau measured by positron emission tomography or cerebrospinal fluid. We selected 37 studies from PubMed and Embase, including 11,694 individuals. More impaired glucose metabolism and DM status were associated with higher tau biomarkers (r=0.11[0.03-0.18], p=0.008; I2=68%), but were not associated with amyloid-β biomarkers (r=-0.06[-0.13-0.01], p=0.08; I2=81%). Meta-regression revealed that glucose metabolism and DM were specifically associated with tau biomarkers in population settings (p=0.001). Furthermore, more impaired glucose metabolism and DM status were associated with lower amyloid-β biomarkers in memory clinic settings (p=0.004), and in studies with a higher prevalence of dementia (p<0.001) or lower cognitive scores (p=0.04). These findings indicate that DM is associated with biomarkers of tau but not with amyloid-β. This knowledge is valuable for improving dementia and DM diagnostics and treatment.

Plasma Biomarkers of Alzheimer's Disease Are Associated with Physical Functioning Outcomes Among Cognitively Normal Adults in the Multiethnic HABS-HD Cohort

In this study, we examined the link between plasma Alzheimer's disease (AD) biomarkers and physical functioning outcomes within a community-dwelling, multiethnic cohort. Data from 1 328 cognitively unimpaired participants (n = 659 Mexican American and n = 669 non-Hispanic White) from the ongoing Health & Aging Brain Study-Health Disparities (HABS-HD) cohort were examined. Plasma AD biomarkers (amyloid beta [Aβ]40, Aβ42, total tau [t-tau], and neurofilament light chain [NfL]) were assayed using the ultra-sensitive Simoa platform. Physical functioning measures were the Timed Up and Go (TUG) and the Short Physical Performance Battery (SPPB). Cross-sectional linear regression analyses revealed that plasma Aβ 40 (p < .001), Aβ 42 (p = .003), and NfL (p < .001) were each significantly associated with TUG time in seconds. Plasma Aβ 40 (p < .001), Aβ 42 (p < .001), t-tau (p = .002), and NfL (p < .001) were each significantly associated with SPPB Total Score. Additional analyses demonstrate that the link between plasma AD biomarkers and physical functioning outcomes were strongest among Mexican Americans. Plasma AD biomarkers are receiving a great deal of attention in the literature and are now available clinically including use in clinical trials. The examination of AD biomarkers and physical functioning may allow for the development of risk profiles, which could stratify a person's risk for neurodegenerative diseases, such as AD, based on plasma AD biomarkers, physical functioning, ethnicity, or a combination of these measures prior to the onset of cognitive impairment.

Comorbidities Incorporated to Improve Prediction for Prevalent Mild Cognitive Impairment and Alzheimer's Disease in the HABS-HD Study

BACKGROUND

Blood biomarkers have the potential to transform Alzheimer's disease (AD) diagnosis and monitoring, yet their integration with common medical comorbidities remains insufficiently explored.

OBJECTIVE

This study aims to enhance blood biomarkers' sensitivity, specificity, and predictive performance by incorporating comorbidities. We assess this integration's efficacy in diagnostic classification using machine learning, hypothesizing that it can identify a confident set of predictive features.

METHODS

We analyzed data from 1,705 participants in the Health and Aging Brain Study-Health Disparities, including 116 AD patients, 261 with mild cognitive impairment, and 1,328 cognitively normal controls. Blood samples were assayed using electrochemiluminescence and single molecule array technology, alongside comorbidity data gathered through clinical interviews and medical records. We visually explored blood biomarker and comorbidity characteristics, developed a Feature Importance and SVM-based Leave-One-Out Recursive Feature Elimination (FI-SVM-RFE-LOO) method to optimize feature selection, and compared four models: Biomarker Only, Comorbidity Only, Biomarker and Comorbidity, and Feature-Selected Biomarker and Comorbidity.

RESULTS

The combination model incorporating 17 blood biomarkers and 12 comorbidity variables outperformed single-modal models, with NPV12 at 92.78%, AUC at 67.59%, and Sensitivity at 65.70%. Feature selection led to 22 chosen features, resulting in the highest performance, with NPV12 at 93.76%, AUC at 69.22%, and Sensitivity at 70.69%. Additionally, interpretative machine learning highlighted factors contributing to improved prediction performance.

CONCLUSIONS

In conclusion, combining feature-selected biomarkers and comorbidities enhances prediction performance, while feature selection optimizes their integration. These findings hold promise for understanding AD pathophysiology and advancing preventive treatments.

Combination of Serum and Plasma Biomarkers Could Improve Prediction Performance for Alzheimer’s Disease

Alzheimer’s disease (AD) can be predicted either by serum or plasma biomarkers, and a combination may increase predictive power, but due to the high complexity of machine learning, it may also incur overfitting problems. In this paper, we investigated whether combining serum and plasma biomarkers with feature selection could improve prediction performance for AD. 150 D patients and 150 normal controls (NCs) were enrolled for a serum test, and 100 patients and 100 NCs were enrolled for the plasma test. Among these, 79 ADs and 65 NCs had serum and plasma samples in common. A 10 times repeated 5-fold cross-validation model and a feature selection method were used to overcome the overfitting problem when serum and plasma biomarkers were combined. First, we tested to see if simply adding serum and plasma biomarkers improved prediction performance but also caused overfitting. Then we employed a feature selection algorithm we developed to overcome the overfitting problem. Lastly, we tested the prediction performance in a 10 times repeated 5-fold cross validation model for training and testing sets. We found that the combined biomarkers improved AD prediction but also caused overfitting. A further feature selection based on the combination of serum and plasma biomarkers solved the problem and produced an even higher prediction performance than either serum or plasma biomarkers on their own. The combined feature-selected serum–plasma biomarkers may have critical implications for understanding the pathophysiology of AD and for developing preventative treatments.

Hyperparameter Tuning with High Performance Computing Machine Learning for Imbalanced Alzheimer’s Disease Data

Accurate detection is still a challenge in machine learning (ML) for Alzheimer’s disease (AD). Class imbalance in imbalanced AD data is another big challenge for machine-learning algorithms working under the assumption that the data are evenly distributed within classes. Here, we present a hyperparameter tuning workflow with high-performance computing (HPC) for imbalanced data related to prevalent mild cognitive impairment (MCI) and AD in the Health and Aging Brain Study-Health Disparities (HABS-HD) project. We applied a single-node multicore parallel mode to hyperparameter tuning of gamma, cost, and class weight using a support vector machine (SVM) model with 10 times repeated fivefold cross-validation. We executed the hyperparameter tuning workflow with R’s bigmemory, foreach, and doParallel packages on Texas Advanced Computing Center (TACC)’s Lonestar6 system. The computational time was dramatically reduced by up to 98.2% for the high-performance SVM hyperparameter tuning model, and the performance of cross-validation was also improved (the positive predictive value and the negative predictive value at base rate 12% were, respectively, 16.42% and 92.72%). Our results show that a single-node multicore parallel structure and high-performance SVM hyperparameter tuning model can deliver efficient and fast computation and achieve outstanding agility, simplicity, and productivity for imbalanced data in AD applications.