Neurodegenerative disease management · 2026
Plasma biomarker and machine learning modeling of clinical progression in cognitively normal individuals at risk for Alzheimer's disease.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Thabtah Fadi, Kamalov Firuz, Spencer Robbie, Abdelhamid Neda, Nami Mohammad, Alzheimer’s Disease Neuroimaging Initiative
Аннотация
AIMS: To evaluate whether plasma biomarkers can distinguish cognitively normal (CN) individuals who remain clinically stable from those who subsequently develop mild cognitive impairment (MCI) or Alzheimer's disease (AD) and to test the predictive performance of classification data‑driven models.
MATERIALS AND METHODS: Plasma biomarker follow-up data from initially CN participants in the Alzheimer's Disease Neuroimaging Initiative (ADNI) were analyzed across Elecsys, Lumipulse, Precivity, and SIMOA assays. Feature selection was done using chi-square, information gain ratio, and ReliefF. K-nearest neighbors (kNN), logistic regression, support vector machine (SVM), and Extreme Gradient Boosting (XGBoost) classifiers were tested using cross-validation. Synthetic Minority Over-sampling Technique (SMOTE) was applied to alleviate class imbalance.
RESULTS: pTau217, pTau181, Aβ40, NfL, and assay-specific amyloid/tau ratios have relatively high feature importance. Class imbalance significantly reduced sensitivity in unbalanced models. Following data balancing, the best‑performing SVM model with selected Elecsys features achieved an accuracy of 71.6%, while the highest sensitivity of all models was 56%.
CONCLUSIONS: Baseline plasma biomarker data, combined with follow-up plasma biomarkers visits can help us identify patterns that are associated with subsequent clinical progression from CN to MCI/AD. However, predictive performance was moderate, and external validation in larger and more diverse cohorts is needed before clinical applications.
MATERIALS AND METHODS: Plasma biomarker follow-up data from initially CN participants in the Alzheimer's Disease Neuroimaging Initiative (ADNI) were analyzed across Elecsys, Lumipulse, Precivity, and SIMOA assays. Feature selection was done using chi-square, information gain ratio, and ReliefF. K-nearest neighbors (kNN), logistic regression, support vector machine (SVM), and Extreme Gradient Boosting (XGBoost) classifiers were tested using cross-validation. Synthetic Minority Over-sampling Technique (SMOTE) was applied to alleviate class imbalance.
RESULTS: pTau217, pTau181, Aβ40, NfL, and assay-specific amyloid/tau ratios have relatively high feature importance. Class imbalance significantly reduced sensitivity in unbalanced models. Following data balancing, the best‑performing SVM model with selected Elecsys features achieved an accuracy of 71.6%, while the highest sensitivity of all models was 56%.
CONCLUSIONS: Baseline plasma biomarker data, combined with follow-up plasma biomarkers visits can help us identify patterns that are associated with subsequent clinical progression from CN to MCI/AD. However, predictive performance was moderate, and external validation in larger and more diverse cohorts is needed before clinical applications.