Dataset Information

Machine Learning Approach for Improved Longitudinal Prediction of Progression from Mild Cognitive Impairment to Alzheimer's Disease.

ABSTRACT: Mild cognitive impairment (MCI) is cognitive decline that can indicate future risk of Alzheimer's disease (AD). We developed and validated a machine learning algorithm (MLA), based on a gradient-boosted tree ensemble method, to analyze phenotypic data for individuals 55-88 years old (n = 493) diagnosed with MCI. Data were analyzed within multiple prediction windows and averaged to predict progression to AD within 24-48 months. The MLA outperformed the mini-mental state examination (MMSE) and three comparison models at all prediction windows on most metrics. Exceptions include sensitivity at 18 months (MLA and MMSE each achieved 0.600); and sensitivity at 30 and 42 months (MMSE marginally better). For all prediction windows, the MLA achieved AUROC ≥ 0.857 and NPV ≥ 0.800. With averaged data for the 24-48-month lookahead timeframe, the MLA outperformed MMSE on all metrics. This study demonstrates that machine learning may provide a more accurate risk assessment than the standard of care. This may facilitate care coordination, decrease healthcare expenditures, and maintain quality of life for patients at risk of progressing from MCI to AD.

SUBMITTER: Adelson RP

PROVIDER: S-EPMC10795823 | biostudies-literature | 2023 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine Learning Approach for Improved Longitudinal Prediction of Progression from Mild Cognitive Impairment to Alzheimer's Disease.

Adelson Robert P RP Garikipati Anurag A Maharjan Jenish J Ciobanu Madalina M Barnes Gina G Singh Navan Preet NP Dinenno Frank A FA Mao Qingqing Q Das Ritankar R

Diagnostics (Basel, Switzerland) 20231220 1

Mild cognitive impairment (MCI) is cognitive decline that can indicate future risk of Alzheimer's disease (AD). We developed and validated a machine learning algorithm (MLA), based on a gradient-boosted tree ensemble method, to analyze phenotypic data for individuals 55-88 years old (<i>n</i> = 493) diagnosed with MCI. Data were analyzed within multiple prediction windows and averaged to predict progression to AD within 24-48 months. The MLA outperformed the mini-mental state examination (MMSE) ...[more]

PMID: 38201322

Similar Datasets

Project description:ObjectiveThis study aimed to build and validate a radiomics-integrated model with whole-brain magnetic resonance imaging (MRI) to predict the progression of mild cognitive impairment (MCI) to Alzheimer's disease (AD).Methods357 patients with MCI were selected from the ADNI database, which is an open-source database for AD with multicentre cooperation, of which 154 progressed to AD during the 48-month follow-up period. Subjects were divided into a training and test group. For each patient, the baseline T1WI MR images were automatically segmented into white matter, gray matter and cerebrospinal fluid (CSF), and radiomics features were extracted from each tissue. Based on the data from the training group, a radiomics signature was built using logistic regression after dimensionality reduction. The radiomics signatures, in combination with the apolipoprotein E4 (APOE4) and baseline neuropsychological scales, were used to build an integrated model using machine learning. The receiver operating characteristics (ROC) curve and data of the test group were used to evaluate the diagnostic accuracy and reliability of the model, respectively. In addition, the clinical prognostic efficacy of the model was evaluated based on the time of progression from MCI to AD.ResultsStepwise logistic regression analysis showed that the APOE4, clinical dementia rating, AD assessment scale, and radiomics signature were independent predictors of MCI progression to AD. The integrated model was constructed based on independent predictors using machine learning. The ROC curve showed that the accuracy of the model in the training and the test sets was 0.814 and 0.807, with a specificity of 0.671 and 0.738, and a sensitivity of 0.822 and 0.745, respectively. In addition, the model had the most significant diagnostic efficacy in predicting MCI progression to AD within 12 months, with an AUC of 0.814, sensitivity of 0.726, and specificity of 0.798.ConclusionThe integrated model based on whole-brain radiomics can accurately identify and predict the high-risk population of MCI patients who may progress to AD. Radiomics biomarkers are practical in the precursory stage of such disease.

Project description:IntroductionBrain age is a machine learning-derived estimate that captures lower brain volume. Previous studies have found that brain age is significantly higher in mild cognitive impairment and Alzheimer's disease (AD) compared to healthy controls. Few studies have investigated changes in brain age longitudinally in MCI and AD. We hypothesized that individuals with MCI and AD would show heightened brain age over time and across the lifespan. We also hypothesized that both MCI and AD would show faster rates of brain aging (higher slopes) over time compared to healthy controls.MethodsWe utilized data from an archival dataset, mainly Alzheimer's disease Neuroimaging Initiative (ADNI) 1 with 3Tesla (3 T) data which totaled 677 scans from 183 participants. This constitutes a secondary data analysis on existing data. We included control participants (healthy controls or HC), individuals with MCI, and individuals with AD. We predicted brain age using a pre-trained model and tested for accuracy. We investigated cross-sectional differences in brain age by group [healthy controls or HC, mild cognitive impairment (MCI), and AD]. We conducted longitudinal modeling of age and brain age by group using time from baseline in one model and chronological age in another model.ResultsWe predicted brain age with a mean absolute error (MAE) < 5 years. Brain age was associated with age across the study and individuals with MCI and AD had greater brain age on average. We found that the MCI group had significantly higher rates of change in brain age over time compared to the HC group regardless of individual chronologic age, while the AD group did not differ in rate of brain age change.DiscussionWe replicated past studies that showed that MCI and AD had greater brain age than HC. We additionally found that this was true over time, both groups showed higher brain age longitudinally. Contrary to our hypothesis, we found that the MCI, but not the AD group, showed faster rates of brain aging. We essentially found that while the MCI group was actively experiencing faster rates of brain aging, the AD group may have already experienced this acceleration (as they show higher brain age). Individuals with MCI may experience higher rates of brain aging than AD and controls. AD may represent a homeostatic endpoint after significant neurodegeneration. Future work may focus on individuals with MCI as one potential therapeutic option is to alter rates of brain aging, which ultimately may slow cognitive decline in the long-term.

Project description:A sensitive marker for monitoring progression of early Alzheimer's disease would help to develop and test new therapeutic strategies. The present study is aimed at investigating brain metabolism changes over time, as a potential monitoring marker, in patients with amnestic mild cognitive impairment, according to their clinical outcome (converters or non-converters), and in relation to their cognitive decline. Seventeen amnestic mild cognitive impairment patients underwent magnetic resonance imaging and 18FDG-positron emission tomography scans both at inclusion and 18 months later. Baseline and follow-up positron emission tomography data were corrected for partial volume effects and spatially normalized using magnetic resonance imaging data, scaled to the vermis and compared using SPM2. 'PET-PAC' maps reflecting metabolic per cent annual changes were created for correlation analyses with cognitive decline. In the whole sample, the greatest metabolic decrease concerned the posterior cingulate-precuneus area. Converters had significantly greater metabolic decrease than non-converters in two ventro-medial prefrontal areas, the subgenual (BA25) and anterior cingulate (BA24/32). PET-PAC in BA25 and BA24/32 combined allowed complete between-group discrimination. BA25 PET-PAC significantly correlated with both cognitive decline and PET-PAC in the hippocampal region and temporal pole, while BA24/32 PET-PAC correlated with posterior cingulate PET-PAC. Finally, the metabolic change in BA8/9/10 was inversely related to that in BA25 and showed relative increase with cognitive decline, suggesting that compensatory processes may occur in this dorso-medial prefrontal region. The observed ventro-medial prefrontal disruption is likely to reflect disconnection from the hippocampus, both indirectly through the cingulum bundle and posterior cingulate cortex for BA24/32, and directly through the uncinate fasciculus for BA25. Altogether, our findings emphasize the potential of 18FDG-positron emission tomography for monitoring early Alzheimer's disease progression.

Dataset Information

Machine Learning Approach for Improved Longitudinal Prediction of Progression from Mild Cognitive Impairment to Alzheimer's Disease.

Publications

Machine Learning Approach for Improved Longitudinal Prediction of Progression from Mild Cognitive Impairment to Alzheimer's Disease.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets