Predictive analytics for cognitive decline and Alzheimer’s disease

认知能力下降和阿尔茨海默病的预测分析

• 批准号: 10626743
• 负责人: Ali Ezzati
• 金额: $ 3.01万
• 依托单位: ALBERT EINSTEIN COLLEGE OF MEDICINE
• 依托单位国家: 美国
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-08-01 至 2023-06-01
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10626743
• 关键词: Affect; Aging; Alzheimer disease detection; Alzheimer' s Disease; Alzheimer' s disease therapy; American; Artificial Intelligence; Big Data; Biological Markers; Biometry; Classification; Clinical; Clinical Research; Clinical Trials; Cognition; Cognitive; Cognitive aging; Complex; Data; Data Collection; Data Set; Data Sources; Decision Making; Dementia; Development; Dimensions; Disease; Disease Progression; Early Diagnosis; Elderly; Enrollment; Ensure; Foundations; Future; Genetic; Genetic Risk; Goals; Healthcare; Image; Impaired cognition; Individual; International; Investigation; K-Series Research Career Programs; Life Style; Longitudinal cohort; Machine Learning; Magnetic Resonance Imaging; Measurement; Measures; Mentors; Mentorship; Methodology; Methods; Modeling; Neurobiology; Neurodegenerative Disorders; Neurologist; Neurology; Neuropsychology; Outcome; Participant; Patients; Performance; Pharmaceutical Preparations; Population; Predictive Analytics; Predictive Value; Prevention trial; Primary Prevention; Research; Research Personnel; Sample Size; Sampling; Secondary Prevention; Statistical Methods; Techniques; Therapeutic Effect; United States; Universities; Washington; Work; aged; amyloid imaging; arm; biological heterogeneity; care burden; career; career development; clinical decision-making; clinical heterogeneity; clinical practice; cognitive change; cohort; comparative; computational neuroscience; cost; data harmonization; data quality; data reduction; demographics; diagnostic accuracy; effective therapy; feature selection; follow-up; functional decline; functional outcomes; global health; high risk; improved; innovation; machine learning framework; machine learning method; machine learning model; machine learning prediction; mild cognitive impairment; multidimensional data; neuroimaging; novel; participant enrollment; patient oriented; pre-clinical; predict clinical outcome; predictive modeling; predictive tools; prevent; prodromal Alzheimer' s disease; prognostic; research study; risk prediction; structured data; success; trial enrollment

Project Summary/Abstract Alzheimer’s disease (AD), the most common cause of dementia in the elderly, is a major global healthcare burden. However, there is still no effective disease modifying therapy for AD and clinical trials with the aim of preventing or stabilizing cognitive impairment have largely failed. Decision making in both clinical practice and research is highly dependent on practical predictive tools, which can effectively predict cognitive or functional outcomes in individuals. Such models could be potentially used in clinical research to boost the power of trials by enrollment of participants who are most likely to show disease progression during the trial’s timeframe. Alternatively, these models could be used to identifying individuals who would benefit from primary or secondary prevention once there are effective treatments for AD. In this project, we aim to provide a framework for practical prediction of cognitive decline with aging and prodromal AD, by applying a novel ML framework to multiple dimensions of data (demographics, genetic risk scores, neuropsychological measures, structural MRI, and amyloid imaging). Our ultimate goal is to arrive at a new “Machine Learning predictive framework for aging and AD” (ML4AD), comprised of dimensions each of which each will add incremental value to the predictive models, hence increasing the performance of predictive models while keeping the costs and burden of research at a minimum. The candidate for this Mentored Patient-Oriented Career Development Award (K23), Dr. Ali Ezzati, is a Neurologist whose career goal is to develop predictive tools to help research and clinical decision making in cognitive aging and dementia. The proposed research will leverage the rich clinical and biomarker dataset available from several ongoing international studies, but will also provide a unique avenue of investigation for the candidate. The candidate's career development will benefit from close mentorship and scientific guidance of outstanding investigators in aging/AD neurobiology (Dr. Lipton), machine learning and computational neuroscience (Dr. Davatzikos), and biostatistics (Dr. Hall). The findings from this study will inform future secondary prevention trials, in which sensitive indicators of early AD will be necessary to identify high-risk subjects and track early clinical decline. This work will serve as the foundation to move forward in independent research focusing on development of predictive tools in AD and related neurodegenerative disorders. Key words: Alzheimer’s Disease, Dementia, Mild Cognitive Impairment, Cognitive neurology, Artificial Intelligence, Machine Learning, Predictive Analytics, Longitudinal Cohort, Big Data

项目概要/摘要 阿尔茨海默病（AD）是导致老年人痴呆的最常见原因，是全球主要的疾病 然而，目前还没有有效的 AD 和疾病缓解疗法。 旨在预防或稳定认知障碍的临床试验基本上失败了。 临床实践和研究中的决策高度依赖于实际预测 工具，可以有效地预测此类模型的认知或功能结果。 可能用于临床研究，通过招募受试者来提高试验的效力 在试验期间最有可能出现疾病进展的参与者。 或者，这些模型可用于识别将从中受益的个人 一旦找到有效的 AD 治疗方法，我们就进行一级或二级预防。 旨在为衰老和前驱期认知衰退的实际预测提供框架 AD，通过将新颖的机器学习框架应用于多个维度的数据（人口统计、遗传风险） 我们的最终目标是（评分、神经心理学测量、结构 MRI 和淀粉样蛋白成像）。 得出一个新的“衰老和AD的机器学习预测框架”（ML4AD），包括 每个维度都会为预测模型增加增量价值，因此 提高预测模型的性能，同时保持研究成本和负担 至少是该“以患者为导向的职业发展奖”的候选人。 (K23)，Ali Ezzati 博士，是一位神经科医生，其职业目标是开发预测工具来帮助 认知衰老和痴呆症的研究和临床决策。 将利用来自多个正在进行的国际项目的丰富的临床和生物标志物数据集 研究，但也将为候选人提供独特的调查途径。 职业发展将受益于优秀人才的密切指导和科学指导 衰老/AD 神经生物学研究人员（Lipton 博士）、机器学习和计算 这项研究的结果将为神经科学（Davatzikos 博士）和生物统计学（Hall 博士）提供信息。 未来的二级预防试验中，需要早期 AD 的敏感指标 识别高风险受试者并跟踪早期临床衰退这项工作将作为基础。 推进独立研究，重点关注 AD 和预测工具的开发 相关的神经退行性疾病。 关键词: 阿尔茨海默病, 痴呆, 轻度认知障碍, 认知神经病学, 人工智能、机器学习、预测分析、纵向队列、大数据