PREDICT-FTD: Multimodal Imaging Prediction of FTLD Subtypes.

PREDICT-FTD：FTLD 亚型的多模态成像预测。

• 批准号: 10915129
• 负责人: HOWARD J ROSEN
• 金额: $ 140.83万
• 依托单位: OHIO STATE UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-09-30 至 2024-09-29
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10915129
• 关键词: Affect; Age; Age of Onset; Alzheimer' s Disease; Anatomy; Atrophic; Brain; C9ORF72; Cerebrospinal Fluid; Clinical; Clinical Trials; DNA Sequence Alteration; Data; Data Set; Dementia; Deterioration; Development; Diffuse Pattern; Diffusion; Digit structure; Disease; Early Intervention; Frontotemporal Lobar Degenerations; Grant; Image; Individual; International; Intervention; Light; Machine Learning; Measures; Methods; Modeling; Monitor; Multimodal Imaging; Mutation; Nerve Degeneration; Neuroanatomy; Neurodegenerative Disorders; Neuronal Injury; Patients; Pattern; Persons; Primary Progressive Aphasia; Process; Progressive Supranuclear Palsy; Risk; Sample Size; Sampling Studies; Semantics; Statistical Models; Surface; Symptoms; Syndrome; Teenagers; Testing; Time; Variant; accurate diagnosis; behavioral variant frontotemporal dementia; clinical heterogeneity; clinical predictors; corticobasal syndrome; effective therapy; gray matter; high risk population; improved; machine learning method; machine learning model; mild cognitive impairment; multimodal neuroimaging; multimodality; mutation carrier; neurofilament; neuroimaging; neuron loss; personalized predictions; predictive modeling; prevent; structural imaging; tool; white matter change

Project Summary Frontotemporal lobar degeneration (FTLD) is a devastating neurodegenerative disorder, and a common cause of dementia in people under the age of 65. Recent advances are offering hope for disease modifying interventions, but such treatments will only be effective if patients are accurately diagnosed, and likely to be most effective early in the course of disease. The ability to accurately predict the type of clinical syndrome is critical for selecting appropriate early interventions that may be targeted for specific symptoms at specific stages. The ability to accurately predict the age at which the symptoms will emerge is critical for clinical trials, which requires accurate measures that can indicate that the individual at risk is likely to develop symptoms within a specific timeframe. Unfortunately, the type of clinical syndrome and age of onset can vary dramatically even within the same mutation. An individual with a given mutation can eventually develop any of the FTLD syndrome subtypes, and the age of onset of clinical syndromes can also vary within the same mutation. This striking clinical heterogeneity in FTLD due to genetic mutation severely limits our ability to predict when and what specific symptoms will emerge. Neuroimaging studies of symptomatic and presymptomatic FTLD mutation carriers have shown detectable gray matter and white matter changes across all mutations. However, the variability in the clinical syndrome trajectories that exist in each mutation has led to mixed findings. A small number of studies of presymptomatic mutation carriers who went on to develop clinical symptoms suggest that presymptomatic structural changes can be used to predict time to dementia. However, their sample sizes were small, often are numbered in the single digits or low teens, thus leaving many unanswered questions, including whether prediction differs by mutation and whether specific syndromes can be predicted, and whether the choice of neuroimaging measures may depend on the stage of the presymptomatic progression. Our proposed study will use large-scale longitudinal multimodal neuroimaging datasets of presymptomatic FTLD mutation carriers to predict the type of syndrome that they will develop and the time of onset. 1) We will focus on the presymptomatic trajectory of eventual clinical syndromes and develop new models that incorporate longitudinal data. 2) We will leverage the largest international consortia studies that follow presymptomatic mutation carriers. 3) We will consider more sensitive imaging measures. 4) Finally, we will employ powerful machine- learning-based methods appropriate for our proposed sample size. If successful, the methods and measures we develop can inform clinical trials that require accurate predictors of timeframe before conversion. If and when effective treatments become available, reliable predictors of when and which syndrome an individual will develop can be used for selecting appropriate early interventions that may be targeted for specific symptoms at specific stages, as well as specific measures that should be monitored in each at-risk individual.

项目总结