PREDICT-FTD: Multimodal Imaging Prediction of FTLD Subtypes.

项目摘要 额颞叶变性(FTLD)是一种破坏性的神经退行性疾病，也是一种常见的 65岁以下人群患痴呆症的原因。最近的进展为疾病的改良带来了希望 干预，但这种治疗只有在患者得到准确诊断的情况下才有效，而且很可能是 在病程早期最有效。准确预测临床症状类型的能力是 对于选择适当的早期干预措施至关重要，这些干预措施可能针对特定症状 各阶段。准确预测出现症状的年龄是临床试验的关键， 这需要准确的测量方法，以表明处于危险中的个人可能会出现症状 在特定的时间范围内。不幸的是，临床症状的类型和发病年龄可能会有很大的不同。 即使在同一变异中也是如此。带有给定突变的个体最终可能会患上任何FTLD 在相同的突变中，症状亚型和临床症状的发病年龄也可能不同。这 由于基因突变导致的FTLD显著的临床异质性严重限制了我们预测何时和 会出现哪些具体症状。有症状和无症状FTLD的神经影像研究 突变携带者在所有突变中都显示出可检测到的灰质和白质变化。然而， 每个突变中存在的临床综合征轨迹的差异性导致了混合的结果。一小块 对后来出现临床症状的症状前突变携带者的多项研究表明 症状前的结构变化可以用来预测痴呆的时间。然而，他们的样本量是 小的，通常是个位数或十几岁的数字，因此留下了许多悬而未决的问题，包括 预测是否因突变而不同，是否可以预测特定的症状，以及 神经影像方法的选择可能取决于症状前进展的阶段。我们的建议 研究将使用症状前FTLD突变的大规模纵向多模式神经成像数据集 携带者可以预测他们将出现的综合征类型和发病时间。1)我们将重点关注 最终临床症状的症状前轨迹，并开发结合纵向的新模型 数据。2)我们将利用最大的国际财团对症状前突变的研究 承运人。3)我们将考虑更敏感的成像措施。4)最后，我们将使用强大的机器- 基于学习的方法适合我们建议的样本量。如果成功，有哪些方法和措施 我们开发的可以为需要在转换之前准确预测时间框架的临床试验提供信息。如果和 当有效的治疗变得可用时，可靠的预测个体将在什么时候以及哪种综合症 开发可用于选择适当的早期干预措施，这些干预措施可能针对以下特定症状 具体阶段，以及应对每个高危个人进行监测的具体措施。

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.