A Novel Informatics System For Craniosynostosis Surgery

摘要 阿尔茨海默病（AD）的特征在于进行性记忆丧失和认知能力下降，脑损害， 老年斑中淀粉样β肽（Aβ）的积累和神经元缠结中过度磷酸化的tau蛋白 （NFT）。由于AD是一种复杂且多因素的疾病，因此具有多种数据类型的大型数据集已被广泛应用于临床。 关键是要确定其风险因素。几十年来，只有载脂蛋白E（APOE）的等位基因4， 在大约一半的迟发性AD（LOAD）患者中存在，已令人信服地证明会影响 即可.然而，不幸的是，目前的治疗方法只是治标不治本，因为它们不能减缓或停止 疾病进展。迫切需要对生物标志物进行更多的研究。 本研究中使用的数据来自阿尔茨海默病神经影像学倡议（ADNI）数据库。 目前，ADNI联盟开放了2,000多名AD患者的MRI成像数据， 和严重的阶段。我们计划应用为颅缝早闭症研究开发的人工智能和机器学习方法 在父R 01 DE 027027中，将ADNI数据与AD成像数据进行比较，并尝试分割和重建AD成像数据， 描述AD患者脑的生物力学特性，然后进一步对AD患者进行分层， 更好的治疗这种想法从未应用于AD研究，这可能是一个潜在的贡献， AD研究。 AD疾病的分期对于设计治疗策略是非常重要的。有许多研究工作 从ADNI数据集成像遗传学和基于生物标志物的分期技术，但这些都没有工作 研究AD发生过程中生物力学特性的变化。许多人都观察到， 研究人员和医生认为，AD组织往往不太僵硬，弹性也不太好。因此，迫切需要 以提高我们对AD脑组织特性与AD分期相关性的理解。我们的近期目标是 开发计算模型以表征AD患者特定的组织弹性和AD阶段。实现 这些目标，我们的具体目标是：（1）开发深度学习框架，以获得大脑体积， AD患者表面;（2）开发直接估计子区域组织硬度的计算技术 从AD成像数据;并预测AD进展的基础上AD脑的生物力学特征。 本NIA供应品的范围在母产品R 01 DE 027027“eSuture system：A”的范围内。 新的颅缝早闭（CSO）手术信息系统。eSuture系统的重点是 开发新的成像信息学和机器学习技术来分割CSO成像数据， 对CSO患者进行分层和分类，并对颅骨组织的生物力学特性进行表征， 非线性有限元模型

Abstract Alzheimer's disease (AD) is characterized by progressive memory loss and cognitive decline, cerebral accumulation of amyloid-β peptide (Aβ) in senile plaques and hyper-phosphorylated tau in neurofibrillary tangles (NFT). Since AD is a complex and multifactorial disease, large datasets with multiple data types have been critical to identify its risk factors. For several decades, only the allele 4 of Apolipoprotein E (APOE), which is present in about half of late-onset AD (LOAD) patients, has been convincingly demonstrated to affect risk for LOAD. However, unfortunately, current treatments are just palliative because they do not slow down or halt the disease progression. More research on biomarkers are urgently needed. Data used in this study were obtained fromthe Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Currently ADNI consortium opened MRI imaging data for over 2,000 AD patients from normal, mild, moderate and severe stages. We plan to apply the AI and machine learning methods developed for craniosynostosis study in the parent R01DE027027 to the ADNI data and try to segment and reconstruct the AD imaging data, characterize the biomechanical property of brain in AD patients, and then further stratify the AD patients for better therapy. This kind of idea was never applied to AD research, which could be a potential contribution to the AD study. Staging the AD disease is very important for design therapy strategy. There are numerous work studied imaging genetics from the ADNI data sets and biomarker based staging technologies, but none of those work studied the biomechanical property changes during the AD development. It has been observed by many researchers and physicians that AD tissues tend to be less stiff and less elastic. Hence, there is an urgent need to improve our understanding of the AD brain tissue property correlated to AD stages. Our immediate goal is to develop computational model to characterize the AD patient specific tissue elasticity and AD stages. To achieve these goals, our Specific Aims are: (1) to develop deep learning framework to obtain the brain volume and surface of AD patients; (2) to develop computational techniques for estimating sub-region tissue stiffness directly from AD imaging data; and to predict AD progression based on the biomechanical features of AD brain. The scope of this NIA suppl. is within the scope of the parent R01DE027027 “eSuture system: A novel informatics system for craniosynostosis (CSO) surgery.” The eSuture system focuses on developing novel imaging informatics and machine leaning technologies to segment CSO imagining data, to stratify and classify CSO patients, and to characterize the biomechanical property of calvarial bone tissue with nonlinear finite element models.