Accurate and rapid assessment of sarcopenia in older adults through electrical impedance myography

项目摘要 42%的老年人（OAs）有一个或多个身体限制，这是必不可少的。 依赖与肌肉萎缩相关的肌肉萎缩和无力（肌肉减少症）是这些物理学的重要贡献者， 生理障碍，包括最明显的步态障碍。此外，肌肉减少症与 福尔斯和其他损伤的风险增加，导致丧失独立性和死亡率增加。医疗保健 OA的提供者需要改进评估下肢肌肉状况的方法， 肌肉减少症MRI、CT和DXA可提供大量信息，但都不是基于办公室的，且仅限于DXA 提供有关瘦体重和脂肪含量的信息。在此应用中，我们将电气 阻抗肌电图（EIM）用于评估OA患者下肢肌肉状况。EIM依赖于 在将定向聚焦的多频率电流施加到特定肌肉或肌肉 组通过以这种局部化的方式应用该技术，它几乎不受水合状态的影响， 通常影响其他生物阻抗方法的其他问题。EIM已经研究了近两年 在神经肌肉疾病领域已有数十年的研究，并已被证明对多种改变敏感， 骨骼肌包括萎缩、退化和单纯的失调。使EIM更进一步， 将机器学习（ML）技术应用于复杂的多频EIM数据集， 可以预测组织学特征，包括肌纤维大小、脂肪和结缔组织含量。鉴于这种 EIM已被证明具有这种能力，它有可能成为一种方便的、基于办公室的评估方法， OAs的肌肉健康。在这一直接进入第2阶段的SBIR应用中，Myolex建议通过以下方式建立EIM 其新设备mScan与基于云的机器学习平台相结合， 获得OA肌肉中的MRI样定量数据。在具体目标1中，与老龄化专家一起 研究人员在3个不同的临床站点，我们将收集MRI和EIM数据的健康老年人队列， 沿着标准功能测量以及通过电刺激肌肉测量比力 收缩。使用这些数据，在具体目标2中，我们将开发预测算法，通过惩罚 Lasso回归技术（最小绝对收缩和选择算子），利用EIM值， 预测肌肉体积、肌肉比力和肌肉脂肪浸润。我们将把这些 将算法转化为基于云的引擎，提供有意义的，易于解释的值。结束时 通过这项拟议的工作，我们将开发出一个准确，强大的系统，临床医生和研究人员可以 用于OA肌肉健康的快速评估。

Project Summary Forty-two percent of older adults (OAs) have one or more physical limitations that are essential for in- dependence. Age-related muscle wasting and weakness (sarcopenia) are important contributors to these phy- sical impairments, including most notably gait impairment. Moreover, sarcopenia is closely associated with an increased risk of falls and other injuries leading to loss of independence and increased mortality. Healthcare providers of OAs need improved means of evaluating lower extremity muscle condition for the development of sarcopenia. MRI, CT, and DXA provide considerable information, but none are office-based and DXA only provides information on lean body mass and fat content. In this application, we advance electrical impedance myography (EIM) for assessment of lower extremity muscle condition in OAs. EIM relies upon application of directionally focused, multi-frequency electrical current to specific muscles or muscle groups. By applying the technology in such a localized fashion, it is virtually unaffected by hydration status or other issues that commonly impact other bioimpedance-based methods. EIM has been studied for nearly two decades in the field of neuromuscular disease and has been shown to be sensitive to a variety of alterations in skeletal muscle including atrophy, degeneration, and simple deconditioning. Taking EIM one step further and applying machine learning (ML) techniques to the complex multifrequency EIM data set, it even becomes possible to predict histological features, including myofiber size, fat, and connective tissue content. Given this demonstrated capability, EIM has the potential to serve as a convenient, office-based approach for assessing muscle health in OAs. In this direct-to-Phase 2 SBIR application, Myolex proposes to establish EIM, via its new device, the mScan, in conjunction with machine learning cloud-based platform, as a means of obtaining MRI-like quantitative data in muscle of OAs. In Specific Aim 1, in conjunction with aging expert researchers at 3 different clinical sites, we will collect MRI and EIM data on a cohort of healthy older adults, along with standard functional measures as well as measuring specific force, via electrically stimulated muscle contraction. Using this data, in Specific Aim 2, we will develop predictive algorithms, via the penalized regression technique of Lasso (least absolute shrinkage and selection operator), leveraging EIM values to predict muscle volume, muscle specific force, and muscle fatty infiltration. We will then incorporate these algorithms into a cloud-based engine that will provide meaningful, easy-to-interpret values. At the conclusion of this proposed work, we will have developed an accurate, powerful system that clinicians and researchers can use for the rapid assessment of OA muscle health.