Tele-CF: A practical platform for remote monitoring of cognitive frailty

摘要 认知脆弱(CF)，即身体脆弱和认知障碍的组合存在，是一种强烈的 随着时间的推移，认知能力下降的独立预测者。国际老年学和老年医学协会 和国际营养与老龄化学会建议使用认知脆弱评估 跟踪轻度认知障碍(MCI)向痴呆症、阿尔茨海默病(AD)的进展，以及 失去独立性。然而，目前还没有实用的工具来评估远程医疗访问期间的CF。这是 尤其重要的是，远程医疗在老年人中越来越受欢迎，并越来越多地被 由于目前的新冠肺炎大流行已显著加速了这些趋势。 因此，对基于软件的解决方案的需求尚未得到满足，该解决方案可在 远程医疗访问，并可集成到现有的远程医疗系统中。 在第一阶段，我们成功地设计了一个新的CF测量工具的原型，称为Tele-CF，它使用 基于深度学习的图像处理用于远程测量CF。TELE-CF算法提取运动学 从一段20秒的肘部屈曲和伸展视频中获得前臂运动的特征，并量化虚弱， 迟缓、僵硬和疲惫(这些都是脆弱的表型)，以生成脆弱指数(FI)。FI范围从 从0到1，值越大表示脆弱性越严重。在双重任务下应用时 条件(例如，在同时执行工作记忆任务的同时)，Tele-CF也可以屏蔽认知 减损。在第一阶段，我们论证了Tele-CF在老年人中使用的可行性和概念有效性 成年人群(n=29，年龄：78.6±6.5)，包括18例MCI或轻度痴呆患者，通过比较结果 针对经过验证的工具(基于传感器的工具，用于临床脆弱性评估、双重任务步态和临床认知 Scale，MMSE)。在成功完成第一阶段的所有里程碑后，我们建议完成 Tele-CF的研制及其远程跟踪能力的临床研究 在12个月期间预测现金流量的下降，并预测12个月的现金流量下降。为了达到研究的临床目的， 我们将招募100名临床确诊为MCI或轻度痴呆的成年人(年龄60岁以上)，其中包括50名没有 体弱者和体弱者50人。所有受试者将在临床上使用常规方法进行评估 认知-运动评估工具，在基线、6个月和12个月，也从远程开始 使用Tele-CF每两个月进行一次基准测试。 Tele-CF为临床医生提供了易于使用的桌面应用程序，用于远程评估CF， 现有的远程医疗平台。Tele-CF桌面应用程序将使临床医生能够识别、记录和跟踪 Cf在老年患者中。此外，Tele-CF还将在视频采集和分析方面有更广泛的应用 远程医疗和临床试验中的生物标记物。

ABSTRACT Cognitive frailty (CF), the combined presence of physical frailty and cognitive impairment, is a strong and independent predictor of cognitive decline over time. The International Association of Gerontology and Geriatrics and the International Academy on Nutrition and Aging have recommended the use of cognitive frailty assessment to track the progression of mild cognitive impairment (MCI) towards dementia, Alzheimer's disease (AD), and loss of independence. However, there is no practical tool for assessment of CF during telehealth visits. This is especially important as telehealth is growing in popularity amongst older adults and increasingly accepted by healthcare payers, and as the ongoing COVID-19 pandemic has significantly accelerated these trends. Therefore, there is an unmet need for a software-based solution that enables remote assessment of CF during telehealth visit and can be integrated into existing telehealth systems. In Phase I, we successfully designed a prototype of a novel CF measurement tool, called Tele-CF, that uses deep learning-based image processing to remotely measure CF. The Tele-CF algorithms extracts kinematic features of the forearm motion from a video of 20-second elbow flexion and extension, and quantifies weakness, slowness, rigidity, and exhaustion (which are phenotypes of frailty) to generate a frailty index (FI). FI ranges from zero to one, with higher values indicating progressively greater severity of frailty. When applied under dual-task conditions (e.g., while simultaneously performing a working memory task), Tele-CF can also screen cognitive impairment. In Phase I, we demonstrated the feasibility and proof of concept validity of Tele-CF for use in older adult population (n=29, age: 78.6±6.5), including 18 subjects with MCI or mild dementia, by comparing the results against validated tools (a sensor-based tool for in-clinic frailty assessment, dual-task gait, and a clinical cognitive scale, MMSE). After successfully achieving all milestones of Phase I, we are proposing to complete the development of Tele-CF and carry out a clinical study to demonstrate the ability of Tele-CF for remote tracking of CF over a 12-month period and to predict decline in CF at 12 months. To achieve the clinical aim of the study, we will recruit 100 adults (age 60+) with clinically confirmed MCI or mild dementia including 50 subjects without physical frailty and 50 subjects with physical frailty. All subjects will be assessed in clinic using conventional cognitive-motor assessment tools, at baseline, 6 months, and 12 months, and also remotely starting from the baseline every two months using Tele-CF. Tele-CF provides an easy-to-use desktop application for remote assessment of CF for clinicians that works with existing telehealth platforms. Tele-CF desktop application will enable clinicians to identify, document, and track CF in older patients. In addition, Tele-CF will have broader applications in collection and analysis of video biomarkers in telehealth and clinical trials.