A breakthrough mobile phone technology that aids in early detection of COPD

项目摘要/摘要 COPD是全球第四大杀手，部分原因是一线药物未能 检测疾病。早期发现不佳导致不良结局和高医疗保健 成本。远程听诊器*或“TS”是一项新技术，它允许从 使用普通手机进行强大的血流动力学和呼吸动力学分析。 TS计划用于初级保健，特别是远程保健，并在计划之前使用 体检。它的目的是提醒初级保健医生注意身体检查的必要性 检查和可能的转介。在这些情况下，没有其他类似的测试 选择。 TS有三个重要的创新。首先，是一项临床应用的创新，它带来了远程， 对任何拥有智能手机的人进行即兴的患者评估。客户发现已验证 工作流程和报销策略都与商业成功保持一致。此外， FDA的指导建议机构批准的可行性。第二个是TS的新方法 制造人工智能算法。TS使用OEM手机麦克风捕获的声学数据进入 描述器官血流动力学和空气动力学的严谨、基于物理的特征。 然后使用特征或自变量来创建二进制分类模型 采用Logistic回归分析。通过依赖物理，TS永远不需要使用神经网络或 过度匹配策略引发了对数据密集度、模型稳定性、监管 透明度和未知的可靠性。第三个创新是，依靠被动 声学数据，TS通过直接观察器官动力学来分类器官功能，而不是 从强制通气量。TS使用了许多物理学的想法，这些想法直到公平地 最近。通过依靠动力学，TS提供了一种新的强制空气方法，最终可能 有助于我们了解呼吸状况，即使在PFT实验室也是如此。 这将是第三项证明心脏疾病分类的大规模人体研究。 肺功能。它是在发表了一项关于COVID检测的研究之后进行的，不久将 发表关于TS重现射血超声心动图估测的能力的研究 分数。听诊器记录的分析表明，COPD的诊断将依赖于 右心侧血流动力学和低频肺血流动力学的组合。一个 提出了大量的研究，验证了移动电话作为一种 具有检测和诊断慢性阻塞性肺疾病能力的听诊器。 这项研究的目的是开发、评估和测试一种允许普通移动设备 检测慢性阻塞性肺疾病的手机。我们的假设是，TS技术可以通过匹配检测到COPD 通过电话录音进行诊断。诊断将由医生在以下情况下确定 考虑黄金标准测试。模型的性能将通过使用以下方法进行评估 常用的统计方法如接收器工作曲线下面积AUC， 敏感性、特异性、准确性。这项研究将有三个里程碑，基于招聘 分别为30例、60例和100例。 我们的方法是基于从分配到医院PFT实验室的患者招聘，因此 录音和金标测试同时进行。我们的团队由一位广泛的成员组成 一系列人才，包括TS的发明者，肺病学专业知识，研究总监 为运行测试、统计专业知识和项目的医院提供急救医学 管理层。 *远程听诊器或“TS”是由弗莱明科学公司开发的正在申请专利的技术。

Project Summary / Abstract COPD is the fourth leading global killer, in part, because of the failure of frontline medicine to detect the disease. Poor early detection leads to unfavorable outcomes and high healthcare costs. Tele-stethoscope* or “TS” is novel technology that allows acoustic data captured from ordinary mobile phones to perform robust analysis of hemodynamics and pneumodynamics. TS is intended to be used in primary care particularly in Tele-health and before scheduled physical examination. Its purpose is to alert primary care physicians to the need for physical examination and possibly referral. In these situations, there are no other comparable testing options. TS has three important innovations. First, is a clinical use innovation which brings remote, impromptu patient evaluation to anyone with a smartphone. Customer discovery has validated both a workflow and reimbursement strategy consistent with commercial success. Further, FDA guidance suggests viability of agency clearance. Second is TS’s novel approach to making AI algorithms. TS uses acoustic data, captured by OEM phone microphones, into rigorous, physics-based features that describe organ hemodynamics and pneumodynamics. The features or independent variables are then used to create binary classification models with logistic regression. By relying on physics, TS never has to use neural networks or overfitting strategies that have raised concerns about data intensity, model stability, regulatory transparency and unknown reliability. The third innovation is that, by relying on passive acoustic data, TS classifies organ function by observing organ dynamics directly rather than from forced air capacity. TS uses many ideas from physics that were not available until fairly recently. By relying on dynamics, TS offers a novel approach to forced air that may eventually be additive to our understanding of respiratory conditions even in PFT labs. This will be the third large scale human study that demonstrates classification of cardio - pulmonary functionality. It follows a published study on COVID detection as well as soon to be published study on the ability of TS to reproduce echocardiogram estimates of ejection fraction. Analysis of stethoscope recordings suggests that the COPD diagnostic will rely on a combination of right heart side hemodynamics and low frequency lung pneumodynamics. A significant body of research is presented which validates the use of mobile phones as an auscultation device with the ability to detect and diagnosis COPD. The Aim of this study is to develop, evaluate, and test an algorithm that allows ordinary mobile phones to detect COPD. Our hypothesis is that TS technology can detect COPD by matching phone acoustic recordings to diagnosis. Diagnosis will be determined by a physician after considering gold standard testing. Performance of the models will be evaluated by using common statistical approaches such as area under the receiver operating curve AUC, sensitivity, specificity, accuracy. The research will have 3 milestones based on recruitment of 30, 60 and 100 patients respectively. Our approach is based on recruitment from patients assigned to a hospital PFT lab so that recordings and gold standard testing are done concurrently. Our team consists of a broad range of talent including the inventor of TS, pulmonology expertise, the research director of emergency medicine for the hospital running the tests, statistical expertise, and project management. * Tele-stethoscope or "TS" is patent pending technology developed by Fleming Scientific.