A Cyber-Physical System for Unified Diagnosis and Treatment of Lung Disease

该项目解决了在诊断和治疗感染性肺部疾病方面的两个主要挑战：(I)对机械通气(MV)危重患者的肺部进行区域性抽样诊断方面的挑战。目前，对MV患者的肺部采样并不规范。事实上，在新冠肺炎大流行之前，重症监护病房(ICU)对疑似肺部感染的高抗生素使用率就是例证。由于感知的风险、支气管镜采样所需的专业知识和时间，经验性抗生素/疗法通常在没有采样的情况下开始。(2)除了缺乏抽样的重复性和准确性外，还需要精确和可重复的药物输送来指导实验性药物的开发。目前，不能准确反映人类传染病的疾病动物模型经常被用于实验药物的开发和迭代/优化。必须开发创新技术，以了解和评估人类的疾病和药物效果。为此，我将汇集机器人学家、人工智能专家、临床医生和翻译治理专家，开发一个基于人工智能的导航平台，该平台可以动态生成肺部地图，用于可重复和准确的采样/药物输送。该地图可用于指导新手在ICU中对危重患者进行可重复、准确和及时的床边肺采样。因此，消除了接触专家操作员的需要，并使ICU肺部采样民主化。此外，导航算法与我们之前开发的用于肺部采样的机器人和半自主控制算法相结合，可以实现准确的药物输送和药物靶标反应分析，以促进针对感染驱动的病理的有效治疗方法的开发。该项目将导致呼吸危重护理领域的根本性创新，使肺部采样/讯问和药物输送民主化，并使未来的研究能够发现、优化和开发治疗严重传染病的新疗法。

This project addresses two main challenges in diagnosis and treatment of infectious lung diseases: (i) Challenges in diagnostic regional sampling of the lung in mechanically ventilated (MV) critically ill patients. Currently, sampling of lung in MV patients is not standardised. Indeed, this was exemplified before the COVID-19 pandemic by the high rate of antibiotic administration in the Intensive Care Unit (ICU) for suspected lung infections. Empirical antibiotics/therapies are often started without sampling due to the perceived risk, required expertise and time for bronchoscopic sampling. (ii) In addition to lack of repeatability and accuracy in sampling, there is a need for precise and repeatable drug delivery to guide experimental drug development. Currently, animal models of disease, which do not accurately reflect human infectious diseases, are often used for the development and iteration/optimisation of experimental drugs. It is essential to develop innovative technologies to understand and evaluate disease and drug effectiveness in humans. To this end, I will bring together roboticists, AI specialists, clinicians, and translational governance experts to develop an AI-based navigation platform that can generate a map of the lung on the fly for repeatable and accurate sampling/drug delivery. The map can be used to guide novice practitioners to perform repeatable, accurate, and timely bedside lung sampling in critically ill patients in ICUs. Thus, obviating the need for access to expert operators and democratising ICU lung sampling. Moreover, the navigation algorithm coupled with our previously developed robot for lung sampling and semi-autonomous control algorithms enables accurate drug delivery and analysis of drug-target responses to facilitate the development of effective therapies for infection-driven pathology. The project will lead to a radical innovation in respiratory critical care to democratise lung sampling/interrogation and drug delivery and enable future research for the discovery, optimisation, and development of novel therapies in severe infectious diseases.