Rising from the pandemic: Towards sustainable development of machine learning technologies for point-of-care Ultrasound Imaging

床旁超声（POCUS）描述了任何医疗从业者使用便携式超声设备在床边诊断病情的过程，无论是在城市医院，农村诊所还是偏远地区。如果使用得当，POCUS可以减少诊断测试的等待时间，使患者治疗途径更有效，简化转移以节省医疗成本，并挽救生命。虽然这一技术越来越便宜，越来越便于携带，但其使用仍然受到操作人员熟练掌握解释和分析所需的高水平培训的严重限制。通过加拿大数字技术超级集群的250万美元投资以及我们与位于卑诗省伯纳比的加拿大领先POCUS制造商Clarius移动的Health的合作，我们在卑诗省开发了IN-POCUS，即POCUS智能网络。通过在全省部署近100台POCUS设备，我们能够在这些关键的大流行时期为前线医生提供超声技术。这些POCUS设备通过我们UBC团队的机器学习（ML）技术进行了增强，最初重点关注心脏成像、产科和Covid-19诊断。这个NSERC联盟任务项目旨在解决Clarius目前在其产品线可持续发展和ML技术维护战略规划中面临的两个关键挑战。具体而言，由于Clarius设备在BC广泛部署，并且由经验不足的POCUS操作员使用，因此在自动检测ML故障模式以及从新数据中持续学习以确保模型保持最新和最新方面存在重大问题。及时应对这些挑战将确保Clarius保持其全球竞争优势，并维持其目前的指数增长。该联盟项目将培养1名博士后研究员，4名硕士，4名本科生，一名研究助理和一名临床研究员。这些HQP将直接与Clarius合作，与临床最终用户一起建立需求，开发原型并在快速周期内验证解决方案，以领导对患者护理的全球影响。

Point-of-care ultrasound (POCUS) describes the process whereby any healthcare practitioner uses a portable ultrasound device to diagnose conditions at the bedside, whether in an urban hospital, rural clinic, or a remote location. When used appropriately, POCUS can reduce wait times for diagnostic tests, make patient treatment pathways more efficient, streamline transfers to save healthcare costs, and save lives. Although this technology is becoming cheaper and more portable, its use still remains significantly constrained by the high levels of training required by its operators to become proficient in interpretation and analysis. Through $2.5M investment from Canada's Digital Technology Supercluster and our partnership with Clarius Mobile Health, the leading Canadian POCUS manufacturer in Burnaby, BC, we have developed IN-POCUS, Intelligent Network for POCUS, in BC. By deploying close to 100 POCUS devices across the province, we have been able to provide ultrasound technology to frontline physicians at these critical pandemic times. These POCUS devices are augmented with machine learning (ML) technologies from our team at UBC, initially with a focus on cardiac imaging, obstetrics and Covid-19 diagnosis. This NSERC Alliance Missions project is to tackle two of the key challenges Clarius is currently facing in their strategic planning for sustainable development and maintenance of ML technologies in their product line. Specifically, as Clarius devices are deployed broadly across BC and are used by less experienced POCUS operators, there are major concerns in terms of automatic detection of ML failure modes, and continuous learning from new data to ensure models stay current and up-to-date. Timely addressing these challenges will ensure that Clarius maintains its competitive edge globally and sustain its current exponential growth. The Alliance project will train 1 postdoctoral fellow, 4 MSc, 4 undergraduate students, a Research Associate and a clinical fellow. These HQP will work directly with Clarius to establish requirements, develop prototypes and validate solutions in a rapid cycle with clinical end-users to lead the global impact on patient care.