Supporting future radiologists’ education during and after COVID-19 through an AI-supported training environment

由于新冠肺炎的流行，许多放射学实习生和学生报告说他们的教育受到了干扰。社交距离措施和旅行限制限制了传统的实习生-教员工作站教学，面对面授课已被现场会议、课堂录音或完全取消。放射学学生已经认识到，尽管各种数字资源提供了理论知识，但以实践为基础的培训的替代方案有限。为了解决这一问题，我们提出了一种人工智能支持的高保真虚拟X射线报告培训环境。培训系统使用真实诊断环境的界面，受训者可以用英文自由文本报告X射线图像，模拟真实的报告实践。人工智能算法解释原始报告，受训者会收到即时反馈。由于该系统是基于网络的，它可以接触到受全球大流行影响的受训者和学生。此外，通过利用以实践为基础的考试培训制度，它可以对教育机构有所帮助。通常，评审员会单独审阅学生的报告并提供反馈。我们的系统允许对多名学生同时进行自动评分和实时反馈。我们是一个由技术专家和医生组成的领域专家团队，专注于开发智能解决方案来支持放射科医生。这种流行病广泛扰乱了人们的生活，包括放射学实习生的生活，并迫使我们改变提供教育的方式。由于这场流行病的结束仍不明朗，我们设想创建一个可持续的解决方案，为放射学教育的未来做出贡献，并提供不断扩大的内容，以支持学员和学生的终身学习旅程。扩大Impact资金使我们能够将我们的支持从放射学实习生扩大到放射技师和其他医疗专业人员。英国正面临放射科医生短缺的局面，新冠肺炎疫情给医院本已有限的报告能力带来了巨大压力。为了更好地管理对放射学服务的需求，我们需要更多的放射技师有资格报告影像研究。我们将开发评估工具，使用标准化的病例包，在不同的医疗专业人员之间使用基准分数来评估图像读取的准确性。这将有助于医院确保报告质量，并自信地将放射技师纳入报告过程。此外，提高其他需要解读医学图像的医疗专业人员的诊断技能，如急诊科医生、初级医生和护士，可以帮助更好地管理不断增长的需求。病人将得到有效的分流，并在最初会诊时提供正确的治疗。有了我们的平台，医院可以有效地提高现有和即将到来的劳动力的技能。它可以帮助医院提高更好地管理需求的能力，并为未来建立一个有弹性的医疗体系。

Due to the COVID-19 pandemic, many trainees and students in radiology are reporting disruption to their education. Social distancing measures and travelling restrictions have limited the traditional trainee-faculty member workstation teaching, and face-to-face lectures have been replaced by live conferences, class recordings, or cancelled altogether. Radiology students have recognised that while various digital resources provide theoretical knowledge, alternatives to practice-based training are limited. To solve the problem, we propose an AI-supported high-fidelity virtual training environment for reporting x-rays. The training system uses an interface of a real diagnostic environment and trainees can report x-ray images with an English free text, emulating the real reporting practice. The AI algorithm interprets the raw report and trainees receive instant feedback. As the system is web-based, it can reach trainees and students whose education has been affected by the global pandemic. Furthermore, it can be of value to educational institutions by utilising the training system for practice-based examinations. Usually, assessors individually review and provide feedback to student's reports. Our system allows automated grading and real-time feedback simultaneously to multiple students. We are a team of domain experts formed of technologists and doctors, focused on developing intelligent solutions to support radiologists. The pandemic has widely disrupted lives, including that of radiology trainees, and forcing changes to how we deliver education. As the end to this pandemic is still unclear, we envision to create a sustainable solution that contributes to the future of radiology education and provide an ever-expanding content to support trainees and students in their life-long learning journey.Extension for Impact funding allows us to expand our support from radiology trainees to radiographers, and other medical professionals. The UK is facing a shortage of radiologists and the COVID-19 pandemic is causing a major strain on the hospital’s already limited reporting capacity. To better manage the demand for radiology services, we need more radiographers to be qualified to report imaging studies. We will develop assessment tools that use standardized packages of cases with benchmark scores across different medical professionals to evaluate the accuracy of image reading. This will help hospitals to ensure reporting quality and confidently integrate radiographers in the reporting process. Moreover, improving the diagnostic skills of other medical professionals required to interpret medical images, such as emergency department doctors, junior doctors, and nurses, can help better manage the increasing demand. Patients will be triaged effectively and correct treatment will be provided at the initial point of consultation. With our platform, hospitals can effectively upskill the current and coming workforce. It can help hospitals increase the capability to better manage the demand and build a resilient healthcare system for the future.