In response to an open call for applications in Fall 2020, the Canadian Radiological Foundation (CRF) and the Canadian Heads of Academic Radiology (CHAR) are pleased to award funding to two research projects related to COVID-19:
- Using AI to Determine Prevalence of Pre- or Asymptomatic COVID-19 in Patients Undergoing Chest CT for a Non-COVID Indication within Eastern Ontario – A Collaborative Multicenter Observational Study
- Using Virtual Reality as a Remote Educational Tool During the COVID-19 Pandemic
The hope is that this directed funding will help our community, and most importantly our patients, during these challenging times. Thank you to Bayer for partially supporting these awards.
Using AI to Determine Prevalence of Pre- or Asymptomatic COVID-19 in Patients Undergoing Chest CT for a Non-COVID Indication within Eastern Ontario – A Collaborative Multicenter Observational Study
Dr. Robert Lim, MB, BCh, BAO, FRCPC
University of Ottawa, The Ottawa Hospital
CT chest has been advocated by some to screen the general population because it may have greater sensitivity than traditional reference standard RT-PCR tests and may detect pre- or asymptomatic disease prior to RT-PCR positivity. A recent WHO expert panel recommendation supported the use of screening CT where the test is performed for reasons other than screening. Early disease detection could reduce transmission by pre- or asymptomatic carriers, have significant impact on “flattening the curve,” and reduce the severity of the second wave.
The first phase of our project will be fine tuning our deep neural network to discriminate COVID-19 from non-COVID-19 infections in our Eastern Ontario and Nunavut population. In the second phase, the tailored neural network will be applied to CT chest studies performed across our region for non-COVID-19 indications. These studies offer an economical and real-time method of detecting and monitoring the prevalence of unexpected CT chest changes that may be consistent with COVID-19 and potentially indicative of pre- or asymptomatic patients.
Positive cases detected by the AI algorithm will be reviewed by two (2) chest radiologists and one (1) respirologist who will classify the patient’s CT scan into the RSNA COVID-19 classification scheme: 1) typical; 2) atypical/indeterminate; 3) negative. Any study positive for typical or atypical/indeterminate features will be flagged with the originating referrer and RT-PCR recommended. Our hypothesis is a deep learning algorithm will identify COVID-19 features with high sensitivity and accuracy and the CT appearances typical of COVID-19 are expected to be proportional to the background prevalence of COVID-19 infections derived from population RT-PCR screening.
Using Virtual Reality as a Remote Educational Tool During the COVID-19 Pandemic
Dr. Yuhao Wu, MD
Supervised by Brent Burbridge, MD, FRCPC
Royal University Hospital, Saskatoon
The COVID-19 pandemic has brought significant challenges to undergraduate radiology education. Many undergraduate radiology electives have been temporarily suspended due to physical distancing measures. Virtual reality (VR) software, an innovative tool which creates an immersive educational experience through the use of specialized headsets, can be used to facilitate physically distanced virtual learning. VR creates a shared simulated 2D/3D space workstation where learners and mentors can collaborate without having to be in the same physical environment.
The project will leverage the benefits of using VR to deliver remote education to undergraduate medical students. We will use the SieVRt software, which features “a reading room in a headset” experience, to generate a physically simulated radiology workstation. Using this tool, students will be able to view and interact with medical images from a database of preselected teaching files. They will also be able to review these cases with staff Radiologists in a physically distanced manner.
Through this project, we aim to show the impact of incorporating VR as a remote learning tool on students’ theoretical and practical knowledge in radiology. We also hope to discover the advantages, as well as barriers and/or limitations, of using VR as an educational tool in radiology.
Congratulations to the award winners.