Point of Care Heart-Lung Imaging for Patients Presenting with COVID‑19 Symptoms: Artificial Intelligence Precision Modeling for Prediction of Outcomes
The global COVID‑19 pandemic has placed a huge burden on health care systems and upended our way of life. It is caused by the highly infectious virus, SARS-CoV-2. Rapid and accurate testing for viral infection is needed to prevent spread of disease in hospitals and the community so that infected people can be identified early and treated effectively. EARLY COVID‑19 symptoms are similar to many other respiratory viruses, like the flu and common cold, making it hard to identify in sick patients.The current approach to SARS-CoV-2 testing can be inaccurate and slow to return results. Improving diagnosis of COVID‑19 will be even more important as flu season approaches. Our experience to date in treating patients with COVID‑19 symptoms has shown that ultrasound imaging of the heart and lung may help identify positive cases earlier. Ultrasound is fast and safe. It can be performed at the bedside [point of care ultrasound (POCUS)] when patients first enter the hospital, by any physician with remote support from POCUS experts. In this study, we will test if adding heart and lung ultrasound images to existing clinical and laboratory tests can improve the accuracy and speed of COVID‑19 diagnosis. A total of 16 hospitals across British Columbia and Ontario will participate in this study, and we will analyze this large data set using artificial intelligence (AI). Our team has already developed a platform to share ultrasound data and applied AI to study other heart diseases, so we are ready to rapidly apply this approach to COVID-19. If POCUS is effective it would allow earlier treatment and isolation of positive cases, reducing exposure of frontline health care workers and the public to the virus, improving both patient care and the distribution of health resources like personal protective equipment, intensive care beds and ventilators. Because POCUS-based COVID‑19 diagnosis can be performed remotely it could also be applied in long term care facilities and in rural areas.