What can imaging tell us about SARS-Cov-2 infection

The COVID-19 pandemic has resulted in hundreds of millions infections and millions of deaths worldwide. Morbidity and mortality are primarily driven by SARS-CoV-2-induced lung injury culminating in Acute Respiratory Distress Syndrome (ARDS). Physicians are faced with the task of rapidly evaluating patients with suspected COVID-19 and deciding whether they can go home, if they need hospitalization and if they do, how sever is infection. Furthermore, early data suggests that even non-critically ill survivors of COVID-19 have significant residual impairments in lung function, even though precise mechanisms contributing to lung injury and promoting repair after SARS-CoV-2 infection are unclear.

Medical imaging provides unique spatial and temporal clues of the severity, progression and long-term consequences of SARS-CoV-2 infection. For these reasons, COVID-19 patients are routinely imaged with computed tomography (CT) or X-ray imaging. Advanced image analytics, especially utilizing hand-crafted or deep learning radiomics extraction of features can help identifying quantitative imaging biomarkers of COVID-19 that can help with clinical decisions.

In this lecture, we will review medical imaging related to COVID-19 and emphaze richness of information that it contains. In particular, we will focus on the need and opportunity for advanced image analytics, which provides unique insight into quantification of the SARS-CoV-2 infection, its extent, and long-term consequences. We will review important achievements of the analyses of ARDS, as well as continued need for further development of new image analytics methodologies that will help managing COVID-19 survivors. We will highlight importance of international collaboration in merging imaging datasets as well as the need for inter-disciplinary approach between physics, radiology and clinical sciences to extract relevant information that helps addressing COVID-19 challenges.