In this competitive revision, within the same scope of developing and deploying algorithms to make a quantum leap in clinical diagnosis as that in our current U01EB021183, we would like to revise the original aims to add a new Aim to leverage our expertise in the areas of algorithm development and clinical translation to make immediate contributions to combat the COVID-19 pandemic. Specifically, we propose to develop and deploy artificial intelligence (AI) methods to enable chest x-ray radiography (CXR) as an alternative diagnostic tool to diagnose COVID-19 pneumonia, to rapidly triage patients for appropriate treatment, to monitor the treatment response in a contained environment, and to optimize the distribution of the limited medical resources during the current COVID-19 crisis.
In this project, our overarching objective is to develop automated artificial intelligence (AI)-based algorithms to help radiologists to differentiate COVID-19 related pneumonia from other non-COVID-19 related pneumonia using CXR images. The advantages of the proposed AI equipped CXR technique include: i) widely available, ii) inexpensive, iii) excellent coronavirus exposure profile for patient, technologist, and equipment, and iv) rapid and automated DL interpretation, which is effectively instantaneous.
