AI algorithms have played a key role in diagnosis and treatment planning for COVID-19. Some of the many ways that AI has played a role in this pandemic include helping in diagnosis and predicting the risk of deterioration. As early as March 2020, researchers were using an AI algorithm that had been trained by data from the 2003 SARS outbreak to predict the number of new infections we could expect to see over time.
Since the beginning of the pandemic, start-ups, universities and biopharma have used AI to better understand the structure of the novel coronavirus, to identify promising new compounds for treatment, to find existing FDA-approved compounds that could be repurposed, and even to design drug molecules that are structurally stable. To study the structure of SARS-CoV-2, the virus that causes COVID-19, researchers at the University of Texas in Austin and the National Institute of Health (NIH) used software called cryoSPARC to create a 3D model of the virus from 2D images captured using cryo-electron microscopy—a technique that can capture molecular structures. The cryoSPARC software, developed by Structural Biotechnology, uses neural networks to tackle the problem of “particle picking” or detecting and isolating protein structures in the microscopic images.
Recursion has used AI to better understand the virus. In a controlled environment, healthy cells were infected with the SARS-CoV-2 virus, and the microscopic images were analyzed using deep learning to identify the physical changes that occur in these cells as a result of the infection. During their research, they discovered some DNA that was distinctly different from earlier coronaviruses, mimicking a protein that helps our bodies to regulate the balance of salt and fluid. Through the use of AI, researchers have been able to pinpoint genetic mutations in COVID-19 that can help with both diagnosis and treatment
Facebook AI Research (FAIR) has created machine learning tools that can process x-rays to help doctors to make predictions about how a given patient’s COVID-19 might progress. One model looked at the amount of supplementary oxygen that a patient might need, while another tried to predict deterioration.
Meanwhile, PhysIQ has been working with an AI-based COVID-19 Decompensation Index (CDI), which uses continuous and multi-parameter vital sign data from wearables so that healthcare providers can remotely survey patients with high-risk COVID-19 from their homes. They can then step in and intervene if the symptoms start to get worse. Many of the symptoms and characteristics of COVID-19, like the loss of smell and its effects on blood coagulation, were discovered through the use of AI. Algorithms are being used to diagnose, manage and treat patients, to watch the spread of the pandemic, to understand its characteristics and to develop new vaccines.