American scientists have used machine learning algorithms together with k-fold cross-validation to better predict the occurrence of severe COVID-19. The researchers used chest computed tomography datasets together with proinflammatory cytokines levels including IL-6, IL-8 and TNF-α from over 150 patients at Mount Sinai Health System in New York. The accumulated results may help clinicians to undertake faster and stronger therapeutic intervention in patients at high risk of death due to COVID-19. The paper published in Helioyn.
https://www.sciencedirect.com/science/article/pii/S2405844022014542?via%3Dihub