Improving Therapeutic Decision-Making through Risk-stratification of Severe COVID-19 Patients

Abstract The advent of cellular therapies, particularly the use of SARS-CoV-2-specific T cells (CoV-2-STs), offers a promising avenue for the treatment of severe COVID-19. Presice stratification of COVID-19 patients is essential to identify those at high risk who may benefit from intensive therapeutic strategies. Utilizing longitudinal biomarker data from a randomized phase 1–2 trial which was implemented during the delta COVID-19 variant and compared the efficacy of treatment with CoV-2-STs plus standard-of-care (SoC) against SoC alone in severe COVID-19 patients, we conducted a post hoc, linear discriminant analysis to identify severely infected patients at increased risk of deterioration. We developed a feature importance strategy to detect key determinants influencing patient outcomes post-treatment. Our results demonstrated that crucial biological classifiers could predict treatment response with over 87% accuracy, validated through multiple-fold cross-validation. This predictive model suggested that the survival of the SoC-only, control group, patients, could have been improved by 30%, if they had received CoV-2-STs therapy. Additionally, in order to aid therapeutic decision-making, we generated a computational tool, capable of identifying those patients in whom an additional to SoC intervention, may be required to avert adverse outcomes. Overall, this computational approach represents a step forward in personalized medicine, offering a new perspective on the stratification and management of severe COVID-19 patients.