Background: Although several studies have been launched towards the prediction of risk factors for mortality and admission in the intensive care unit (ICU) in COVID-19, none of them focuses on the development of explainable AI models to define an ICU scoring index using dynamically associated biological markers. Methods: We propose a multimodal approach which combines explainable AI models with dynamic modeling methods to shed light into the clinical features of COVID-19. Dynamic Bayesian networks were used to seek associations among cytokines across four time intervals after hospitalization. Explainable gradient boosting trees were trained to predict the risk for ICU admission and mortality towards the development of an ICU scoring index. Results: Our results highlight LDH, IL-6, IL-8, Cr, number of monocytes, lymphocyte count, TNF as risk predictors for ICU admission and survival along with LDH, age, CRP, Cr, WBC, lymphocyte count for mortality in the ICU, with prediction accuracy 0.79 and 0.81, respectively. These risk factors were combined with dynamically associated biological markers to develop an ICU scoring index with accuracy 0.9. Conclusions: to our knowledge, this is the first multimodal and explainable AI model which quantifies the risk of intensive care with accuracy up to 0.9 across multiple timepoints.

A central paradigm of immunity is that interferon (IFN) mediated antiviral responses precede the pro-inflammatory ones, optimizing host protection and minimizing collateral damage ^1,2 . Here, we report that for COVID-19 this does not apply. By investigating temporal IFN and inflammatory cytokine patterns in 32 COVID-19 patients hospitalized for pneumonia and longitudinally followed for the development of respiratory failure and death, we reveal that IFN-λ and type I IFN production is both diminished and delayed, induced only in a fraction of patients as they become critically ill. On the contrary, pro-inflammatory cytokines such as TNF, IL-6 and IL-8 are produced before IFNs, in all patients, and persist for a prolonged time. By comparison, in 16 flu patients hospitalized for pneumonia with similar clinicopathological characteristics to COVID-19 and 24 milder non-hospitalized flu patients IFN-λ and type I IFN are robustly induced, earlier, at higher levels and independently of disease severity, while pro-inflammatory cytokines are only acutely and transiently produced. Notably, higher IFN-λ levels in COVID-19 patients correlate with lower viral load in bronchial aspirates and faster viral clearance, and a higher IFN-λ:type I IFN ratio with improved outcome of critically ill patients. Moreover, altered cytokine patterns in COVID-19 patients correlate with longer hospitalization time and higher incidence of critical disease and mortality compared to flu. These data point to an untuned antiviral response in COVID-19 contributing to persistent viral presence, hyperinflammation and respiratory failure.