RESUMO
Background: During COVID-19 epidemics several artificial-intelligence neural networks (ANN) systems were developed classify the risk of disease progression to respiratory failure and death, providing aid for clinical decision. However, for optimal results these models should link multiple medical data in a simple model. In this study we analyse the in-hospital mortality and mechanical ventilation risk using combination ANN based rapid computed tomography assessment tool and selected clinical variables. Method(s): Data of 4317 COVID-19 hospitalized patients including 266 cases required mechanical ventilation were analysed using newly constructed and locally trained ANN algorithm. Demographic, clinical, laboratory, and ANNbased lung inflammation data were analysed using proportional Cox Hazards model and estimate in-hospital mortality and intensive care admission risk. Result(s): Overall in-hospital mortality associated with ANN-assigned percentage of the lung involvement (HR 5.72 (95%CI: 4.4-7.43), p< 0.001 for the patients with >50% of lung tissue affected by COVID-19 pneumonia), age category (HR 5.34 (95%CI: 3.32-8.59) for cases >80 years, p< 0.001), procalcitonin > 2 (HR: 2.1 (95%CI: 1.59-2.76) ng/ml p< 0.001, C-reactive protein level category (max. HR 2.11 (95%CI: 1.25-3.56) for CRP >100 mg/dL, p=0.004), estimated glomerular filtration rate (max HR 1.82 (95%CI: 1,37-2,42), p< 0.001 for eGFR < 30 ml/min) and troponin increase above upper limit normal level (HR: 2.14 (95%: 1.69-2.72, p< 0.001) (Figure 1). Furthermore, risk of mechanical ventilation also associated with ANN-based percentage of lung inflammation (HR 13.2 (8.65-20.4), p< 0.001 for patients with >50% involvement), age, procalcitonin > 2 ng/ml (HR: 1.91 (95%CI: 1.14-3.2), p=0.14 estimated glomerular filtration rate (HR 1.82 (1.2-2.74), p=0.004 for eGFR < 30 ml/min) but also clinical variables, including (HR: 2.5 (95%CI: 1.91-3.27), p< 0.001), cardiovascular and cerebrovascular disease (HR: 3.16 (95%CI: 2.38-4.2), p< 0.001), and chronic pulmonary disease (HR: 2.31 (95%CI: 1.44-3.7), p< 0.001). Conclusion(s): ANN-based lung tissue involvement was the strongest predictor of unfavorable outcomes in COVID-19, and represent valuable support tools for clinical decisions. (Figure Presented).