Vasileios C Pezoulas1, Konstantina D Kourou1, Costas Papaloukas1,2, Vassiliki Triantafyllia3, Vicky Lampropoulou3, Eleni Siouti3, Maria Papadaki3, Maria Salagianni3, Evangelia Koukaki4, Nikoletta Rovina4, Antonia Koutsoukou4, Evangelos Andreakos3, Dimitrios I Fotiadis1,5. 1. Unit of Medical Technology and Intelligent Information Systems, Department of Materials Science and Engineering, University of Ioannina, GR45110 Ioannina, Greece. 2. Department of Biological Applications and Technology, University of Ioannina, GR45100 Ioannina, Greece. 3. Laboratory of Immunobiology, Center for Clinical, Experimental Surgery and Translational Research, Biomedical Research Foundation of the Academy of Athens, GR11527 Athens, Greece. 4. Intensive Care Unit (ICU), 1st Department of Respiratory Medicine, Medical School, National and Kapodistrian University of Athens, 'Sotiria' General Hospital of Chest Diseases, GR11527 Athens, Greece. 5. Department of Biomedical Research, Foundation for Research and Technology-Hellas, Institute of Molecular Biology and Biotechnology (FORTH-IMBB), GR45110 Ioannina, Greece.
Abstract
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.
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.
Authors: Vasileios C Pezoulas; Konstantina D Kourou; Fanis Kalatzis; Themis P Exarchos; Aliki Venetsanopoulou; Evi Zampeli; Saviana Gandolfo; Fotini Skopouli; Salvatore De Vita; Athanasios G Tzioufas; Dimitrios I Fotiadis Journal: Comput Biol Med Date: 2019-03-07 Impact factor: 4.589
Authors: Benjamin Gallo Marin; Ghazal Aghagoli; Katya Lavine; Lanbo Yang; Emily J Siff; Silvia S Chiang; Thais P Salazar-Mather; Luba Dumenco; Michael C Savaria; Su N Aung; Timothy Flanigan; Ian C Michelow Journal: Rev Med Virol Date: 2020-07-30 Impact factor: 6.989
Authors: Adam J Singer; Eric J Morley; Kristen Meyers; Rafael Fernandes; Alison L Rowe; Peter Viccellio; Henry C Thode; Alexander Bracey; Mark C Henry Journal: Ann Emerg Med Date: 2020-05-11 Impact factor: 5.721
Authors: Brandon Michael Henry; Gaurav Aggarwal; Johnny Wong; Stefanie Benoit; Jens Vikse; Mario Plebani; Giuseppe Lippi Journal: Am J Emerg Med Date: 2020-05-27 Impact factor: 2.469
Authors: Zirun Zhao; Anne Chen; Wei Hou; James M Graham; Haifang Li; Paul S Richman; Henry C Thode; Adam J Singer; Tim Q Duong Journal: PLoS One Date: 2020-07-30 Impact factor: 3.240
Authors: J Zhang; X Wang; X Jia; J Li; K Hu; G Chen; J Wei; Z Gong; C Zhou; H Yu; M Yu; H Lei; F Cheng; B Zhang; Y Xu; G Wang; W Dong Journal: Clin Microbiol Infect Date: 2020-04-15 Impact factor: 8.067
Authors: Enrico Longato; Mario Luca Morieri; Giovanni Sparacino; Barbara Di Camillo; Annamaria Cattelan; Sara Lo Menzo; Marco Trevenzoli; Andrea Vianello; Gabriella Guarnieri; Federico Lionello; Angelo Avogaro; Paola Fioretto; Roberto Vettor; Gian Paolo Fadini Journal: Comput Methods Programs Biomed Date: 2022-05-11 Impact factor: 7.027