THANKS to artificial intelligence, the development of a certain disease related to coronavirus infection can be predicted. The project is from the University of Padua which has developed a “COVID-19 AVI score” app in which to enter the clinical data of patients, so as to immediately personalize the care of Covid-19 patients. The app can be consulted and used by the specialist doctor on the Padua University website. The goal is to arrive through the examination of the variables of each patient, a better medical history and a more appropriate therapeutic management.
The ‘virtual’ virus mimics Covid and knows how the epidemic travels
by Noemi Penna
The study started from the analysis of 432 morphological variables starting from the consideration that there are three pathological phenotypes that a patient infected with Sars-CoV-2 can present: a prevalent alveolar damage, or a vascular damage and a third with the coexistence of two injuries. Through computational statistics and artificial intelligence, the so-called “machine learning”, the team of researchers from the hospital-university of Padua distinguished the pathological phenotypes in these three groups. The goal is to arrive through the examination of the variables of each patient, a better medical history and a more appropriate therapeutic management.
Covid: the algorithm that keeps us away from hospitals
by Giuliano Aluffi
“The statistical correlations of pathological phenotypes with numerous clinical, laboratory, immunological and radiological data – comments Fiorella Calabrese, professor of Anatomy Pathology – identified the vascular phenotype as the one closely linked to viral infection, while in the forms in which alveolar damage prevails, a correlation was found with other factors, such as the coexistence of previous pulmonary diseases, co-infections , of longer duration of the disease. The ‘Random Forest’ statistical algorithm has identified some clinical variables (body mass index, body temperature, respiratory rate) and laboratory (white blood cell, lymphocyte and platelet count, D-dimer and PaCO2 value) which, collected at the time of hospital admission, could stratify patients into different clinical subgroups at risk of developing a given pathological phenotype, alveolar or vascular “.
.