Neonatal asphyxia – predicting its occurrence and neurodevelopment of children after hypothermia treatment in the context of machine learning models – literature review

Abstract

Introduction and Objective. Perinatal asphyxia (PA) is a condition of impaired blood gas exchange which, if persistent, leads to progressive hypoxemia and hypercapnia. PA is characterized by high neonatal mortality or neurodevelopmental disorders in the child's later life. PA is closely related to hypoxic-ischemic encephalopathy (HIE), which is a complication of perinatal hypoxia. In recent years, there has been a growing interest in the use of machine learning (ML) and artificial intelligence (AI) in medicine. The development of machine learning models can improve patient care by applying them to diagnosis, prognosis, decision support, and treatment recommendations. Review Methods. The data for the article was found using the Web of Science, PubMed, Scopus and Google Scholar websites which were thematically selected for work. Brief description of the state of knowledge. Currently, no sufficiently effective prophylactic and therapeutic methods can prevent HIE or death in newborns with PA. Hypothermia is currently the only available method that seems to improve the prognosis in neonates with PA; however, it is not completely effective. In recent years, research has been conducted in the use of machine learning models to predict the occurrence of neonatal asphyxia, based on the occurrence of specific risk factors. Additionally, research has been carried out on the use of neuroimaging in predicting the neurodevelopmental condition of children after neonatal asphyxia. Summary. Presently, there are no preventive and therapeutic methods that can prevent HIE or death in newborns with severe perinatal hypoxia. In addition to basic knowledge about asphyxia, the review presents issues that could become the beginning of future research in the use of machine learning in neonatology.