摘要
In recent years,the concept of phenotyping in acute respiratory distress syndrome(ARDS)–where ARDS is viewed as a broad phenotype encompassing multiple sub-phenotypes–has gained considerable interest.[1]Machine learning and statistical clustering analyses of data from five randomized controlled trials(RCTs)have identified distinct ARDS subtypes–primarily hyperinflammatory and hypoinflammatory profiles–based on clinical characteristics and blood biomarkers associated with coagulopathy,endothelial injury,and inflammation.
