Description

Acute lung injury/ acute respiratory distress syndrome is one of the most common and complex critical illnesses in clinical practice, with a high mortality rate of 45% to 50%. Currently, effective therapeutic strategies for this condition are still lacking. Increasing evidence suggests that the significant heterogeneity of this disease plays a crucial role in the poor treatment outcomes and high mortality rates observed in patients. Therefore, this study aims to analyze the heterogeneity of acute lung injury/ acute respiratory distress syndrome patients and establish a clinical classification system for acute lung and extrapulmonary organ injuries.

The objectives of this study include establishing a nationwide clinical database and biobank for acute lung injury / acute respiratory distress syndrome by collecting clinical data and biological samples from various provinces. By overcoming the barriers posed by diverse and heterogeneous data sources, mathematical and machine learning models will be utilized to construct clinical, physiological, and biological classification systems for acute lung and extrapulmonary organ injuries. The proposed classification model will be validated multiple times using international public databases and prospective acute lung injury/acute respiratory distress syndrome cohorts to ensure its stability and generalizability. The mapping relationship between different classifications and patient prognosis as well as treatment responsiveness will be explored.

Moreover, a machine learning-based supervised technique will be applied to develop a bedside simplified model (Point-of-Care model) and establish a bedside clinical classification decision system. Ultimately, this research aims to provide a foundation for standardized and precision-guided clinical diagnostic and therapeutic pathways, promoting improved treatment outcomes and overall prognosis in acute lung injury/ acute respiratory distress syndrome.