RESUMEN
The concept and pathogenesis of acute respiratory distress syndrome (ARDS)were introduced,and the methods for identifying ARDS based on noninvasive parameters in recent years were retrospectively reviewed including regression method, nonlinear fitting method and multi parameter method.The above methods had their advantages and disadvantages summarized. It's suggested that multi noninvasive parameters and machine learning algorithms such as neural network, support vector machine and decision tree be involved in model construction to promote PaO2/FiO2assessment based on noninvasive parameters,so that the rapid diagnosis and real-time monitoring of ARDS can be realized based on noninvasive parameters while there were no need for blood gas analysis.