The following is a summary of “Automatic ARDS surveillance with chest X-ray recognition using convolutional neural networks,” published in the August 2024 issue of Critical Care by Ye et al.

Researchers conducted a retrospective study to design, validate, and evaluate the accuracy of a deep-learning model capable of determining chest X-rays between pneumonia, acute respiratory distress syndrome (ARDS), and normal lungs.

They performed a diagnostic study using chest X-ray images from adult patients admitted to a medical intensive care unit between January 2003 and November 2014. X-ray images from 15,899 patients were assigned to 3 categories: ARDS, Pneumonia, and Normal.

The results showed a two-step convolutional neural network (CNN) pipeline was designed and tested to differentiate between the 3 patterns with sensitivity ranging from 91.8% to 97.8% and specificity from 96.6% to 98.8%. The CNN model was validated with a sensitivity of 96.3% and specificity of 96.6% by a previous dataset of patients with Acute Lung Injury (ALI)/ARDS. 

Investigators concluded the deep learning model based on chest X-ray pattern recognition could help distinguish patients with ARDS from patients with normal lungs, providing faster results than digital surveillance tools based on text reports.

Source: sciencedirect.com/science/article/abs/pii/S0883944124002818