A recent study published in Frontiers in Medicine investigated the development and evaluation of a deep learning model for predicting COVID-19 acute respiratory distress syndrome (ARDS) in Critically Ill Patients. COVID-19 can lead to severe pneumonia, with about 33% of patients at risk of developing severe symptoms and high mortality.

The researchers focused on prevention by identifying high-risk patients through early prediction of COVID-19 ARDS. Artificial intelligence (AI) can aid in disease diagnostics, prognostics, and personalized treatment by handling vast data. Combining computed tomography (CT) scans and clinical data can improve the precision of predicting COVID-19 ARDS and enhance patient outcomes.

The study included patients admitted to the intensive care unit (ICU) of Shanghai Renji Hospital between April and June 2022. Patients aged 18 and above diagnosed with COVID-19 ARDS were included, while those diagnosed on the first day of admission, with missing clinical data or without CT scan results, were excluded.

COVID-19 ARDS was diagnosed based on clinical history, epidemiological contact, positive SARS-CoV-2 test, and the Berlin definition of ARDS. The study collected chest clinical data and CT images post-admission, including comorbidity conditions, demographic information, symptoms, respiratory support methods, vital signs, and various tests.

Four machine learning algorithms were used to establish predictive models for COVID-19 ARDS. The models were trained and validated, and hyperparameters were fine-tuned to avoid overfitting. CT slices were labeled manually or with a deep learning framework based on visual geometry group (VGG)-16.

The study found five independent risk factors associated with COVID-19 ARDS: age, PaO2/FiO2 ratio, C-reactive protein, total T lymphocyte count, and interleukin-6. The XGBoost model had the highest area under the receiver operating characteristics (ROC) curve, indicating its effectiveness in predicting COVID-19 ARDS. A convolutional neural network (CNN) model achieved high accuracy in distinguishing between normal and abnormal CT slices.

An integrated deep learning model combining the XGBoost model and CNN model demonstrated higher accuracy in predicting COVID-19 ARDS compared to individual models based on clinical features or CT images. The findings highlight the potential of deep learning models in predicting and managing COVID-19 ARDS in critically ill patients.

Shanghai Renji Hospital: [link]http://renjihospital.shsmu.edu.cn/

Frontiers in Medicine: [link]https://www.frontiersin.org/journals/medicine

Berlin definition of ARDS: [link]https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7095080/

XGBoost: [link]https://xgboost.readthedocs.io/en/latest/

Convolutional neural network (CNN): [link]https://en.wikipedia.org/wiki/Convolutional_neural_network

The post A Deep Learning Model for Predicting COVID-19 ARDS in Critically Ill Patients appeared first on TS2 SPACE.