Deep Learning With Chest Radiographs for Making Prognoses in Patients With COVID-19: Retrospective Cohort Study.

BACKGROUND: An artificial intelligence (AI) model using chest radiography (CXR) may provide good performance in making prognoses for COVID-19. OBJECTIVE: We aimed to develop and validate a prediction model using CXR based on an AI model and clinical variables to predict clinical outcomes in patients with COVID-19. METHODS: This retrospective longitudinal study included patients hospitalized for COVID-19 at multiple COVID-19 medical centers between February 2020 and October 2020. Patients at Boramae Medical Center were randomly classified into training, validation, and internal testing sets (at a ratio of 8:1:1, respectively). An AI model using initial CXR images as input, a logistic regression model using clinical information, and a combined model using the output of the AI model (as CXR score) and clinical information were developed and trained to predict hospital length of stay (LOS) ≤2 weeks, need for oxygen supplementation, and acute respiratory distress syndrome (ARDS). The models were externally validated in the Korean Imaging Cohort of COVID-19 data set for discrimination and calibration. RESULTS: The AI model using CXR and the logistic regression model using clinical variables were suboptimal to predict hospital LOS ≤2 weeks or the need for oxygen supplementation but performed acceptably in the prediction of ARDS (AI model area under the curve [AUC] 0.782, 95% CI 0.720-0.845; logistic regression model AUC 0.878, 95% CI 0.838-0.919). The combined model performed better in predicting the need for oxygen supplementation (AUC 0.704, 95% CI 0.646-0.762) and ARDS (AUC 0.890, 95% CI 0.853-0.928) compared to the CXR score alone. Both the AI and combined models showed good calibration for predicting ARDS (P=.079 and P=.859). CONCLUSIONS: The combined prediction model, comprising the CXR score and clinical information, was externally validated as having acceptable performance in predicting severe illness and excellent performance in predicting ARDS in patients with COVID-19.

Respiratory Protection Effect of Ear-loop-type KF94 Masks according to the Wearing Method in COVID-19 Pandemic: a Randomized, Open-label Study.

BACKGROUND: Ear-loop-type Korean Filter 94 masks (KF94 masks, equivalent to the N95 and FFP2) are broadly used in health care settings in Korea for the coronavirus disease 2019 pandemic. METHODS: A prospective randomized open-label study was designed to identify differences in the fitting performance between mask wearing methods in three different types of KF94 mask with ear loops between January to March 2021. General-fitting involved wearing an ear-loop-type KF94 mask, and tight-fitting involved wearing a mask aided by a clip connecting the ear loops. Each of the 30 participants wore three types of masks according to a randomly assigned order in both methods and performed a total of six quantitative fit tests (QNFTs) according to the occupational safety and health administration protocol. RESULTS: All fit factors (FFs) measured by the QNFT were significantly higher for tight-fitting method with the clip in all KF94 masks (P < 0.001). However, the total FFs were very low, with a median (interquartile range) of 6 (3-23) and 29 (9-116) for general-fitting and tight-fitting, respectively. When wearing tightly, the horizontal 3-fold type mask with adjustable ear-loop length had the highest FF, with a median of 125, and the QNFT pass rate (FF ≥ 100) increased significantly from 4 (13%) to 18 (60%). CONCLUSION: Even with sufficient filter efficiency, ear-loop-type-KF94 masks do not provide adequate protection. However, in relatively low-risk environments, wearing a face-seal adjustable KF94 mask and tight wearing with a clip can improve respiratory protection for healthcare workers. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT04794556.