Evaluation of lung involvement in COVID-19 pneumonia based on ultrasound images

• Published in: Biomedical engineering online Vol. 20; no. 1; pp. 27 - 15
• Main Authors: Hu, Zhaoyu, Liu, Zhenhua, Dong, Yijie, Liu, Jianjian, Huang, Bin, Liu, Aihua, Huang, Jingjing, Pu, Xujuan, Shi, Xia, Yu, Jinhua, Xiao, Yang, Zhang, Hui, Zhou, Jianqiao
• Format: Journal Article
• Language: English
• Published: London BioMed Central 20.03.2021; BioMed Central Ltd; Springer Nature B.V; BMC
• Subjects: Accuracy; Airway management; Algorithms; Bacterial pneumonia; Biomaterials; Biomedical Engineering and Bioengineering; Biomedical Engineering/Biotechnology; Biotechnology; Classification; Computer-aided medical diagnosis; Coronaviruses; COVID-19; COVID-19 - diagnostic imaging; Databases, Factual; Datasets; Deep learning; Diagnosis; Diagnosis, Ultrasonic; Disease; Engineering; Evaluation; False Positive Reactions; Humans; Image Processing, Computer-Assisted - methods; Lung - diagnostic imaging; Lung involvement; Lungs; Machine learning; Methods; Model accuracy; Model testing; Models, Statistical; Neural network; Neural Networks, Computer; Patients; Pneumonia; Pneumonia - diagnostic imaging; Programming Languages; Receptive field; Rehabilitation; Reproducibility of Results; Sensitivity and Specificity; Software; Sonograms; Ultrasonic imaging; Ultrasonography; Ultrasound; China; COVID-19; Lung involvement; Neural network; Ultrasound; Classification
• ISSN: 1475-925X; 1475-925X
• DOI: 10.1186/s12938-021-00863-x

Background Lung ultrasound (LUS) can be an important imaging tool for the diagnosis and assessment of lung involvement. Ultrasound sonograms have been confirmed to illustrate damage to a person’s lungs, which means that the correct classification and scoring of a patient’s sonogram can be used to assess lung involvement. Methods The purpose of this study was to establish a lung involvement assessment model based on deep learning. A novel multimodal channel and receptive field attention network combined with ResNeXt (MCRFNet) was proposed to classify sonograms, and the network can automatically fuse shallow features and determine the importance of different channels and respective fields. Finally, sonogram classes were transformed into scores to evaluate lung involvement from the initial diagnosis to rehabilitation. Results and conclusion Using multicenter and multimodal ultrasound data from 104 patients, the diagnostic model achieved 94.39% accuracy, 82.28% precision, 76.27% sensitivity, and 96.44% specificity. The lung involvement severity and the trend of COVID-19 pneumonia were evaluated quantitatively.