Ho, Thao Thi; Park, Jongmin; Kim, Taewoo; Park, Byunggeon; Lee, Jaehee; Kim, Jin Young; Kim, Ki Beom; Choi, Sooyoung; Kim, Young Hwan; Lim, Jae-Kwang; Choi, Sanghun
JMIR medical informatics
2021Jan ; 9 ( 1 ) :e24973.
PMID : 33455900
ÀúÀÚ »ó¼¼Á¤º¸
Ho, Thao Thi - School of Mechanical Engineering, Kyungpook National University, Daegu, Republic of Korea.
Park, Jongmin - Department of Radiology, School of Medicine, Kyungpook National University, Daegu, Republic of Korea.
Kim, Taewoo - School of Mechanical Engineering, Kyungpook National University, Daegu, Republic of Korea.
Park, Byunggeon - Department of Radiology, School of Medicine, Kyungpook National University, Daegu, Republic of Korea.
Lee, Jaehee - Department of Internal Medicine, School of Medicine, Kyungpook National University, Daegu, Republic of Korea.
Kim, Jin Young - Department of Radiology, Keimyung University Dongsan Hospital, Daegu, Republic of Korea.
Kim, Ki Beom - Department of Radiology, Daegu Fatima Hospital, Daegu, Republic of Korea.
Choi, Sooyoung - Department of Radiology, Yeungnam University Medical Center, Daegu, Republic of Korea.
Kim, Young Hwan - Department of Radiology, School of Medicine, Daegu Catholic University, Daegu, Republic of Korea.
Lim, Jae-Kwang - Department of Radiology, School of Medicine, Kyungpook National University, Daegu, Republic of Korea.
Choi, Sanghun - School of Mechanical Engineering, Kyungpook National University, Daegu, Republic of Korea.
ABSTRACT
BACKGROUND: Many COVID-19 patients rapidly progress to respiratory failure with a broad range of severities. Identification of high-risk cases is critical for early intervention. OBJECTIVE: The aim of this study is to develop deep learning models that can rapidly identify high-risk COVID-19 patients based on computed tomography (CT) images and clinical data.
METHODS: We analyzed 297 COVID-19 patients from five hospitals in Daegu, South Korea. A mixed artificial convolutional neural network (ACNN) model, combining an artificial neural network for clinical data and a convolutional neural network for 3D CT imaging data, was developed to classify these cases as either high risk of severe progression (ie, event) or low risk (ie, event-free).
RESULTS: Using the mixed ACNN model, we were able to obtain high classification performance using novel coronavirus pneumonia lesion images (ie, 93.9% accuracy, 80.8% sensitivity, 96.9% specificity, and 0.916 area under the curve [AUC] score) and lung segmentation images (ie, 94.3% accuracy, 74.7% sensitivity, 95.9% specificity, and 0.928 AUC score) for event versus event-free groups.
CONCLUSIONS: Our study successfully differentiated high-risk cases among COVID-19 patients using imaging and clinical features. The developed model can be used as a predictive tool for interventions in aggressive therapies. CI - ?’Thao Thi Ho, Jongmin Park, Taewoo Kim, Byunggeon Park, Jaehee Lee, Jin Young Kim, Ki Beom Kim, Sooyoung Choi, Young Hwan Kim, Jae-Kwang Lim, Sanghun Choi. Originally published in JMIR Medical Informatics (http://medinform.jmir.org), 28.01.2021.
keyword
COVID-19; artificial neural network; convolutional neural network; deep learning; lung CT