Jiaxing Sun1,2, Ximing Liao1, Yusheng Yan3, Xin Zhang4, Jian Sun5, Weixiong Tan6, Baiyun Liu6, Jiangfen Wu6, Qian Guo1, Shaoyong Gao1, Zhang Li7, Kun Wang8, Qiang Li9. 1. Department of Pulmonary and Critical Care Medicine, Shanghai East Hospital, Tongji University School of Medicine, No. 150 Jimo Road, Pudong, Shanghai, China. 2. Department of Pulmonary and Critical Care Medicine, The Affiliated Hospital of Qingdao University, Qingdao, China. 3. Department of Pulmonary and Critical Care Medicine, Changsha First Hospital, Changsha, China. 4. Department of Pulmonary and Critical Care Medicine, People's Liberation Army Joint Logistic Support Force 920th Hospital, Kunming, Yunnan, China. 5. Department of Pulmonary and Critical Care Medicine, Shandong Provincial Hospital, Jinan, China. 6. Infervision, Beijing, China. 7. College of Aerospace Science and Engineering, National University of Defense Technology, Changsha, China. 8. Department of Pulmonary and Critical Care Medicine, Shanghai East Hospital, Tongji University School of Medicine, No. 150 Jimo Road, Pudong, Shanghai, China. Dr_Wangk@tongji.edu.cn. 9. Department of Pulmonary and Critical Care Medicine, Shanghai East Hospital, Tongji University School of Medicine, No. 150 Jimo Road, Pudong, Shanghai, China. liqressh1962@163.com.
Abstract
OBJECTIVES: Chronic obstructive pulmonary disease (COPD) is underdiagnosed globally. The present study aimed to develop weakly supervised deep learning (DL) models that utilize computed tomography (CT) image data for the automated detection and staging of spirometry-defined COPD. METHODS: A large, highly heterogeneous dataset was established, consisting of 1393 participants retrospectively recruited from outpatient, inpatient, and physical examination center settings of four large public hospitals in China. All participants underwent both inspiratory chest CT scans and pulmonary function tests. CT images, spirometry data, demographic information, and clinical information of each participant were collected. An attention-based multi-instance learning (MIL) model for COPD detection was trained using CT scans from 837 participants. External validation of the COPD detection was performed with 620 low-dose CT (LDCT) scans acquired from the National Lung Screening Trial (NLST) cohort. A multi-channel 3D residual network was further developed to categorize GOLD stages among confirmed COPD patients. RESULTS: The attention-based MIL model used for COPD detection achieved an area under the receiver operating characteristic curve (AUC) of 0.934 (95% CI: 0.903, 0.961) on the internal test set and 0.866 (95% CI: 0.805, 0.928) on the LDCT subset acquired from the NLST. The multi-channel 3D residual network was able to correctly grade 76.4% of COPD patients in the test set (423/553) using the GOLD scale. CONCLUSIONS: The proposed chest CT-DL approach can automatically identify spirometry-defined COPD and categorize patients according to the GOLD scale. As such, this approach may be an effective case-finding tool for COPD diagnosis and staging. KEY POINTS: • Chronic obstructive pulmonary disease is underdiagnosed globally, particularly in developing countries. • The proposed chest computed tomography (CT)-based deep learning (DL) approaches could accurately identify spirometry-defined COPD and categorize patients according to the GOLD scale. • The chest CT-DL approach may be an alternative case-finding tool for COPD identification and evaluation.
OBJECTIVES: Chronic obstructive pulmonary disease (COPD) is underdiagnosed globally. The present study aimed to develop weakly supervised deep learning (DL) models that utilize computed tomography (CT) image data for the automated detection and staging of spirometry-defined COPD. METHODS: A large, highly heterogeneous dataset was established, consisting of 1393 participants retrospectively recruited from outpatient, inpatient, and physical examination center settings of four large public hospitals in China. All participants underwent both inspiratory chest CT scans and pulmonary function tests. CT images, spirometry data, demographic information, and clinical information of each participant were collected. An attention-based multi-instance learning (MIL) model for COPD detection was trained using CT scans from 837 participants. External validation of the COPD detection was performed with 620 low-dose CT (LDCT) scans acquired from the National Lung Screening Trial (NLST) cohort. A multi-channel 3D residual network was further developed to categorize GOLD stages among confirmed COPD patients. RESULTS: The attention-based MIL model used for COPD detection achieved an area under the receiver operating characteristic curve (AUC) of 0.934 (95% CI: 0.903, 0.961) on the internal test set and 0.866 (95% CI: 0.805, 0.928) on the LDCT subset acquired from the NLST. The multi-channel 3D residual network was able to correctly grade 76.4% of COPD patients in the test set (423/553) using the GOLD scale. CONCLUSIONS: The proposed chest CT-DL approach can automatically identify spirometry-defined COPD and categorize patients according to the GOLD scale. As such, this approach may be an effective case-finding tool for COPD diagnosis and staging. KEY POINTS: • Chronic obstructive pulmonary disease is underdiagnosed globally, particularly in developing countries. • The proposed chest computed tomography (CT)-based deep learning (DL) approaches could accurately identify spirometry-defined COPD and categorize patients according to the GOLD scale. • The chest CT-DL approach may be an alternative case-finding tool for COPD identification and evaluation.
Authors: Kendra A Young; Matthew Strand; Margaret F Ragland; Gregory L Kinney; Erin E Austin; Elizabeth A Regan; Katherine E Lowe; Barry J Make; Edwin K Silverman; James D Crapo; John E Hokanson Journal: Chronic Obstr Pulm Dis Date: 2019-11
Authors: Katherine E Lowe; Elizabeth A Regan; Antonio Anzueto; Erin Austin; John H M Austin; Terri H Beaty; Panayiotis V Benos; Christopher J Benway; Surya P Bhatt; Eugene R Bleecker; Sandeep Bodduluri; Jessica Bon; Aladin M Boriek; Adel Re Boueiz; Russell P Bowler; Matthew Budoff; Richard Casaburi; Peter J Castaldi; Jean-Paul Charbonnier; Michael H Cho; Alejandro Comellas; Douglas Conrad; Corinne Costa Davis; Gerard J Criner; Douglas Curran-Everett; Jeffrey L Curtis; Dawn L DeMeo; Alejandro A Diaz; Mark T Dransfield; Jennifer G Dy; Ashraf Fawzy; Margaret Fleming; Eric L Flenaugh; Marilyn G Foreman; Spyridon Fortis; Hirut Gebrekristos; Sarah Grant; Philippe A Grenier; Tian Gu; Abhya Gupta; MeiLan K Han; Nicola A Hanania; Nadia N Hansel; Lystra P Hayden; Craig P Hersh; Brian D Hobbs; Eric A Hoffman; James C Hogg; John E Hokanson; Karin F Hoth; Albert Hsiao; Stephen Humphries; Kathleen Jacobs; Francine L Jacobson; Ella A Kazerooni; Victor Kim; Woo Jin Kim; Gregory L Kinney; Harald Koegler; Sharon M Lutz; David A Lynch; Neil R MacIntye; Barry J Make; Nathaniel Marchetti; Fernando J Martinez; Diego J Maselli; Anne M Mathews; Meredith C McCormack; Merry-Lynn N McDonald; Charlene E McEvoy; Matthew Moll; Sarah S Molye; Susan Murray; Hrudaya Nath; John D Newell; Mariaelena Occhipinti; Matteo Paoletti; Trisha Parekh; Massimo Pistolesi; Katherine A Pratte; Nirupama Putcha; Margaret Ragland; Joseph M Reinhardt; Stephen I Rennard; Richard A Rosiello; James C Ross; Harry B Rossiter; Ingo Ruczinski; Raul San Jose Estepar; Frank C Sciurba; Jessica C Sieren; Harjinder Singh; Xavier Soler; Robert M Steiner; Matthew J Strand; William W Stringer; Ruth Tal-Singer; Byron Thomashow; Gonzalo Vegas Sánchez-Ferrero; John W Walsh; Emily S Wan; George R Washko; J Michael Wells; Chris H Wendt; Gloria Westney; Ava Wilson; Robert A Wise; Andrew Yen; Kendra Young; Jeong Yun; Edwin K Silverman; James D Crapo Journal: Chronic Obstr Pulm Dis Date: 2019-11