2019 Novel Coronavirus-Infected Pneumonia on CT: A Feasibility Study of Few-Shot Learning for Computerized Diagnosis of Emergency Diseases

Yaoming Lai; Guangming Li; Dongmei Wu; Wanmin Lian; Cheng Li; Junzhang Tian; Xiaofen Ma; Hui Chen; Wen Xu; Jun Wei; Yaqin Zhang; Guihua Jiang

doi:10.1109/ACCESS.2020.3033069

2019 Novel Coronavirus-Infected Pneumonia on CT: A Feasibility Study of Few-Shot Learning for Computerized Diagnosis of Emergency Diseases

IEEE Access. 2020 Oct 22:8:194158-194165. doi: 10.1109/ACCESS.2020.3033069. eCollection 2020.

Authors

Yaoming Lai¹, Guangming Li², Dongmei Wu³, Wanmin Lian⁴, Cheng Li⁴, Junzhang Tian⁴, Xiaofen Ma⁴, Hui Chen², Wen Xu², Jun Wei¹, Yaqin Zhang⁵, Guihua Jiang⁴

Affiliations

¹ Guangzhou Perception Vision Medical Technology Inc. Guangzhou 510000 China.
² Xiangyang Central HospitalAffiliated Hospital of Hubei University of Arts and Science Xiangyang 441000 China.
³ Nanxishan Hospital of Guangxi Zhuang Autonomous Region Guilin 541000 China.
⁴ Department of Medical ImagingGuangdong Second Provincial General Hospital Guangzhou 510000 China.
⁵ Department of RadiologyThe Fifth Affiliated HospitalSun Yat-sen University Zhuhai 519000 China.

Abstract

COVID-19 is an emerging disease with transmissibility and severity. So far, there are no effective therapeutic drugs or vaccines for COVID-19. The most serious complication of COVID-19 is a type of pneumonia called 2019 novel coronavirus-infected pneumonia (NCIP) with about 4.3% mortality rate. Comparing to chest Digital Radiography (DR), it is recently reported that chest Computed Tomography (CT) is more useful to serve as the early screening and diagnosis tool for NCIP. In this study, aimed to help physicians make the diagnostic decision, we develop a machine learning (ML) approach for automated diagnosis of NCIP on chest CT. Different from most ML approaches which often require training on thousands or millions of samples, we design a few-shot learning approach, in which we combine few-shot learning with weakly supervised model training, for computerized NCIP diagnosis. A total of 824 patients are retrospectively collected from two Hospitals with IRB approval. We first use 9 patients with clinically confirmed NCIP and 20 patients without known lung diseases for training a location detector which is a multitask deep convolutional neural network (DCNN) designed to output a probability of NCIP and the segmentation of targeted lesion area. An experienced radiologist manually localizes the potential locations of NCIPs on chest CTs of 9 COVID-19 patients and interactively segments the area of the NCIP lesions as the reference standard. Then, the multitask DCNN is furtherly fine-tuned by a weakly supervised learning scheme with 291 case-level labeled samples without lesion labels. A test set of 293 patients is independently collected for evaluation. With our NCIP-Net, the test AUC is 0.91. Our system has potential to serve as the NCIP screening and diagnosis tools for the fight of COVID-19's endemic and pandemic.

Keywords: 2019 novel coronavirus-infected pneumonia; COVID-19; chest CT; few-shot learning.

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.

Grants and funding

This work was supported in part by the National Natural Science Foundation of China under Grant 81860523, Grant 81801809; in part by Guangdong Basic and Applied Basic Research Foundation under Grant 2020A1515010572, in part by the Guangzhou Red Cotton Planning Venture Project under Key Grant HMJH-2019-0010, in part by Guangzhou Economic and Technological Development Zone Science and Technology Innovation Leader Project, in part by the Guangzhou Science and Technology Planning Project under Key Grant 201906010014, and in part by the Hospital Basic Foundation of Nanxishan Hospital of Guangxi Zhuang Autonomous Region.