SCIENCE CHINA Information Sciences, Volume 63 , Issue 7 : 172103(2020) https://doi.org/10.1007/s11432-020-2849-3

CT radiomics can help screen the coronavirus disease 2019 (COVID-19): a preliminary study

More info
  • ReceivedMar 1, 2020
  • AcceptedMar 23, 2020
  • PublishedApr 15, 2020


The coronavirus disease 2019 (COVID-19) is raging across the world. The radiomics, which explores huge amounts of features from medical image for disease diagnosis, may help the screen of the COVID-19. In this study, we aim to develop a radiomic signature to screen COVID-19 from CT images. We retrospectively collect 75 pneumonia patients from Beijing Youan Hospital, including 46 patients with COVID-19 and 29 other types of pneumonias. These patients are divided into training set ($n=50$) and test set ($n=25$) at random. We segment the lung lesions from the CT images, and extract 77 radiomic features from the lesions. Then unsupervised consensus clustering and multiple cross-validation are utilized to select the key features that are associated with the COVID-19. In the experiments, while twenty-three radiomic features are found to be highly associated with COVID-19, four key features are screened and used as the inputs of support vector machine to build the radiomic signature. We use area under the receiver operating characteristic curve (AUC) and calibration curve to assess the performance of our model. It yields AUCs of 0.862 and 0.826 in the training set and the test set respectively. We also perform the stratified analysis and find that its predictive ability is not affected by gender, age, chronic disease and degree of severity. In conclusion, we investigate the value of radiomics in screening COVID-19, and the experimental results suggest the radiomic signature could be a potential tool for diagnosis of COVID-19.


This work was supported by National Key RD Program of China (Grant Nos. 2017YFC1308700, 2017YFA0205200, 2017YFC1309100, 2017YFA0700401), National Natural Science Foundation of China (Grant Nos. 819300- 53, 91959130, 81971776, 81771924, 81930053, 81227901).


[1] Chan J F W, Yuan S, Kok K H. A familial cluster of pneumonia associated with the 2019 novel coronavirus indicating person-to-person transmission: a study of a family cluster. Lancet, 2020, 395: 514-523 CrossRef Google Scholar

[2] Zhu N, Zhang D, Wang W. A Novel Coronavirus from Patients with Pneumonia in China, 2019.. N Engl J Med, 2020, 382: 727-733 CrossRef PubMed Google Scholar

[3] Paules C I, Marston H D, Fauci A S. Coronavirus Infections-More Than Just the Common Cold.. JAMA, 2020, 323: 707-708 CrossRef PubMed Google Scholar

[4] Holshue M L, DeBolt C, Lindquist S. First Case of 2019 Novel Coronavirus in the United States.. N Engl J Med, 2020, 382: 929-936 CrossRef PubMed Google Scholar

[5] Chen N, Zhou M, Dong X. Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study. Lancet, 2020, 395: 507-513 CrossRef Google Scholar

[6] Yang X, Yu Y, Xu J. Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study. Lancet Respiratory Med, 2020, CrossRef Google Scholar

[7] Rodriguez-Morales AJ, Cardona-Ospina JA, Gutierrez-Ocampo E, et al. Clinical, laboratory and imaging features of COVID-19: a systematic review and meta-analysis. Preprints, 2020, 2020020378, DOI: 10.20944/preprints202002.0378.v1. Google Scholar

[8] Wang D, Hu B, Hu C. Clinical Characteristics of 138 Hospitalized Patients With 2019 Novel Coronavirus-Infected Pneumonia in Wuhan, China.. JAMA, 2020, 323: 1061 CrossRef PubMed Google Scholar

[9] Huang C, Wang Y, Li X. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. Lancet, 2020, 395: 497-506 CrossRef Google Scholar

[10] Corman V M, Landt O, Kaiser M. Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-PCR.. 349 CrossRef PubMed Google Scholar

[11] Lu R, Zhao X, Li J. Genomic characterisation and epidemiology of 2019 novel coronavirus: implications for virus origins and receptor binding. Lancet, 2020, 395: 565-574 CrossRef Google Scholar

[12] Lee E Y P, Ng M Y, Khong P L. COVID-19 pneumonia: what has CT taught us?. Lancet Infect Dis, 2020, 20: 384-385 CrossRef Google Scholar

[13] Pan F, Ye T, Sun P. Time Course of Lung Changes On Chest CT During Recovery From 2019 Novel Coronavirus (COVID-19) Pneumonia.. Radiology, 2020, : 200370 CrossRef PubMed Google Scholar

[14] Xu X, Yu C, Zhang L. Imaging features of 2019 novel coronavirus pneumonia.. Eur J Nucl Med Mol Imag, 2020, 47: 1022-1023 CrossRef PubMed Google Scholar

[15] Fang Y, Zhang H, Xu Y. CT Manifestations of Two Cases of 2019 Novel Coronavirus (2019-nCoV) Pneumonia.. Radiology, 2020, 295: 208-209 CrossRef PubMed Google Scholar

[16] Shi H, Han X, Jiang N. Radiological findings from 81 patients with COVID-19 pneumonia in Wuhan, China: a descriptive study. Lancet Infect Dis, 2020, 20: 425-434 CrossRef Google Scholar

[17] Dong D, Zhang F, Zhong L Z. Development and validation of a novel MR imaging predictor of response to induction chemotherapy in locoregionally advanced nasopharyngeal cancer: a randomized controlled trial substudy (NCT01245959).. BMC Med, 2019, 17: 190 CrossRef PubMed Google Scholar

[18] Lambin P, Rios-Velazquez E, Leijenaar R. Radiomics: extracting more information from medical images using advanced feature analysis.. Eur J Cancer, 2012, 48: 441-446 CrossRef PubMed Google Scholar

[19] Dong D, Tang L, Li Z Y. Development and validation of an individualized nomogram to identify occult peritoneal metastasis in patients with advanced gastric cancer. Ann Oncology, 2019, 30: 431-438 CrossRef Google Scholar

[20] Bi W L, Hosny A, Schabath M B. Artificial intelligence in cancer imaging: Clinical challenges and applications.. CA A Cancer J Clin, 2019, 16: caac.21552 CrossRef PubMed Google Scholar

[21] Song J, Shi J, Dong D. A New Approach to Predict Progression-free Survival in Stage IV EGFR-mutant NSCLC Patients with EGFR-TKI Therapy.. Clin Cancer Res, 2018, 24: 3583-3592 CrossRef PubMed Google Scholar

[22] Zhang L, Chen B, Liu X. Quantitative Biomarkers for Prediction of Epidermal Growth Factor Receptor Mutation in Non-Small Cell Lung Cancer.. Translational Oncology, 2018, 11: 94-101 CrossRef PubMed Google Scholar

[23] Peng H, Dong D, Fang M J. Prognostic Value of Deep Learning PET/CT-Based Radiomics: Potential Role for Future Individual Induction Chemotherapy in Advanced Nasopharyngeal Carcinoma.. Clin Cancer Res, 2019, 25: 4271-4279 CrossRef PubMed Google Scholar

[24] Cheng S, Fang M, Cui C. LGE-CMR-derived texture features reflect poor prognosis in hypertrophic cardiomyopathy patients with systolic dysfunction: preliminary results.. Eur Radiol, 2018, 28: 4615-4624 CrossRef PubMed Google Scholar

[25] Kolossváry M, Karády J, Szilveszter B. Radiomic Features Are Superior to Conventional Quantitative Computed Tomographic Metrics to Identify Coronary Plaques With Napkin-Ring Sign.. Circ Cardiovasc Imag, 2017, 10 CrossRef PubMed Google Scholar

[26] Wang B, Li M, Ma H. Computed tomography-based predictive nomogram for differentiating primary progressive pulmonary tuberculosis from community-acquired pneumonia in children.. BMC Med Imag, 2019, 19: 63 CrossRef PubMed Google Scholar

  • Figure 1

    (Color online) Radiomic heatmap on the overall set. Unsupervised clustering of patients ($n=75$) and radiomics features ($n=77$) reveal clusters of patients with similar radiomic expression patterns.

  • Figure 2

    Results of consensus clustering analysis for radiomic features. The curve depicts the minimum Spearman correlation coefficient between the medoid features with their intra-cluster features.

  • Figure 3

    (Color online) Bar plot of the radiomic signature scores for patients with COVID-19 and other pneumonias types in the (a) training and (b) test sets.

  • Figure 4

    (Color online) ROC curves (a) and calibration curves (b) of the radiomic signature in each set.

  • Figure 5

    (Color online) ROC curves of radiomic signature for each subgroup stratified by gender (a), age (b), with/without chronic disease (c) and degree of severity (d).

  • Table 1   Characteristics of patients in the training and test sets
    Characteristics Total Training set Test set
    Age (mean$\pm$SD, years) 47.8$\pm$20.2 46.2$\pm$20.3 51.1$\pm$19.9
    Gender (male/female, No.) 40/35 26/24 14/11
    Pneumonias type (COVID-19/other types, No.) 46/29 30/20 16/9
  • Table 2   Features and coefficients of the radiomic signature$^{\rm~a)}$
    Name Group Data AUC
    S_GLRLM_GLN GLRLM feature S 0.633
    O_GLCM_Contrast GLCM feature O 0.577
    S_I_Mean Intensity-based statistical feature S 0.552
    O_I_Krutosis Intensity-based statistical feature O 0.508
    S_I_Range Intensity-based statistical feature S 0.605
    O_GLCM_Variance GLCM feature O 0.728
    O_I_Minimum Intensity-based statistical feature O 0.538
    S_GLRLM_LGLRE GLRLM feature S 0.558
    S_I_Krutosis Intensity-based statistical feature S 0.607
    S_I_Skewness Intensity-based statistical feature S 0.582
    O_GLCM_Energy GLCM feature O 0.628
    O_GLRLM_LRLGLE GLRLM feature O 0.487
    S_GLCM_Cluster_shade GLCM feature S 0.605
    O_GLRLM_RP GLRLM feature O 0.490
    O_GLRLM_SRLGLE GLRLM feature O 0.528
    S_GLRLM_SRLGLE GLRLM feature S 0.518
    S_GLRLM_LRHGLE GLRLM feature S 0.513


Copyright 2020  CHINA SCIENCE PUBLISHING & MEDIA LTD.  中国科技出版传媒股份有限公司  版权所有

京ICP备14028887号-23       京公网安备11010102003388号