Automatic Classification of Centrilobular Emphysema on CT Using Deep Learning: Comparison with Visual Scoring

  • Stephen M. HumphriesEmail author
  • Aleena M. Notary
  • Juan Pablo Centeno
  • David A. Lynch
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11040)


The presence and severity of emphysema, scored visually on computed tomography (CT) using a classification system developed by the Fleischner Society, is a clinically significant index of disease severity. Since visual assessment can be subjective and is time consuming, our purpose was to evaluate the potential of a deep learning method for automatic grading of emphysema. The study cohort included 8213 subjects enrolled in the COPDGene study. Baseline CT and visual scores on 2500 subjects were used to train a deep learning model for classification of centrilobular emphysema according to the Fleischner system. The model was then used to predict emphysema scores on 5713 subjects not included in the training set. Predictions were compared with visual emphysema scores, pulmonary function tests (PFTs), smoking history and St. George Respiratory Questionnaire (SGRQ). Agreement between visual emphysema scores and those generated automatically was moderate (weighted \(\kappa \) = 0.60, p < 0.0001). Emphysema scores predicted by the deep learning model showed significant associations with PFTs, smoking history and SGRQ, similar to those seen in comparison with visual scores.


Computed tomography Emphysema Deep learning 


  1. 1.
    Donahue, J., et al.: Long-term recurrent convolutional networks for visual recognition and description. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 2625–2634 (2015)Google Scholar
  2. 2.
    Esteva, A., et al.: Dermatologist-level classification of skin cancer with deep neural networks. Nature 542(7639), 115 (2017)CrossRefGoogle Scholar
  3. 3.
    González, G., et al.: Disease staging and prognosis in smokers using deep learning in chest computed tomography. Am. J. Respir. Crit. Care Med. 197(2), 193–203 (2018)CrossRefGoogle Scholar
  4. 4.
    González, G., Washko, G.R., San José Estépar, R.a.: Deep learning for biomarker regression: application to osteoporosis and emphysema on chest CT scans. In: SPIE Medical Imaging, vol. 10574 (2018)Google Scholar
  5. 5.
    Halper-Stromberg, E., et al.: Visual assessment of chest computed tomographic images is independently useful for genetic association analysis in studies of chronic obstructive pulmonary disease. Ann. Am. Thorac. Soc. 14(1), 33–40 (2017)CrossRefGoogle Scholar
  6. 6.
    Lynch, D.A., et al.: CT-definable subtypes of chronic obstructive pulmonary disease: a statement of the fleischner society. Radiology 277(1), 192–205 (2015)CrossRefGoogle Scholar
  7. 7.
    Lynch, D.A.: CT-based visual classification of emphysema: association with mortality in the COPDGene study. Radiology 288, 859–866 (2018)CrossRefGoogle Scholar
  8. 8.
    Paszke, A., et al.: Automatic Differentiation in PyTorch (2017)Google Scholar
  9. 9.
    Regan, E.A., et al.: Genetic epidemiology of COPD (COPDGene) study design. COPD J. Chronic Obstructive Pulm. Dis. 7(1), 32–43 (2011)CrossRefGoogle Scholar
  10. 10.
    Ronneberger, O., Fischer, P., Brox, T.: U-Net: convolutional networks for biomedical image segmentation. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) MICCAI 2015. LNCS, vol. 9351, pp. 234–241. Springer, Cham (2015). Scholar
  11. 11.
    Ross, J.C., et al.: Lung extraction, lobe segmentation and hierarchical region assessment for quantitative analysis on high resolution computed tomography images. In: Yang, G.-Z., Hawkes, D., Rueckert, D., Noble, A., Taylor, C. (eds.) MICCAI 2009. LNCS, vol. 5762, pp. 690–698. Springer, Heidelberg (2009). Scholar
  12. 12.
    Ross, J., Harmouche, R., Onieva, J., Diaz, A., Washko, G., Estepar, R.S.J.: Chest imaging platform: an open-source library and workstation for quantitative chest imaging. Am. J. Respir. Crit. Care Med. 191, A4975 (2015)Google Scholar
  13. 13.
    Schroeder, J.D., et al.: Relationships between airflow obstruction and quantitative CT measurements of emphysema, air trapping, and airways in subjects with and without chronic obstructive pulmonary disease. Am. J. Roentgenol. 201(3), W460–W470 (2013)CrossRefGoogle Scholar
  14. 14.
    Setio, A.A.A., et al.: Pulmonary nodule detection in CT images: false positive reduction using multi-view convolutional networks. IEEE Trans. Med. Imaging 35(5), 1160–1169 (2016)CrossRefGoogle Scholar
  15. 15.
    Vestbo, J.: Evaluation of COPD longitudinally to identify predictive surrogate endpoints (ECLIPSE). Eur. Respir. J. 31, 869–873 (2008)CrossRefGoogle Scholar
  16. 16.
    Zhou, S.K., Greenspan, H., Shen, D.: Deep Learning for Medical Image Analysis. Academic Press, San Diego (2017)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  • Stephen M. Humphries
    • 1
    Email author
  • Aleena M. Notary
    • 1
  • Juan Pablo Centeno
    • 1
  • David A. Lynch
    • 1
  1. 1.National Jewish HealthDenverUSA

Personalised recommendations