Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

A shallow convolutional neural network predicts prognosis of lung cancer patients in multi-institutional computed tomography image datasets

Abstract

Lung cancer is the most common fatal malignancy in adults worldwide, and non-small-cell lung cancer (NSCLC) accounts for 85% of lung cancer diagnoses. Computed tomography is routinely used in clinical practice to determine lung cancer treatment and assess prognosis. Here, we developed LungNet, a shallow convolutional neural network for predicting outcomes of patients with NSCLC. We trained and evaluated LungNet on four independent cohorts of patients with NSCLC from four medical centres: Stanford Hospital (n = 129), H. Lee Moffitt Cancer Center and Research Institute (n = 185), MAASTRO Clinic (n = 311) and Charité – Universitätsmedizin, Berlin (n = 84). We show that outcomes from LungNet are predictive of overall survival in all four independent survival cohorts as measured by concordance indices of 0.62, 0.62, 0.62 and 0.58 on cohorts 1, 2, 3 and 4, respectively. Furthermore, the survival model can be used, via transfer learning, for classifying benign versus malignant nodules on the Lung Image Database Consortium (n = 1,010), with improved performance (AUC = 0.85) versus training from scratch (AUC = 0.82). LungNet can be used as a non-invasive predictor for prognosis in patients with NSCLC and can facilitate interpretation of computed tomography images for lung cancer stratification and prognostication.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Fig. 1: Illustration of LungNet’s CNN architecture.
Fig. 2: Illustration of the proposed computational framework.
Fig. 3: Kaplan–Meier analysis of LungNet.
Fig. 4: Kaplan–Meier survival performance of LungNet on early-stage cancers.
Fig. 5: Transfer learning for malignancy prediction.
Fig. 6: Visualization of lung nodules and their survival outcomes in 2D space using t-SNE.

Data availability

The data for cohort 1 (Stanford Hospital, n = 129) are publicly available on The Cancer Imaging Archive (TCIA) at https://doi.org/10.7937/K9/TCIA.2017.7hs46erv (ref. 64). A portion of the data (54/185) for cohort 2 (H. Lee Moffitt Cancer Center and Research Institute, n = 185) is available from TCIA at https://doi.org/10.7937/K9/TCIA.2015.NPGZYZBZ (ref. 65) and https://doi.org/10.7937/K9/TCIA.2015.A6V7JIWX (ref. 66). The data for cohort 3 (MAASTRO Clinic, the Netherlands, n = 311) are publicly available on TCIA at https://doi.org/10.7937/K9/TCIA.2015.PF0M9REI (ref. 11). The data for cohort 4 (Charité – Universitätsmedizin, Berlin, n = 84) are not publicly available yet. The data for LIDC–IDRI (n = 1,010) are available on TCIA at https://doi.org/10.7937/K9/TCIA.2015.LO9QL9SX (ref. 67).

Code availability

Code for LungNet is available at https://doi.org/10.24433/CO.0612256.v1 (ref. 68).

References

  1. 1.

    Ferlay, J. et al. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int. J. Cancer 136, E359–E386 (2015).

    Article  Google Scholar 

  2. 2.

    Hirsch, F. R. et al. Lung cancer: current therapies and new targeted treatments. Lancet 389, 299–311 (2017).

    Article  Google Scholar 

  3. 3.

    Swensen, S. J. et al. CT screening for lung cancer: five-year prospective experience. Radiology 235, 259–265 (2005).

    Article  Google Scholar 

  4. 4.

    Swensen, S. J. et al. Lung cancer screening with CT: Mayo Clinic experience. Radiology 226, 756–761 (2003).

    Article  Google Scholar 

  5. 5.

    McWilliams, A. et al. Probability of cancer in pulmonary nodules detected on first screening CT. N. Engl. J. Med. 369, 910–919 (2013).

    Article  Google Scholar 

  6. 6.

    Henschke, C. I. et al. Early lung cancer action project: overall design and findings from baseline screening. Lancet 354, 99–105 (1999).

    Article  Google Scholar 

  7. 7.

    Gillies, R. J., Kinahan, P. E. & Hricak, H. Radiomics: images are more than pictures, they are data. Radiology 278, 563–577 (2015).

    Article  Google Scholar 

  8. 8.

    Lambin, P. et al. Radiomics: the bridge between medical imaging and personalized medicine. Nat. Rev. Clin. Oncol. 14, 749–762 (2017).

    Article  Google Scholar 

  9. 9.

    Thawani, R. et al. Radiomics and radiogenomics in lung cancer: a review for the clinician. Lung Cancer 115, 34–41 (2018).

    Article  Google Scholar 

  10. 10.

    Zhou, M. et al. Non–small cell lung cancer radiogenomics map identifies relationships between molecular and imaging phenotypes with prognostic implications. Radiology (2017).

  11. 11.

    Aerts, H. J. W. L. et al. Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nat. Commun. 5, 4006 (2014).

    Article  Google Scholar 

  12. 12.

    Shen, C. et al. 2D and 3D CT radiomics features prognostic performance comparison in non-small cell lung cancer. Transl. Oncol. 10, 886–894 (2017).

    Article  Google Scholar 

  13. 13.

    Mattonen, S. A. et al. [18F] FDG positron emission tomography (PET) tumor and penumbra imaging features predict recurrence in non-small cell lung cancer. Tomography 5, 145–153 (2019).

    Article  Google Scholar 

  14. 14.

    Napel, S., Mu, W., Jardim-Perassi, B. V., Aerts, H. J. W. L. & Gillies, R. J. Quantitative imaging of cancer in the postgenomic era: radio(geno)mics, deep learning, and habitats. Cancer 124, 4633–4649 (2018).

    Article  Google Scholar 

  15. 15.

    Minamimoto, R. et al. Prediction of EGFR and KRAS mutation in non-small cell lung cancer using quantitative 18F FDG-PET/CT metrics. Oncotarget 8, 52792–52801 (2017).

    Article  Google Scholar 

  16. 16.

    Gevaert, O. et al. Predictive radiogenomics modeling of EGFR mutation status in lung cancer. Sci. Rep. 7, 41674 (2017).

    Article  Google Scholar 

  17. 17.

    van Griethuysen, J. J. M. et al. Computational radiomics system to decode the radiographic phenotype. Cancer Res. 77, e104–e107 (2017).

    Article  Google Scholar 

  18. 18.

    Aerts, H. J. W. L. Data science in radiology: a path forward. Clin. Cancer Res. 24, 532–534 (2018).

    Article  Google Scholar 

  19. 19.

    Hosny, A., Parmar, C., Quackenbush, J., Schwartz, L. H. & Aerts, H. J. W. L. Artificial intelligence in radiology. Nat. Rev. Cancer 18, 500–510 (2018).

    Article  Google Scholar 

  20. 20.

    Dehmeshki, J., Amin, H., Valdivieso, M. & Ye, X. Segmentation of pulmonary nodules in thoracic CT scans: a region growing approach. IEEE Trans. Med. Imaging 27, 467–80 (2008).

    Article  Google Scholar 

  21. 21.

    Lee, Y., Hara, T., Fujita, H., Itoh, S. & Ishigaki, T. Automated detection of pulmonary nodules in helical CT images based on an improved template-matching technique. IEEE Trans. Med. Imaging 20, 595–604 (2001).

    Article  Google Scholar 

  22. 22.

    Shen, W., Zhou, M., Yang, F., Yang, C. & Tian, J. in Information Processing in Medical Imaging (eds Ourselin S. et al.) (Springer, 2015).

  23. 23.

    Xu, Y. et al. Deep learning predicts lung cancer treatment response from serial medical imaging. Clin. Cancer Res. https://doi.org/10.1158/1078-0432.CCR-18-2495 (2019).

  24. 24.

    Krizhevsky, A., Sutskever, I. & Hinton, G. E. ImageNet classification with deep convolutional neural networks. In Commun. ACM 60, 84–90 (2017).

  25. 25.

    Esteva, A. et al. Dermatologist-level classification of skin cancer with deep neural networks. Nature 542, 115–118 (2017).

    Article  Google Scholar 

  26. 26.

    Bi, W. L. et al. Artificial intelligence in cancer imaging: clinical challenges and applications. CA Cancer J. Clin. 69, 127–157 (2019).

    Google Scholar 

  27. 27.

    Shin, H. C. et al. Deep convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and transfer learning. IEEE Trans. Med. Imaging 35, 1285–1298 (2016).

    Article  Google Scholar 

  28. 28.

    Ardila, D. et al. End-to-end lung cancer screening with three-dimensional deep learning on low-dose chest computed tomography. Nat. Med. 25, 954–961 (2019).

    Article  Google Scholar 

  29. 29.

    National Lung Screening Trial Research Team et al. The National Lung Screening Trial: overview and study design. Radiology 258, 243–253 (2011).

  30. 30.

    National Lung Screening Trial Research Team et al. Results of initial low-dose computed tomographic screening for lung cancer. N. Engl. J. Med. 368, 1980–1991 (2013).

  31. 31.

    Hanley, J. A. & McNeil, B. J. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 148, 839–843 (2014).

    Article  Google Scholar 

  32. 32.

    van der Maaten, L. & Hinton, G. Visualizing data using t-SNE. J. Mach. Learn. Res. 9, 2579–2605 (2008).

    MATH  Google Scholar 

  33. 33.

    Jamal-Hanjani, M. et al. Tracking the evolution of non–small-cell lung cancer. N. Engl. J. Med. 376, 2109–2121 (2017).

    Article  Google Scholar 

  34. 34.

    Parmar, C., Grossmann, P., Bussink, J., Lambin, P. & Aerts, H. J. W. L. Machine learning methods for quantitative radiomic biomarkers. Sci. Rep. 5, 13087 (2015).

    Article  Google Scholar 

  35. 35.

    Deng, J. et al. ImageNet: a large-scale hierarchical image database. In Proc. 2009 IEEE Conference on Computer Vision and Pattern Recognition 248–255 (IEEE, 2009).

  36. 36.

    Szegedy, C. et al. Going deeper with convolutions. In Proc. 2015 IEEE Conference on Computer Vision and Pattern Recognition 1–9 (IEEE, 2015).

  37. 37.

    He, K., Zhang, X., Ren S. & Sun, J. Deep residual learning for image recognition. In IEEE Conf. Computer Vision and Pattern Recognition 770–778 (IEEE, 2016).

  38. 38.

    Szegedy, C., Ioffe, S., Vanhoucke, V. & Alemi, A. A. Inception-v4, inception-ResNet and the impact of residual connections on learning. In Proc. 31st AAAI Conference on Artificial Intelligence 4278–4284 (AAAI, 2017).

  39. 39.

    Huang, G., Liu, Z., Van Der Maaten, L. & Weinberger, K. Q. Densely connected convolutional networks. In Proc. 2017 IEEE Conference on Computer Vision and Pattern Recognition 2261–2269 (IEEE, 2017).

  40. 40.

    Hara, K., Kataoka, H. & Satoh, Y. Can spatiotemporal 3D CNNs retrace the history of 2D CNNs and ImageNet? In Proc. 2018 IEEE Conference on Computer Vision and Pattern Recognition 6546–6555 (IEEE, 2018).

  41. 41.

    Raghu, M., Zhang, C., Kleinberg, J. & Bengio, S. Transfusion: understanding Transfer learning for medical imaging. In Advances in Neural Information Processing Systems Vol. 32 (eds H. Wallach et al.) (Curran Associates, 2019).

  42. 42.

    Causey, J. L. et al. Highly accurate model for prediction of lung nodule malignancy with CT scans. Sci. Rep. 8, 9286 (2018).

    Article  Google Scholar 

  43. 43.

    Wang, S. et al. Central focused convolutional neural networks: developing a data-driven model for lung nodule segmentation. Med. Image Anal. 40, 172–183 (2017).

    Article  Google Scholar 

  44. 44.

    Zhu, W., Liu, C., Fan, W. & Xie, X. DeepLung: deep 3D dual path nets for automated pulmonary nodule detection and classification. In Proc. 2018 IEEE Winter Conference on Applications of Computer Vision 673–681 (IEEE, 2018).

  45. 45.

    Shen, W. et al. Multi-crop convolutional neural networks for lung nodule malignancy suspiciousness classification. Pattern Recognit. 61, 663–673 (2017).

    Article  Google Scholar 

  46. 46.

    Cao, H. et al. Dual-branch residual network for lung nodule segmentation. Appl. Soft Comput. 86, 105934 (2020).

    Article  Google Scholar 

  47. 47.

    Liu, H. et al. A cascaded dual-pathway residual network for lung nodule segmentation in CT images. Phys. Med. 63, 112–121 (2019).

    Article  Google Scholar 

  48. 48.

    Hosny, A. et al. Deep learning for lung cancer prognostication: a retrospective multi-cohort radiomics study. PLoS Med. 15, e1002711 (2018).

    Article  Google Scholar 

  49. 49.

    Gentles, A. J. et al. Integrating tumor and stromal gene expression signatures with clinical indices for survival stratification of early-stage non-small cell lung cancer. J. Natl Cancer Inst. 107, djv211 (2015).

    Article  Google Scholar 

  50. 50.

    Liang, C. et al. Radiomics signature: a potential biomarker for the prediction of disease-free survival in early-stage (I or II) non—small cell lung cancer. Radiology 281, 947–957 (2016).

    Article  Google Scholar 

  51. 51.

    Shedden, K. et al. Gene expression-based survival prediction in lung adenocarcinoma: a multi-site, blinded validation study. Nat. Med. 14, 822–827 (2008).

    Article  Google Scholar 

  52. 52.

    Guo, N. L. et al. Confirmation of gene expression-based prediction of survival in non-small cell lung cancer. Clin. Cancer Res. 14, 8213–8220 (2008).

    Article  Google Scholar 

  53. 53.

    Armato, S. G. et al. The Lung Image Database Consortium (LIDC) and Image Database Resource Initiative (IDRI): A completed reference: database of lung nodules on CT scans. Med. Phys. 38, 915–931 (2011).

    Article  Google Scholar 

  54. 54.

    Cox, D. R. Regression models and life-tables. J. R. Stat. Soc. Ser. B 34, 187–220 (1972).

    MathSciNet  MATH  Google Scholar 

  55. 55.

    De Boer, P. T., Kroese, D. P., Mannor, S. & Rubinstein, R. Y. A tutorial on the cross-entropy method. Ann. Oper. Res. 134, 19–67 (2005).

    MathSciNet  MATH  Article  Google Scholar 

  56. 56.

    Smith, L. N. Cyclical learning rates for training neural networks. In Proc. 2017 IEEE Winter Conference on Applications of Computer Vision 464–472 (IEEE, 2017).

  57. 57.

    Abadi, M. et al. TensorFlow: a system for large-scale machine learning. Proc. 12th USENIX Conference on Operating Systems Design and Implementation 265–283 (USENIX, 2016).

  58. 58.

    Lambin, P. et al. Radiomics: extracting more information from medical images using advanced feature analysis. Eur. J. Cancer 48, 441–446 (2012).

    Article  Google Scholar 

  59. 59.

    Gevaert, O. et al. Non–small cell lung cancer: identifying prognostic imaging biomarkers by leveraging public gene expression microarray data—methods and preliminary results. Radiology 264, 387–396 (2012).

    Article  Google Scholar 

  60. 60.

    Gevaert, O. et al. Glioblastoma multiforme: exploratory radiogenomic analysis by using quantitative image features. Radiology 273, 168–174 (2015).

    Article  Google Scholar 

  61. 61.

    Huang, C. et al. Development and validation of radiomic signatures of head and neck squamous cell carcinoma molecular features and subtypes. EBioMedicine 45, 70–80 (2019).

    Article  Google Scholar 

  62. 62.

    Goeman, J. J. L1 penalized estimation in the Cox proportional hazards model. Biom. J. 2, 70–84 (2010).

    MATH  Google Scholar 

  63. 63.

    Davidson-Pilon, C. et al. CamDavidsonPilon/lifelines v0.21.1 (Zenodo, 2019); https://doi.org/10.5281/ZENODO.2652543

  64. 64.

    Bakr, S. et al. Data descriptor: a radiogenomic dataset of non-small cell lung cancer. Sci. Data 5, 180202 (2018).

    Article  Google Scholar 

  65. 65.

    Kalpathy-Cramer, J. et al. A comparison of lung nodule segmentation algorithms: methods and results from a multi-institutional study. J. Digit. Imaging 29, 476–487 (2016).

    Article  Google Scholar 

  66. 66.

    Grove, O. et al. Quantitative computed tomographic descriptors associate tumor shape complexity and intratumor heterogeneity with prognosis in lung adenocarcinoma. PLoS ONE 10, e0118261 (2015).

    Article  Google Scholar 

  67. 67.

    Armato, S. G. et al. The Lung Image Database Consortium (LIDC) and Image Database Resource Initiative (IDRI): a completed reference database of lung NODULES on CT scans. Med. Phys. 38, 915–931 (2011).

    Article  Google Scholar 

  68. 68.

    Mukherjee, P., Zhou, M., Lee, E. & Gevaert, O. LungNet: a shallow convolutional neural network predicts prognosis of lung cancer patients in multi-institutional CT-image data. Code Ocean https://codeocean.com/capsule/5978670/tree/v1 (2020).

Download references

Acknowledgements

Research reported in this publication was supported by the National Institute of Biomedical Imaging and Bioengineering (NIBIB) of the National Institutes of Health under award number R01EB020527 and R56EB020527. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. A Titan X Pascal used for this research was donated by the NVIDIA Corporation.

Author information

Affiliations

Authors

Contributions

Conception design: M.Z., E.L. and O.G. Provision of data: O.G., Y.B., S.N. and R.G. Data analysis and interpretation: P.M., M.Z., E.L. and O.G. Writing: all authors. Computation resource: O.G.

Corresponding author

Correspondence to Olivier Gevaert.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Supplementary information on the features extracted for radiomic analysis.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Mukherjee, P., Zhou, M., Lee, E. et al. A shallow convolutional neural network predicts prognosis of lung cancer patients in multi-institutional computed tomography image datasets. Nat Mach Intell 2, 274–282 (2020). https://doi.org/10.1038/s42256-020-0173-6

Download citation

Further reading

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing