Automated Detection of Mediastinal Lymph Nodes for Assistance of Transbronchial Needle Aspiration

  • Takayuki Kitasaka
  • Mitsuhiro Kishimoto
  • Masahiro Oda
  • Shingo Iwano
  • Kensaku Mori
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7815)


This paper proposes an automated method for detecting mediastinal lymph nodes from contrasted 3D chest CT images to assist transbronchial needle aspiration (TBNA). In lung cancer treatment, it is vital to check whether metastasis occurs in lymph nodes. For this purpose, two-phase contrasted CT images are used for precise and accurate image diagnosis in the clinical field. In the proposed method, two input CT images are first registered and then the processing area is calculated on the basis of intensity information and an atlas for the lymph node map. Lymph node metastasis candidates are extracted roughly by using local intensity structure analysis. Lymph node metastasis regions are obtained by using a shape modification process based on using the shape features of candidates to achieve region growth and false positive (FP) reduction. To provide assistance for TBNA, detection results are visualized through the semi-translucent bronchial wall by virtual bronchoscopy. In experimentally applying the method to 42 pairs of contrasted chest CT images, the results show that the method detected 92.9% of the lymph node metastases, while a radiologist detected 57.1%. The lymph nodes were clearly visualized in the virtual bronchoscopy as the simulation of TBNA guidance.


Lymph Node Mediastinal Lymph Node Enlarge Lymph Node Lung Cancer Treatment Detect Lymph Node Metastasis 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Pass, H.I.: Mediastinal Staging 2005: Pictures, Scopes, and Scalpels. Seminars in Oncology 32, 269–278 (2005)CrossRefGoogle Scholar
  2. 2.
    Feuerstein, M., Glocker, B., Kitasaka, T., Nakamura, Y., Iwano, S., Mori, K.: Mediastinal Atlas Creation from 3-D Chest Computed Tomography Images: Application to Automated Detection and Station Mapping of Lymph Nodes. Medical Image Analysis 16(1), 63–74 (2012)CrossRefGoogle Scholar
  3. 3.
    Feulner, J., Kevin Zhou, S., Hammon, M., Hornegger, J., Comaniciu, D.: Segmentation Based Features for Lymph Node Detection from 3-D Chest CT. In: Suzuki, K., Wang, F., Shen, D., Yan, P. (eds.) MLMI 2011. LNCS, vol. 7009, pp. 91–99. Springer, Heidelberg (2011)CrossRefGoogle Scholar
  4. 4.
    Glocker, B., Komodakis, N., Tziritas, G., Navab, N., Paragios, N.: Dense image registration through MRFs and efficient linear programming. Medical Image Analysis 12(6), 731–741 (2008)CrossRefGoogle Scholar
  5. 5.
    Feuerstein, M., Kitasaka, T., Mori, K.: Automated Anatomical Likelihood Driven Extraction and Branching Detection of Aortic Arch in 3-D Chest CT. In: The Second International Workshop on Pulmonary Image Analysis, MICCAI 2009, pp. 49–60 (2009)Google Scholar
  6. 6.
    Sato, Y., Westin, C.F., Bhalerao, A., Nakagima, S., Shiraga, N., Tamura, S., Kikinis, R.: Tissue Classification Based on 3D Local Intensity Structures for Volume Rendering. IEEE Transactions on Visualization and Computer Graphics 6(2), 160–180 (2000)CrossRefGoogle Scholar
  7. 7.
    Nagao, M., Matsuyama, T.: Edge preserving smoothing. Computer Graphics and Image Processing 9(4), 394–407 (1979)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2013

Authors and Affiliations

  • Takayuki Kitasaka
    • 1
  • Mitsuhiro Kishimoto
    • 2
  • Masahiro Oda
    • 2
  • Shingo Iwano
    • 3
  • Kensaku Mori
    • 4
  1. 1.School of Information ScienceAichi Institute of TechnologyJapan
  2. 2.Graduate School of Information ScienceNagoya UniversityJapan
  3. 3.Graduate School of MedicineNagoya UniversityJapan
  4. 4.Strategy Office, Information and Communications HeadquartersNagoya UniversityJapan

Personalised recommendations