Elsevier

Academic Radiology

Volume 11, Issue 9, September 2004, Pages 1011-1021
Academic Radiology

Computer-assisted radiology and surgery
Original investigations
Automated lung segmentation for thoracic CT: Impact on computer-aided diagnosis1

Presented in part at the 1999 annual meeting of the RSNA, the 1999 and 2000 annual meetings of the AAPM, and the 2003 CARS meeting.
https://doi.org/10.1016/j.acra.2004.06.005Get rights and content

Rationale and objectives

Automated lung segmentation in thoracic computed tomography scans is essential for the development of computer-aided diagnostic (CAD) methods. A core segmentation method may be developed for general application; however, modifications may be required for specific clinical tasks.

Materials and methods

An automated lung segmentation method has been applied (1) as preprocessing for automated lung nodule detection and (2) as the foundation for computer-assisted measurements of pleural mesothelioma tumor thickness. The core method uses gray-level thresholding to segment the lungs within each computed tomography section. The segmentation is revised through separation of right and left lungs along the anterior junction line, elimination of the trachea and main bronchi from the lung segmentation regions, and suppression of the diaphragm. Segmentation modifications required for nodule detection include a rolling ball algorithm to include juxtapleural nodules and morphologic erosion to eliminate partial volume pixels at the boundary of the segmentation regions.

Results

For automated lung nodule detection, 4 of 82 actual nodules (4.9%) were excluded from the lung segmentation regions when the core segmentation method was modified compared with 14 nodules (17.1%) excluded without modifications. The computer-assisted quantification of mesothelioma method achieved a correlation coefficient of 0.990 with 134 manual measurements when the core segmentation method was used alone; correlation was reduced to 0.977 when the segmentation modifications, as adapted for the lung nodule detection task, were applied to the mesothelioma measurement task.

Conclusion

Different CAD applications impose different requirements on the automated lung segmentation process. The specific approach to lung segmentation must be adapted to the particular CAD task.

Section snippets

Core segmentation method

The automated lung segmentation method is depicted in Fig 1. Aspects of this method have been reported previously (2, 10, 30, 31). With the exception of trachea and main bronchi elimination, which makes use of the contiguity of the airways between adjacent sections, each step is performed on a section-by-section basis. The overall method includes core techniques and also allows for the implementation of additional task-specific modifications to the lung segmentation regions.

As visualized on a

Nodule detection results

The automated lung nodule detection method achieved 71% nodule detection sensitivity with an average of 0.4 false-positive detections per section on a database of 38 CT scans (34). The effect of lung segmentation on these results may be appreciated through the observation that of the 82 actual lung nodules in this 38-case database, only 4 nodules (4.9%) were excluded from the lung segmentation regions when the core lung segmentation method was modified by the rolling ball algorithm and the

Conclusion

We have developed automated lung segmentation methods that provide an important part of our CAD research for thoracic CT scans. These methods have been successfully applied to the automated detection of lung nodules and to the computer-assisted measurement of mesothelioma. The present study demonstrates that the lung segmentation method used as preprocessing for CAD applications may affect CAD results. The core lung segmentation method yielded lung regions that excluded 17.1% of the lung

Acknowledgment

The authors would like to thank Heber MacMahon, MD, Masha Kocherginsky, PhD, Maryellen L. Giger, PhD, Nicholas J. Vogelzang, MD, Hedy L. Kindler, MD, Carl J. Vyborny, MD, PhD, John Fennessy, MD, Feng Li, MD, PhD, Geoffrey R. Oxnard, BA, Michael B. Altman, BA, Adam Starkey, and Tamara Thompson, RN, for their assistance.

References (36)

  • M. Fiebich et al.

    Automatic detection of pulmonary nodules in low-dose screening thoracic CT examinations

    Proc SPIE

    (1999)
  • H. Taguchi et al.

    Lung cancer detection based on helical CT images using curved surface morphology analysis

    Proc SPIE

    (1999)
  • S.-L. Lou et al.

    Object-based deformation technique for 3-D CT lung nodule detection

    Proc SPIE

    (1999)
  • S.G. Armato et al.

    Automated detection of lung nodules in CT scanspreliminary results

    Med Phys

    (2001)
  • Y. Masutani et al.

    Computerized detection of pulmonary embolism in spiral CT angiography based on volumetric image analysis

    IEEE Trans Med Imaging

    (2002)
  • M.S. Brown et al.

    . Knowledge-based segmentation of thoracic computed tomography images for assessment of split lung function

    Med Phys

    (2000)
  • M. Morgan

    Detection and quantification of pulmonary emphysema by computed tomographya window of opportunity

    Thorax

    (1992)
  • R. Uppaluri et al.

    Computer recognition of regional lung disease patterns

    Am J Respir Crit Care Med

    (1999)
  • Cited by (268)

    • Deep Residual Separable Convolutional Neural Network for lung tumor segmentation

      2022, Computers in Biology and Medicine
      Citation Excerpt :

      Armato and Sensakovic were able to prove the importance of accurate lung segmentation. They showed that 5–17% of small size tumors were missed in their test data due to incorrect lung segmentation [9]. This can affect the tumor segmentation, if the pipeline of tumor segmentation method includes lung segmentation process.

    View all citing articles on Scopus

    Supported in part by USPHS Grant CA83908, a grant from the Mesothelioma Applied Research Foundation, a grant from the American Lung Association of Metropolitan Chicago, and funding from The University of Chicago Cancer Research Center.

    View full text