Primary lung tumor segmentation from PET–CT volumes with spatial–topological constraint

  • Hui Cui
  • Xiuying Wang
  • Weiran Lin
  • Jianlong Zhou
  • Stefan Eberl
  • Dagan Feng
  • Michael Fulham
Original Article
First Online:
Received:
Accepted:

Abstract

Purpose

Accurate lung tumor segmentation is a prerequisite for effective radiation therapy and surgical planning. However, tumor delineation is challenging when the tumor boundaries are indistinct on PET or CT. To address this problem, we developed a segmentation method to improve the delineation of primary lung tumors from PET–CT images.

Methods

We formulated the segmentation problem as a label information propagation process in an iterative manner. Our model incorporates spatial–topological information from PET and local intensity changes from CT. The topological information of the regions was extracted based on the metabolic activity of different tissues. The spatial–topological information moderates the amount of label information that a pixel receives: The label information attenuates as the spatial distance increases and when crossing different topological regions. Thus, the spatial–topological constraint assists accurate tumor delineation and separation. The label information propagation and transition model are solved under a random walk framework.

Results

Our method achieved an average DSC of \(0.848 \pm 0.036\) and HD (mm) of \(8.652 \pm 4.532\) on 40 patients with lung cancer. The t test showed a significant improvement (p value \(<\) 0.05) in segmentation accuracy when compared to eight other methods. Our method was better able to delineate tumors that had heterogeneous FDG uptake and which abutted adjacent structures that had similar densities.

Conclusions

Our method, using a spatial–topological constraint, provided better lung tumor delineation, in particular, when the tumor involved or abutted the chest wall and the mediastinum.

Keywords

PET/CT Segmentation NSCLC  Graph topology 
This is a preview of subscription content, log in to check access.

Notes

Conflict of interest

Hui Cui, Xiuying Wang, Weiran Lin, Jianlong Zhou, Stefan Eberl, Dagan Feng, and Michael Fulham declare that they have no conflict of interest.

References

  1. 1.
    MacManus M et al (2009) Use of PET and PET/CT for radiation therapy planning: IAEA expert report 2006–2007. Radiother Oncol 91(1):85–94CrossRefPubMedGoogle Scholar
  2. 2.
    Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D (2011) Global cancer statistics. Cancer J Clin 61(2):69–90CrossRefGoogle Scholar
  3. 3.
    Kirov A, Fanchon L (2014) Pathology-validated PET image data sets and their role in PET segmentation. Clin Transl Imaging 2(3):253–267CrossRefGoogle Scholar
  4. 4.
    Simona B-H, Peter E (2009) 18F-FDG PET and PET/CT in the evaluation of cancer treatment response. J Nucl Med 50(1):88–99Google Scholar
  5. 5.
    Boellaard R (2012) Mutatis mutandis: harmonize the standard!. J Nucl Med 53(1):1–3CrossRefPubMedGoogle Scholar
  6. 6.
    Hatt M et al (2009) A fuzzy locally adaptive Bayesian segmentation approach for volume determination in PET. IEEE Trans Med Imaging 28(6):881–893PubMedCentralCrossRefPubMedGoogle Scholar
  7. 7.
    Hatt M et al (2010) Accurate automatic delineation of heterogeneous functional volumes in positron emission tomography for oncology applications. Int J Radiat Oncol Biol Phys 77(1):301–308CrossRefPubMedGoogle Scholar
  8. 8.
    Bagci U et al (2013) Joint segmentation of anatomical and functional images: applications in quantification of lesions from PET, PET–CT, MRI–PET, and MRI–PET–CT images. Med Image Anal 17(8):929–945PubMedCentralCrossRefPubMedGoogle Scholar
  9. 9.
    Foster B et al (2014) Segmentation of PET images for computer-aided functional quantification of tuberculosis in small animal models. IEEE Trans Biomed Eng 61(3):711–724PubMedCentralCrossRefPubMedGoogle Scholar
  10. 10.
    Foster B et al (2014) A review on segmentation of positron emission tomography images. Comput Biol Med 50:76–96CrossRefPubMedGoogle Scholar
  11. 11.
    Wang X et al (2014) Lung tumor delineation based on novel tumor-background likelihood models in PET–CT images. IEEE Trans Nucl Sci 61(1):218–224CrossRefGoogle Scholar
  12. 12.
    Han D et al (2011) Globally optimal tumor segmentation in PET-CT images: a graph-based co-segmentation method. Inf Process Med Imaging 22:245–256PubMedCentralPubMedGoogle Scholar
  13. 13.
    Xu Z et al (2014) Fuzzy connectedness image co-segmentation for hybrid PET/MRI and PET/CT scans. In: Computational methods for molecular imaging workshop, medical image computing and computer-assisted intervention (MICCAI), BostonGoogle Scholar
  14. 14.
    Boykov Y, Veksler O, Zabih R (2001) Fast approximate energy minimization via graph cuts. IEEE Trans Pattern Anal Mach Intell 23(11):1222–1239CrossRefGoogle Scholar
  15. 15.
    Grady L (2006) Random walks for image segmentation. IEEE Trans Pattern Anal Mach Intell 28(11):1768–1783CrossRefPubMedGoogle Scholar
  16. 16.
    Qi S et al (2013) Optimal co-segmentation of tumor in PET–CT images with context information. IEEE Trans Med Imaging 32(9):1685–1697CrossRefGoogle Scholar
  17. 17.
    Grady L et al (2005) Random walks for interactive organ segmentation in two and three dimensions: implementation and validation. In: Medical image computing and computer-assisted intervention, pp 773–780Google Scholar
  18. 18.
    Chen M et al (2011) Random walk-based automated segmentation for the prognosis of malignant pleural mesothelioma. In: IEEE international symposium on biomedical imaging: from nano to macro, pp 1978–1981Google Scholar
  19. 19.
    Carr H, Snoeyink J, Axen U (2003) Computing contour trees in all dimensions. Comput Geom Theory Appl 24(2):75–94CrossRefGoogle Scholar
  20. 20.
    Zhou J (2012) Gaining insights into volumetric data visualization: a semi-automatic transfer function generation approach using contour tree analyses. LAP LAMBERT Academic Publishing, GermanyGoogle Scholar
  21. 21.
    Cui H et al (2014) Topology constraint graph-based model for non-small-cell lung tumor segmentation from PET volumes. In: IEEE international symposium on biomedical imaging (ISBI), pp 1243–1246Google Scholar
  22. 22.
    RIDER Phantom PET–CT: University of Washington. https://wiki.cancerimagingarchive.net/display/Public/RIDER+Collections
  23. 23.
    Black QC et al (2004) Defining a radiotherapy target with positron emission tomography. Int J Radiat Oncol Biol Phys 60(4):1272–1282CrossRefPubMedGoogle Scholar
  24. 24.
    Ballangan C et al (2011) Automated delineation of lung tumors in PET images based on monotonicity and a tumor-customized criterion. IEEE Trans Inf Technol Biomed 15(5):691–702CrossRefPubMedGoogle Scholar
  25. 25.
    Kim TH, Lee KM, Lee SU (2010) Nonparametric higher-order learning for interactive segmentation. In: IEEE conference on computer vision and pattern recognition, pp 3201–3208Google Scholar
  26. 26.
    Grady L (2003) The graph analysis toolbox: image processing on arbitrary graphs 2003. Boston University, Boston, MA, Tech. Rep. TR-03-021Google Scholar
  27. 27.
    Kim TH. Generative image segmentation using random walks with restart. http://cv.snu.ac.kr/thkim/RWR_ECCV08/index.html
  28. 28.
    Kim TH, Lee KM, Lee SU (2008) Generative image segmentation using random walks with restart. In: European conference on computer vision, pp 264–275Google Scholar
  29. 29.
    Kinahan PE, Doot RK, Wanner-Roybal M, Bidaut LM, Armato SG, Meyer CR, McLennan G (2009) PET/CT assessment of response to therapy: tumor change measurement, truth data, and error. Transl Oncol 2(4):223–230PubMedCentralCrossRefPubMedGoogle Scholar
  30. 30.
    Shah B et al (2012) Intra-reader reliability of FDG PET volumetric tumor parameters: effects of primary tumor size and segmentation methods. Ann Nucl Med 26(9):707–714CrossRefPubMedGoogle Scholar
  31. 31.
    Zhou J, Xiao C, Takatsuka M (2013) A multi-dimensional importance metric for contour tree simplification. J Vis 16(4):341–349CrossRefGoogle Scholar

Copyright information

© CARS 2015

Authors and Affiliations

  • Hui Cui
    • 1
  • Xiuying Wang
    • 1
  • Weiran Lin
    • 1
  • Jianlong Zhou
    • 2
  • Stefan Eberl
    • 3
  • Dagan Feng
    • 1
    • 4
  • Michael Fulham
    • 3
    • 5
  1. 1.Biomedical and Multimedia Information Technology (BMIT) Research Group, School of Information TechnologiesUniversity of SydneySydneyAustralia
  2. 2.National ICTSydneyAustralia
  3. 3.Department of PET and Nuclear MedicineRoyal Prince Alfred HospitalSydneyAustralia
  4. 4.Med-X Research InstituteShanghai Jiao Tong UniversityShanghaiChina
  5. 5.Sydney Medical SchoolUniversity of SydneySydneyAustralia

Personalised recommendations