Voxel classification and graph cuts for automated segmentation of pathological periprosthetic hip anatomy
- 980 Downloads
Automated patient-specific image-based segmentation of tissues surrounding aseptically loose hip prostheses is desired. For this we present an automated segmentation pipeline that labels periprosthetic tissues in computed tomography (CT). The intended application of this pipeline is in pre-operative planning.
Individual voxels were classified based on a set of automatically extracted image features. Minimum-cost graph cuts were computed on the classification results. The graph-cut step enabled us to enforce geometrical containment constraints, such as cortical bone sheathing the femur’s interior. The solution’s novelty lies in the combination of voxel classification with multilabel graph cuts and in the way label costs were defined to enforce containment constraints.
The segmentation pipeline was tested on a set of twelve manually segmented clinical CT volumes. The distribution of healthy tissue and bone cement was automatically determined with sensitivities greater than 82% and pathological fibrous interface tissue with a sensitivity exceeding 73%. Specificity exceeded 96% for all tissues.
The addition of a graph-cut step improved segmentation compared to voxel classification alone. The pipeline described in this paper represents a practical approach to segmenting multitissue regions from CT.
KeywordsSegmentation Graph cut Voxel classification Osteolysis Computed tomography
This research is supported by the Dutch Technology Foundation STW, which is the applied science division of NWO, and the Technology Programme of the Ministry of Economic Affairs (project number LKG 7943). The authors sincerely thank Professor Rob Nelissen for verifying the correctness of the manually segmented ground truth and Noeska Smit for the re-segmentation of several of the data sets for evaluating inter-observer variability.We furthermore thank David Tax and Marco Loog from the pattern recognition group at Delft University of Technology for their helpful input.
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.
- 3.Raaijmaakers M, Mulier M (2010) Percutaneous in situ cementation of a loose femoral stem. J Arthroplast 25:1169.e21–1169.e24. doi: 10.1016/j.arth.2009.03.027
- 13.Yokota F, Okada T, Takao M, Sugano N, Tada Y, Sato Y (2009) Automated segmentation of the femur and pelvis from 3D CT data of diseased hip using hierarchical statistical shape model of joint structure. Med Image Comput Comput Assist Interv 12: 811–818. doi: 10.1007/978-3-642-04271-3_98 PubMedGoogle Scholar
- 14.Shlens J (2005) A tutorial on principal component analysis. Systems Neurobiology Laboratory, Salk Institute for Biological Studies. http://www.snl.salk.edu/~shlens/pca.pdf. Accessed 16 Nov 2011
- 16.Malan DF, Botha CP, Nelissen RGHH, Valstar ER (2010) Voxel classification of periprosthetic tissues in clinical computed tomography of loosened hip prostheses. In: Proceedings of the IEEE international symposium on biomedical imaging: from nano to macro, Rotterdam, The Netherlands, pp 1341–1344Google Scholar
- 18.Delong A, Boykov Y (2009) Globally optimal segmentation of multi-region objects. In: Proceedings of the IEEE 12th international computer vision conference, Kyoto, Japan, pp 285–292Google Scholar
- 19.Veksler O (2010) Code: multi-label optimization. University of Western Ontario. http://vision.csd.uwo.ca/code/. Accessed 22 Sept 2010
- 24.Botha CP (2008) Hybrid scheduling in the DeVIDE dataflow visualisation environment. In: Hauser H, Strassburger S, Theisel H (eds) Proceedings of simulation and visualization, pp 309–322Google Scholar
- 28.Paclik P, Lai C (2011) PRSD Studio. PR Sys Design, Delft, The Netherlands. http://www.prsdstudio.com/. Accessed 16 Nov 2011
- 29.Greig DM, Porteous BT (1989) Exact maximum a posteriori estimation for binary images. J R Stat Soc Ser B Stat Methodol 51: 271–279Google Scholar
- 30.Ahuja RK, Magnanti TL, Orlin JB (1993) Maximum flows: polynomial algorithms. In: Janzow P, Peterson M (eds) Network flows. Prentice Hall, Englewood Cliffs, New Jersey, p 240Google Scholar
- 33.Egger J, Colen RR, Freisleben B, Nimsky C (2011) Manual refinement system for graph-based segmentation results in the medical domain. J Med Syst. doi: 10.1007/s10916-011-9761-7
- 34.Schwarz EM, Campbell D, Totterman S, Boyd A, O’Keefe RJ (2003) Use of volumetric computerized tomography as a primary outcome measure to evaluate drug efficacy in the prevention of peri-prosthetic osteolysis: a 1-year clinical pilot of etanercept vs. placebo. J Orthop Res 21: 1049–1055. doi: 10.1016/S0736-0266(03)00093-7 PubMedCrossRefGoogle Scholar
- 35.Malan DF, Botha CP, Kraaij G, Joemai RM, van der Heide HJL, Nelissen RGHH, Valstar ER (2011) Measuring femoral lesions despite CT metal artefacts: a cadaveric study. Skeletal Radiol. doi: 10.1007/s00256-011-1223-2