Impact of Atlas Selection on Clinical Practice

There have been numerous studies investigating atlas-based contouring within the radiotherapy domain. Most of these studies only evaluate quantitative accuracy, yet a number have also considered the clinical impact with respect to contouring time. Table 3.2 shows studies where editing time has been investigated for atlas contouring. Although the atlas selection methods are not described in detail for commercial software, it is known that selection is being performed using mutual information as a similarity criteria by some manufacturers [26, 27].

It can be observed in Table 3.2 that for the prostate there are two studies where larger numbers of atlases have been used together with atlas selection, and two with fewer atlases and no selection. However, the reported percentage time saving is similar regardless of the number of atlases or the use of selection. While timing was not investigated, Lee et al. [28] found no improvement in performance using quantitative measures for a commercial system when increasing the size of an atlas database size from 20 to 100 atlases in steps of 20 atlases. Although that study only investigated two organs in the head and neck, these organs (mandible and thyroid) were chosen as laborious to draw but also represent organs with significant difference in contrast and appearance. Thus, the available evidence suggests that the selection methods implemented by some manufacturers have no impact on the resulting clinical workflow.

Summary and Recommendations for Future Research

In this chapter the background to atlas selection has been reviewed, seeing that since the outset the underlying assumption has been that an atlas with a more similar image to the test case will have better contouring performance than a less similar atlas. It was also seen that while there was some evidence to back up this assumption, it was not fully examined. An experiment was presented to test this assumption whereby a number of image-based atlas selection methods were evaluated and compared to perfect atlas selection. This study reveals that image-based methods fall some way short of perfect selection in terms of rank, and consequently atlas selection is having negligible impact in terms of clinical efficiency.

Notwithstanding this rather negative finding, there is some cause for optimism. Works exploring learned similarity measures have also shown encouraging improvements [15] over the direct image- based measures explored here, and the search for the optimum combination of atlases to generate the best consensus has only recently been considered [11, 24].

Schipaanboord et al., using extreme value theory, suggested that contouring performance equivalent to clinical standards could be possible using a large atlas database in the presence of perfect atlas selection [2]. Thus, there is scope for significant gains to be made with further research into atlas selection. However, as demonstrated in this chapter methods proposed in the future should evaluate their performance not with respect to where they have come from (i.e. older reference implementations) but with respect to where they want to get to (i.e. perfect selection).


  • 1. T Rohlfing, “Image similarity and tissue overlaps as surrogates for image registration accuracy: widely used but unreliable,” IEEE Trans. Med. Imaging, vol. 31. no. 2. pp. 153-163. Feb. 2012.
  • 2. В Schipaanboord et ah, “Can atlas-based auto-segmentation ever be perfect? Insights from extreme value theory.” IEEE Trans. Med. Imaging, vol. 38. no. 1. pp. 99-106, 2019.
  • 3. T Rohlfing, R Brandt, R Menzel, and C Maurer, “Evaluation of atlas selection strategies for atlas-based image segmentation with application to confocal microscopy images of bee brains,” Neuroimage, vol. 21, no. 4. pp. 1428-1442, 2004.
  • 4. T Rohlfing, DB Russakoff, and CR Maurer, “Performance-based classifier combination in atlas- based image segmentation using expectation-maximization parameter estimation,” IEEE Trans. Med. Imaging, vol. 23, no. 8. pp. 983-994, 2004.
  • 5. О Commowick, and G Malandain, “Efficient selection of the most similar image in a database for critical structures segmentation,” Led. Notes Comput. Sci. (including Subser. Led. Notes Artif. Intell. Led. Notes Bioinformatics), vol. 4792 LNCS, no. PART 2. pp. 203-210, 2007.
  • 6. M Wu, C Rosano, P Lopez-Garcia, CS Carter, and HJ Aizenstein, “Optimum template selection for atlas-based segmentation,” Neuroimage, vol. 34, no. 4, pp. 1612-1618,2007.
  • 7. A Akinyemi, C Plakas, J Piper, C Roberts, and I Poole, “Optimal atlas selection using image similarities in a trained regression model to predict performance,” Proc. Int. Symp. Bioined. Imaging, pp. 1264-1267, 2012.
  • 8. TR Langerak, FF Berendsen, UA Van Der Heide. ANTJ Kotte, and JPW Pluim, “Multiatlas-based segmentation with preregistration atlas selection,” Med. Phys., vol. 40, no. 9, pp. 1-17, 2013.
  • 9. P Aljabar, RA Heckemann, A Hammers, JV Hajnal, and D Rueckert, “Multi-atlas based segmentation of brain images: atlas selection and its effect on accuracy,” Neuroimage, vol. 46, no. 3, pp. 726-738, 2009.
  • 10. TR Langerak, UA Van Der Heide, ANTJ Kotte. MA Viergever, M Van Vulpen. and JPW Pluim, “Label fusion in atlas-based segmentation using a selective and iterative method for performance level estimation (SIMPLE),” IEEE Trans. Med. Imaging, vol. 29, no. 12, pp. 2000-2008, 2010.
  • 11. M Antonelli et ah, “GAS: a genetic atlas selection strategy in multi-atlas segmentation framework,” Med. Image Anal., vol. 52. pp. 97-108, 2019.
  • 12. Y Cao, Y Yuan, X Li, and P Yan, “Putting images on a manifold for atlas-based image segmentation,” Proc. - Int. Conf. Image Process. ICIP, No. May 2016, pp. 289-292, 2011.
  • 13. R Wolz, P Aljabar, JV Hajnal, A Hammers, and D Rueckert, “LEAP: learning embeddings for atlas propagation,” Neuroimage, vol. 49, no. 2, pp. 1316-1325, 2010.
  • 14. AKHoang Due et ah, “Using manifold learning for atlas selection in multi-atlas segmentation,” PLoS One, vol. 8. no. 8, 2013.
  • 15. G Sanroma, G Wu, Y Gao, and D Shen, “Learning to rank atlases for multiple-atlas segmentation,” IEEE Trans. Med. Imaging, vol. 33, no. 10, pp. 1939-1953, Oct. 2014.
  • 16. L Ramus, and G Malandain, “Assessing selection methods in the context of multi-atlas based segmentation,” in 2010 IEEE International Symposium on Biomedical Imaging: From Nano to Macro, 2010, pp. 1321-1324.
  • 17. В Schipaanboord et al., “An evaluation of atlas selection methods for atlas-based automatic segmentation in radiotherapy treatment planning,” IEEE Trans. Med. Imaging, vol. 38, no. 11, pp. 2654-2664, 2019.
  • 18. J Yang et al., “Autosegmentation for thoracic radiation treatment planning: a grand challenge at AAPM 2017,” Med. Phys., vol. 45, no. 10, pp. 4568-4581. Oct. 2018.
  • 19. S Gorthi, and M Cuadra, “Multi-Atlas based segmentation of head and neck CT images using active contour framework," MICCAI Work. 3D Segmentation Chall. Clin. Appl, pp. 313-321, 2010.
  • 20. LR Dice, “Measures of the amount of ecologic association between species,” Ecology, vol. 26, no. 3, pp. 297-302, 1945.
  • 21. JM Lotjonen et al., “Fast and robust multi-atlas segmentation of brain magnetic resonance images,” Neuroimage, vol. 49, no. 3, pp. 2352-2365, 2010.
  • 22. P Raudaschl. К Fritscher. P Zaffino. GC Sharp. MF Spadea, and R Schubert, “A novel atlas-selection approach for multi-atlas based segmentation using the correlation of inter-atlas similarities,” in Proceedings of Image-Guided Adaptive Radiation Therapy Workshop, 2014, pp. 53-60.
  • 23. T Zhao, and D Ruan, “Learning image based surrogate relevance criterion for atlas selection in segmentation,” Phys. Med. Biol., vol. 61. no. 11. pp. 4223-4234. 2016.
  • 24. P Zaffino et al., “Multi atlas based segmentation: should we prefer the best atlas group over the group of best atlases?” Phys. Med. Biol., vol. 63, no. 12, 2018.
  • 25. G Sharp et al., “Vision 20/20: perspectives on automated image segmentation for radiotherapy,” Med. Phys., vol. 41, no. 5. pp. 1-13, 2014.
  • 26. DN Teguh et al., “Clinical validation of atlas-based auto-segmentation of multiple target volumes and normal tissue (swallowing/mastication) structures in the head and neck,” Int. J. Radiat. Oncol. Biol. Phys., vol. 81. no. 4, pp. 950-957, 2011.
  • 27. AV Young, A Wortham, I Wernick, A Evans, and RD Ennis, “Atlas-based segmentation improves consistency and decreases time required for contouring postoperative endometrial cancer nodal volumes,” Int. J. Radiat. Oncol. Biol. Phys., vol. 79, no. 3, pp. 943-947, 2011.
  • 28. H Lee et al., “Clinical evaluation of commercial atlas-based auto-segmentation in the head and neck region,” Front. Oncol., vol. 9, pp. 1-9, 2019.
  • 29. S Klein, UA Van Der Heide, IM Lips, M Van Vulpen, M Staring, and JPW Pluim, “Automatic segmentation of the prostate in 3D MR images by atlas matching using localized mutual information,” Med. Phys., vol. 35, no. 4. pp. 1407-1417, 2008.
  • 30. MR Sabuncu, SK Bald, ME Shenton, and P Golland, “Image-driven population analysis through mixture modeling,” IEEE Trans. Med. Imaging, vol. 28. no. 9. pp. 1473-1487. Sep. 2009.
  • 31.1 Isgum, M Staring, A Rutten, M Prokop, MA Viergever, and В Van Ginneken, “Multi-atlas-based segmentation with local decision fusion-application to cardiac and aortic segmentation in CT scans,” IEEE Trans. Med. Imaging, vol. 28, no. 7, pp. 1000-1010, 2009.
  • 32. L Ramus, and G Malandain, “Multi-atlas based segmentation: application to the head and neck region for radiotherapy planning,” Med. Image Anal. Clin., pp. 281-288,2010.
  • 33. EM van Rikxoort et al., “Adaptive local multi-atlas segmentation: application to the heart and the caudate nucleus,” Med. Image Anal., vol. 14, no. 1, pp. 39-49, 2010.
  • 34. J Yang, Y Zhang, L Zhang, and L Dong, “Automatic segmentation of parotids from CT scans using multiple atlases,” in Medical Image Analysis for the Clinic: A Grand Challenge, 2010, pp. 323-330. www.a
  • 35. JA Dowling et al., “Fast automatic multi-atlas segmentation of the prostate from 3D MR images,” Led. Notes Comput. Sci. (including Suhser. Led. Notes Art if Intell. Led. Notes Bioinformatics), vol. 6963 LNCS.pp. 10-21,2011.
  • 36. R Wolz, C Chu, К Misawa, M Fujiwara, К Mori, and D Rueckert, “Automated abdominal multi-organ segmentation with subject-specific atlas generation,” IEEE Trans. Med. Imaging, vol. 32, no. 9, pp. 1723-1730, 2013.
  • 37. AJ Asman, FW Bryan, SA Smith, DS Reich, and BA Landman, “Groupwise multi-atlas segmentation of the spinal cord’s internal structure,” Med. Image Anal., vol. 18, no. 3, pp. 460-471,2014.
  • 38. H Wang, JW Suh, SR Das, JB Pluta, C Craige, and PA Yushkevich, “Multi-atlas segmentation with joint label fusion,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 35, no. 3. pp. 611-623, 2013.
  • 39. AJ Asman, Y Huo, AJ Plassard, and BA Landman, “Multi-atlas learner fusion: an efficient segmentation approach for large-scale data,” Med. Image Anal., vol. 26, no. 1, pp. 82-91, Dec. 2015.
  • 40. TR Langerak, UA Van Der Heide, ANTJ Kotte, FF Berendsen, and JPW Pluirn, “Improving label fusion in multi-atlas based segmentation by locally combining atlas selection and performance estimation,” Comput. Vis. Image Undent., vol. 130, pp. 71-79, 2015.
  • 41. Z Xu et al., “Efficient multi-atlas abdominal segmentation on clinically acquired CT with SIMPLE context learning,” Med. Image Anal., vol. 24, no. 1, pp. 18-27, 2015.
  • 42. P Yan, Y Cao, Y Yuan, В Turkbey, and PL Choyke, “Label image constrained multiatlas selection,” IEEE Trans. Cybern., vol. 45. no. 6, pp. 1158-1168, 2015.
  • 43. К Karasawa et al., “Multi-atlas pancreas segmentation: atlas selection based on vessel structure,” Med. Image Anal., vol. 39, pp. 18-28, 2017.
  • 44. J Yang et al., “Atlas ranking and selection for automatic segmentation of the esophagus from CT scans,” Phys. Med. Biol., vol. 62, no. 23, pp. 9140-9158, 2017.
  • 45. LJ Stapleford et al., “Evaluation of automatic atlas-based lymph node segmentation for head-and-neck cancer,” Int. J. Radiat. Oncol. Biol. Phys., vol. 77, no. 3. pp. 959-966, 2010.
  • 46. MA Gambacorta et al., “Clinical validation of atlas-based auto-segmentation of pelvic volumes and normal tissue in rectal tumors using auto-segmentation computed system,” Acta Oncol. (Madr)., vol. 52, no. 8. pp. 1676-1681,2013.
  • 47. J Hwee et al., “Technology assessment of automated atlas based segmentation in prostate bed contouring,” Rad. Oncol., vol. 6, no. 1. pp. 1-9, 2011.
  • 48. A Lin, G Kubicek, JW Piper, AS Nelson, AP Dicker, and RK Valicenti, “Atlas-based segmentation in prostate IMRT: timesavings in the clinical workflow,” Int. J. Radiat. Oncol., vol. 72, no. 1, pp. S328- S329, 2008.
  • 49. C Granberg, Clinical Evaluation of Atlas Based Segmentation for Radiotherapy of Prostate Tumours, M.S. Thesis, pp. 1-66, 2011.
  • 50. KA Langmack, C Perry, C Sinstead, J Mills, and D Saunders, “The utility of atlas-assisted segmentation in the male pelvis is dependent on the interobserver agreement of the structures segmented,” Br. J. Radiol., vol. 87, no. 1043, 2014.
< Prev   CONTENTS   Source   Next >