Fast GPU-based computation of spatial multigrid multiframe LMEM for PET | Medical & Biological Engineering & Computing Skip to main content
Springer Nature Link
Log in

Fast GPU-based computation of spatial multigrid multiframe LMEM for PET

  • Original Article
  • Published:
Medical & Biological Engineering & Computing Aims and scope Submit manuscript

Abstract

Significant efforts were invested during the last decade to accelerate PET list-mode reconstructions, notably with GPU devices. However, the computation time per event is still relatively long, and the list-mode efficiency on the GPU is well below the histogram-mode efficiency. Since list-mode data are not arranged in any regular pattern, costly accesses to the GPU global memory can hardly be optimized and geometrical symmetries cannot be used. To overcome obstacles that limit the acceleration of reconstruction from list-mode on the GPU, a multigrid and multiframe approach of an expectation-maximization algorithm was developed. The reconstruction process is started during data acquisition, and calculations are executed concurrently on the GPU and the CPU, while the system matrix is computed on-the-fly. A new convergence criterion also was introduced, which is computationally more efficient on the GPU. The implementation was tested on a Tesla C2050 GPU device for a Gemini GXL PET system geometry. The results show that the proposed algorithm (multigrid and multiframe list-mode expectation-maximization, MGMF-LMEM) converges to the same solution as the LMEM algorithm more than three times faster. The execution time of the MGMF-LMEM algorithm was 1.1 s per million of events on the Tesla C2050 hardware used, for a reconstructed space of \(188 \times 188\times 57\) voxels of \(2\times 2\times 3.15\,\hbox {mm}^3\). For 17- and 22-mm simulated hot lesions, the MGMF-LMEM algorithm led on the first iteration to contrast recovery coefficients (CRC) of more than 75 % of the maximum CRC while achieving a minimum in the relative mean square error. Therefore, the MGMF-LMEM algorithm can be used as a one-pass method to perform real-time reconstructions for low-count acquisitions, as in list-mode gated studies. The computation time for one iteration and 60 millions of events was approximately 66 s.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Subscribe and save

Springer+ Basic
¥17,985 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price includes VAT (Japan)

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7

Similar content being viewed by others

References

  1. Vandenberghe S, Daube-Witherspoon ME, Lewitt RM, Karp JS (2006) Fast reconstruction of 3D time-of-flight PET data by axial rebinning and transverse mashing. Phys Med Biol 51(6):1603

    Article  PubMed  Google Scholar 

  2. Qi J, Leahy RM, Hsu C, Farquhar T, Cherry SR (1998) Fully 3D Bayesian image reconstruction for the ECAT EXACT HR+. IEEE Trans Nucl Sci 45(3):1096

    Article  Google Scholar 

  3. Nassiri MA, Hissoiny S, Carrier JF, Després P (2012) Fast GPU-based computation of the sensitivity matrix for a PET list-mode OSEM algorithm. Phys Med Biol 57(19):6279

    Article  PubMed  Google Scholar 

  4. Scheins J, Herzog H, Shah N (2011) Fully-3D PET image reconstruction using scanner-independent, adaptive projection data and highly rotation-symmetric voxel assemblies. IEEE Trans Med Imaging 30(3):879 (ISSN 0278–0062)

    Article  CAS  PubMed  Google Scholar 

  5. Zhou J, Qi J (2011) Fast and efficient fully 3D PET image reconstruction using sparse system matrix factorization with GPU acceleration. Phys Med Biol 56(20):6739

    Article  PubMed Central  PubMed  Google Scholar 

  6. Hu Z, Wang W, Gualtieri EE, Hsieh YL, Karp JS, Matej S, Parma M, Tung C, Walsh E, Werner M, Gagnon D (2007) An LOR-based fully-3D PET image reconstruction using a blob-basis function. In: Nuclear Science Symposium Conference Record, 2007. NSS ’07. IEEE, volume 6, pp. 4415–4418. ISSN 1095–7863

  7. Hong I, Chung S, Kim H, Kim Y, Son Y, Cho Z (2007) Ultra fast symmetry and SIMD-based projection-backprojection (SSP) algorithm for 3-D PET image reconstruction. IEEE Trans Med Imaging 26(6):789

    Article  CAS  PubMed  Google Scholar 

  8. Jacobson MS, Levkovitz R, Ben-Tal A, Thielemans K, Spinks T, Belluzzo D, Pagani E, Bettinardi V, Gilardi M, Zverovich A et al (2000) Enhanced 3D PET OSEM reconstruction using inter-update Metz filtering. Phys Med Biol 45(8):2417

    Article  CAS  PubMed  Google Scholar 

  9. Ranganath M, Dhawan A, Mullani N (1988) A multigrid expectation maximization reconstruction algorithm for positron emission tomography. IEEE Trans Med Imaging 7(4):273

    Article  CAS  PubMed  Google Scholar 

  10. Raheja A, Doniere T, Dhawan A (1999) Multiresolution expectation maximization reconstruction algorithm for positron emission tomography using wavelet processing. IEEE Trans Nucl Sci 46(3):594

    Article  Google Scholar 

  11. Oh S, Milstein AB, Bouman CA, Webb KJ (2005) Multiresolution expectation maximization reconstruction algorithm for positron emission tomography using wavelet processing. IEEE Trans Image Process 14(1):125

    Article  PubMed  Google Scholar 

  12. Oh S, Bouman CA, Webb KJ (2006) Multigrid tomographic inversion with variable resolution data and image spaces. IEEE Trans Image Process 15(9):2805

    Article  PubMed  Google Scholar 

  13. Mendes L, Ferreira N, Comtat C (2011) A multiscale-multiframe approach to 3D PET data reconstruction. In: Proceedings of The 11th International Meeting on Fully Three-Dimensional Image Reconstruction in Radiology and Nuclear Medicine and the 3rd Workshop on High Performance Image Reconstruction, pp. 411–413

  14. Rahmim A, Cheng JC, Blinder S, Camborde ML, Sossi V (2005) Statistical dynamic image reconstruction in state-of-the-art high-resolution PET. Phys Med Biol 50:4887

    Article  PubMed  Google Scholar 

  15. Matej S, Surti S, Jayanthi S, Daube-Witherspoon ME, Lewitt RM, Karp JS (2009) Efficient 3-D TOF PET reconstruction using view-grouped histo-images: direct image reconstruction for TOF. IEEE Trans Med Imaging 28(5):739

    Article  PubMed Central  PubMed  Google Scholar 

  16. Grotus N, Reader AJ, Stute S, Rosenwald J, Giraud P, Buvat I (2009) Fully 4D list-mode reconstruction applied to respiratory-gated PET scans. Phys Med Biol 54:1705

    Article  CAS  PubMed  Google Scholar 

  17. Qi J (2006) Calculation of the sensitivity image in list-mode reconstruction for PET. IEEE Trans Nucl Sci 53(5):2746

    Article  Google Scholar 

  18. Rahmim A, Lenox M, Reader AJ, Michel C, Burbar Z, Ruth TJ, Sossi V (2004) Statistical list-mode image reconstruction for the high resolution research tomograph. Phys Med Biol 49(18):4239

    Article  CAS  PubMed  Google Scholar 

  19. Carson R, Barker W, Liow J, Johnson C (2003) Design of a motion compensation OSEM list-mode algorithm for resolution-recovery reconstruction for the HRRT. In: Nuclear Science Symposium Conference Record, 2003 IEEE, volume 5, pp. 3281–3285 IEEE

  20. Autret A, Bert J, Strauss O, Visvikis D et al (2013) Fully 3D PET list-mode reconstruction including an accurate detector modeling on GPU architecture. In: FULLY 3D 2013: International meeting on fully three dimensional image reconstruction in radiology and nuclear medecine

  21. Autret A, Bert J, Zakaria B, Strauss O, Visvikis D, et al (2013) Accurate fully 3D list-mode PET reconstruction on multi-GPUs. In: RITS 2013: Colloque Recherche en imagerie et technologies pour la santé

  22. Bahi Z, Bert J, Autret A, Visvikis D (2012) High performance multi-GPU acceleration for fully 3D list-mode PET reconstruction. In: IEEE Nuclear Science Symposium and Medical Imaging Conference Record (NSS/MIC) pp. 3390–3393. IEEE

  23. Bert J, Visvikis D (2011) A fast CPU/GPU ray projector for fully 3d list-mode PET reconstruction. In: Nuclear Science Symposium and Medical Imaging Conference (NSS/MIC), 2011 IEEE, pp. 4126–4130. IEEE

  24. Cui J, Pratx G, Prevrhal S, Levin C (2011) Fully 3D list-mode time-of-flight PET image reconstruction on GPUs using CUDA. Med Phys 38(12):6775

    Article  PubMed  Google Scholar 

  25. Pratx G, Surti S, Levin C (2011) Fast list-mode reconstruction for time-of-flight PET using graphics hardware. IEEE Trans Nucl Sci 58(1):105

    Article  CAS  Google Scholar 

  26. Pratx G, Chinn G, Olcott PD, Levin C (2009) Fast, accurate and shift-varying line projections for iterative reconstruction using the GPU. IEEE Trans Med Imaging 28(3):435

    Article  PubMed Central  PubMed  Google Scholar 

  27. Pratx G, Chinn G, Habte F, Olcott PD, Levin C (2006) Fully 3-D list-mode OSEM accelerated by graphics processing units. In: Nuclear Science Symposium Conference Record, 2006. IEEE, volume. 4 pp. 2196–2202. IEEE

  28. Schellmann M, Gorlatch S, Meilnder D, Ksters T, Schfers K, Wbbeling F, Burger M (2009) Parallel medical image reconstruction: from graphics processors to grids. In: Malyshkin V (ed) Proceedings of the 10th International Conference on Parallel Computing Technologies, Volume 598 of Lecture Notes in Computer Science, pp. 457–473. Springer, Berlin

  29. Felder T, Blume M, Oliver JF, Rafecas M (2009) ML-EM implementation on a GPU: avoiding simultaneous read-modify-write processes. In: Proceedings of 10th Fully 3D Meeting and 2nd HPIR Workshop, pp. 65–68

  30. Walker MD, Asselin MC, Julyan P, Feldmann M, Talbot PS, Jones T, Matthews JC (2011) Bias in iterative reconstruction of low-statistics PET data: benefits of a resolution model. Phys Med Biol 56(4):931

    Article  CAS  PubMed  Google Scholar 

  31. Qi J, Leahy RM (2006) Iterative reconstruction techniques in emission computed tomography. Phys Med Biol 51:R541

    Article  PubMed  Google Scholar 

  32. Liu X, Comtat C, Michel C, Kinahan PE, Defrise M, Townsend D (2001) Comparison of 3-D reconstruction with 3D-OSEM and with FORE+OSEM for PET. IEEE Trans Med Imaging 20(8):804 ISSN 0278–0062

    Article  CAS  PubMed  Google Scholar 

  33. Huang SC (1999) Image oscillation reduction and convergence acceleration for OSEM reconstruction PET imaging. IEEE Trans Nucl Sci 46(3):603 ISSN 0018–9499

    Article  Google Scholar 

  34. Browne J, De Pierro A (1996) A row-action alternative to the EM algorithm for maximizing likelihood in emission tomography. IEEE Trans Med Imaging 15(5):687 ISSN 0278–0062

    Article  CAS  PubMed  Google Scholar 

  35. Eklund A, Dufort P, Forsberg D, LaConte SM (2013) Medical image processing on the GPU-Past, present and future. Med Image Anal 17(8):1073

    Article  PubMed  Google Scholar 

  36. Pratx G, Xing L (2011) GPU computing in medical physics: a review. Med phys 38(5):2685

    Article  PubMed  Google Scholar 

  37. Mendes L, Ferreira N, Comtat C (2012) An overview on the multiscale/multiframe reconstruction for positron emission tomography. In: Bioengineering (ENBENG), 2012 IEEE 2nd Portuguese Meeting in, pp. 1–4

  38. Lehmann T, Gonner C, Spitzer K (1999) Survey: interpolation methods in medical image processing. IEEE Trans Med Imaging 18(11):1049

    Article  CAS  PubMed  Google Scholar 

  39. Ruijters D, ter Haar Romeny B, Suetens P (2008) Efficient GPU-based texture interpolation using uniform B-splines. Suetens J Gr GPU Game Tools 13(4):61

    Article  Google Scholar 

  40. Siddon RL (1985) Fast calculation of the exact radiological path for a three-dimensional CT array. Med Phys 12(2):252

    Article  CAS  PubMed  Google Scholar 

  41. Barrett HH, White T, Parra L (1997) List-mode likelihood. JOSA A 14(11):2914

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  42. Reader AJ, Erlandsson K, Flower MA, Ott RJ (1998) Fast accurate iterative reconstruction for low-statistics positron volume imaging. Phys Med Biol 43:835

    Article  CAS  PubMed  Google Scholar 

  43. Huesman R, Klein G, Moses W, Qi J, Reutter B, Virador PRG (2000) List-mode maximum-likelihood reconstruction applied to positron emission mammography (PEM) with irregular sampling. IEEE Trans Med Imaging 19(5):532

    Article  CAS  PubMed  Google Scholar 

  44. Gaitanis A, Kontaxakis G, Spyrou G, Panayiotakis G, Tzanakos G (2010) PET image reconstruction: a stopping rule for the MLEM algorithm based on properties of the updating coefficients. Comput Med Imaging Gr 34(2):131

    Article  Google Scholar 

  45. Bissantz N, Mair BA, Munk A (2006) A multi-scale stopping criterion for MLEM reconstructions in PET. In: Nuclear Science Symposium Conference Record, 2006. IEEE, volume 6, pp. 3376–3379

  46. Llacer J, Veklerov E (1989) Feasible images and practical stopping rules for iterative algorithms in emission tomography. IEEE Trans Med Imaging 8(2):186

    Article  CAS  PubMed  Google Scholar 

  47. Parker J, Kenyon R, Troxel D (1983) Comparison of interpolating methods for image resampling. IEEE Trans Med Imaging 2(1):31

    Article  CAS  PubMed  Google Scholar 

  48. Kastis GA, Gaitanis A, Fernandez Y, Kontaxakis G, Fokas AS (2010) Evaluation of a spline reconstruction technique: Comparison with FBP, MLEM and OSEM. In: Nuclear Science Symposium Conference Record (NSS/MIC), 2010 IEEE, pp. 3282–3287 IEEE

  49. Tang J, Rahmim A (2009) Bayesian PET image reconstruction incorporating anato-functional joint entropy. Phys Med Biol 54(23):7063

    Article  PubMed Central  PubMed  Google Scholar 

  50. Chang JH, Anderson JM, Votaw J (2004) Regularized image reconstruction algorithms for positron emission tomography. IEEE Trans Med Imaging 23(9):1165

    Article  PubMed  Google Scholar 

  51. Kadrmas DJ (2001) Statistically regulated and adaptive EM reconstruction for emission computed tomography. IEEE Trans Nucl Sci 48(3):790

    Article  Google Scholar 

  52. Wilson DW, Tsui BM, Barrett HH (1994) Noise properties of the EM algorithm. II. Monte Carlo simulations. Phys Med Biol 39(5):847

    Article  CAS  PubMed  Google Scholar 

  53. Hebert T, Leahy R (1989) A generalized EM algorithm for 3-D Bayesian reconstruction from Poisson data using Gibbs priors. IEEE Trans Med Imaging 8(2):194

    Article  CAS  PubMed  Google Scholar 

  54. Nassiri MA, Hissoiny S, Carrier JF, Després P (2011) Fast GPU-based computation of the sensitivity matrix for a PET list-mode OSEM algorithm. In: Proceedings of the 11th International Meeting on Fully Three-Dimensional Image Reconstruction in Radiology and Nuclear Medicine, 11-15 July. Potsdam, Germany

  55. Jan S et al (2004) GATE: a simulation toolkit for PET and SPECT. Phys Med Biol 49(19):4543

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  56. Lamare F, Turzo A, Bizais Y, Rest CCL, Visvikis D (2006) Validation of a Monte Carlo simulation of the Philips Allegro/GEMINI PET systems using GATE. Phys Med Biol 51(4):943

    Article  CAS  PubMed  Google Scholar 

  57. McLennan A, Reilhac A, Brady M (2009) SORTEO: Monte carlo-based simulator with list-mode capabilities. In: Engineering in Medicine and Biology Society, 2009. EMBC 2009. Annual International Conference of the IEEE, pp. 3751–3754. ISSN 1557–170X

  58. De Beenhouwer J, Staelens S, Kruecker D, Ferrer L, DAsseler Y, Lemahieu I, Rannou FR (2007) Cluster computing software for GATE simulations. Med Phys 34(6):1926

    Article  PubMed  Google Scholar 

  59. National Electrical Manufacturers Association, NEMA NU 2–2007. Performance measurements of positron emission tomographs (2007)

  60. Cui J, Pratx G, Meng B, Levin C (2013) Distributed MLEM: an iterative tomographic image reconstruction algorithm for distributed memory architectures. IEEE Trans Med Imaging 32(5):957 ISSN 0278–0062

    Article  CAS  PubMed  Google Scholar 

  61. Wang W, Hu Z, Gualtieri E, Parma M, Walsh E, Sebok D, Hsieh YL, Tung C, Song X, Griesmer J et al (2006) Systematic and distributed time-of-flight list mode PET reconstruction. In: Nuclear Science Symposium Conference Record, 2006. IEEE, volume 3, pp. 1715–1722. IEEE

  62. Hissoiny S, Ozell B, Bouchard H, Desprs P (2011) GPUMCD: A new GPU-oriented Monte Carlo dose calculation platform. Med Phys 38(2):754

  63. Varrone A, Sjholm N, Eriksson L, Gulys B, Halldin C, Farde L (2009) Advancement in PET quantification using 3D-OP-OSEM point spread function reconstruction with the HRRT. Eur J Nucl Med Mol Imaging 36:1639

    Article  PubMed  Google Scholar 

Download references

Acknowledgments

This work was supported by the Fonds de recherche du Québec – Nature et technologies (FRQ-NT) and by the Natural Sciences and Engineering Research Council of Canada (NSERC). NVIDIA Corporation kindly donated material to conduct this study.

Author information

Authors and Affiliations

  1. Département de radio-oncologie, Centre hospitalier de l’Université de Montréal (CHUM), Montréal, QC, Canada

    Moulay Ali Nassiri & Jean-François Carrier

  2. Département de radio-oncologie and Centre de recherche du CHU de Québec, Québec, QC, Canada

    Philippe Després

  3. Département de physique, de génie physique et d’optique and Centre de recherche sur le cancer, Université Laval, Québec, QC, Canada

    Philippe Després

Authors
  1. Moulay Ali Nassiri
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Jean-François Carrier
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Philippe Després
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Moulay Ali Nassiri.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Nassiri, M.A., Carrier, JF. & Després, P. Fast GPU-based computation of spatial multigrid multiframe LMEM for PET. Med Biol Eng Comput 53, 791–803 (2015). https://doi.org/10.1007/s11517-015-1284-9

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11517-015-1284-9

Keywords

Search

Navigation