Accelerated SPECT image reconstruction with FBP and an image enhancement convolutional neural network
Tóm tắt
Monte Carlo-based iterative reconstruction to correct for photon scatter and collimator effects has been proven to be superior over analytical correction schemes in single-photon emission computed tomography (SPECT/CT), but it is currently not commonly used in daily clinical practice due to the long associated reconstruction times. We propose to use a convolutional neural network (CNN) to upgrade fast filtered back projection (FBP) image quality so that reconstructions comparable in quality to the Monte Carlo-based reconstruction can be obtained within seconds. A total of 128 technetium-99m macroaggregated albumin pre-treatment SPECT/CT scans used to guide hepatic radioembolization were available. Four reconstruction methods were compared: FBP, clinical reconstruction, Monte Carlo-based reconstruction, and the neural network approach. The CNN generated reconstructions in 5 sec, whereas clinical reconstruction took 5 min and the Monte Carlo-based reconstruction took 19 min. The mean squared error of the neural network approach in the validation set was between that of the Monte Carlo-based and clinical reconstruction, and the lung shunting fraction difference was lower than 2 percent point. A phantom experiment showed that quantitative measures required in radioembolization were accurately retrieved from the CNN-generated reconstructions. FBP with an image enhancement neural network provides SPECT reconstructions with quality close to that obtained with Monte Carlo-based reconstruction within seconds.
Tài liệu tham khảo
Cherry SR, Sorenson JA, Phelps ME. Tomographic reconstruction in nuclear medicine. In: Physics in Nuclear Medicine. 4th ed: Saunders; 2012. p. 270–1.
citation_journal_title=Phys Med Biol; citation_title=Monte Carlo-based SPECT reconstruction within the SIMIND framework; citation_author=J Gustafsson, G Brolin, M Ljungberg; citation_volume=63; citation_publication_date=2018; citation_pages=245012; citation_doi=10.1088/1361-6560/aaf0f1; citation_id=CR2
citation_journal_title=Med Phys; citation_title=Improved quantitative 90Y bremsstrahlung SPECT/CT reconstruction with Monte Carlo scatter modeling; citation_author=YK Dewaraja, SY Chun, RN Srinivasa; citation_volume=44; citation_issue=12; citation_publication_date=2017; citation_pages=6364-6376; citation_doi=10.1002/mp.12597; citation_id=CR3
citation_journal_title=Med Image Anal.; citation_title=A survey on deep learning in medical image analysis; citation_author=G Litjens, T Kooi, BE Bejnordi; citation_volume=42; citation_publication_date=2017; citation_pages=60-88; citation_doi=10.1016/j.media.2017.07.005; citation_id=CR4
citation_journal_title=Am J Clin Oncol Cancer Clin Trials.; citation_title=Radioembolization for the treatment of liver tumors: general principles; citation_author=A Kennedy, D Coldwell, B Sangro, H Wasan, R Salem; citation_volume=35; citation_issue=1; citation_publication_date=2012; citation_pages=91-99; citation_doi=10.1097/COC.0b013e3181f47583; citation_id=CR5
citation_journal_title=Eur J Nucl Med Mol Imaging.; citation_title=Same-day 90Y radioembolization: implementing a new treatment paradigm; citation_author=A Gabr, JR Kallini, VL Gates; citation_volume=43; citation_issue=13; citation_publication_date=2016; citation_pages=2353-2359; citation_doi=10.1007/s00259-016-3438-x; citation_id=CR6
citation_journal_title=Radiology; citation_title=A dual-layer detector for simultaneous fluoroscopic and nuclear imaging; citation_author=S Velden, B Kunnen, WJC Koppert; citation_volume=290; citation_issue=3; citation_publication_date=2019; citation_pages=833-838; citation_doi=10.1148/radiol.2018180796; citation_id=CR7
citation_journal_title=Phys Med Biol.; citation_title=Performance of a dual-layer scanner for hybrid SPECT/CBCT; citation_author=MMA Dietze, B Kunnen, S Velden, JHL Steenbergen, WJC Koppert, MA Viergever, HWAM Jong; citation_volume=64; citation_publication_date=2019; citation_pages=105020; citation_doi=10.1088/1361-6560/ab15f6; citation_id=CR8
Dietze MMA, Bastiaannet R, Kunnen B, Van der Velden S, Lam MGEH, Viergever MA, De Jong HWAM. Respiratory motion compensation in interventional liver SPECT using simultaneous fluoroscopic and nuclear imaging. Med Phys. 2019.
citation_journal_title=IEEE Trans Nucl Sci.; citation_title=Acceleration of Monte Carlo SPECT simulation using convolution-based forced detection; citation_author=HWAM Jong, ETP Slijpen, FJ Beekman; citation_volume=48; citation_publication_date=2001; citation_pages=58-64; citation_doi=10.1109/23.910833; citation_id=CR10
citation_journal_title=EJNMMI Phys; citation_title=Fast quantitative reconstruction with focusing collimators for liver SPECT; citation_author=MMA Dietze, S Velden, MGEH Lam, MA Viergever, HWAM Jong; citation_volume=5; citation_issue=1; citation_publication_date=2018; citation_pages=28; citation_doi=10.1186/s40658-018-0228-5; citation_id=CR11
citation_journal_title=J Nucl Med.; citation_title=Evaluation of 3D Monte Carlo-based scatter correction for 99mTc cardiac perfusion SPECT; citation_author=J Xiao, TC Wit, SG Staelens, FJ Beekman; citation_volume=47; citation_issue=10; citation_publication_date=2006; citation_pages=1662-1669; citation_id=CR12
citation_journal_title=J Nucl Med.; citation_title=Quantitative Monte Carlo-based 90Y SPECT reconstruction; citation_author=M Elschot, MGEH Lam, MAAJ Bosch, MA Viergever, HWAM Jong; citation_volume=54; citation_issue=9; citation_publication_date=2013; citation_pages=1557-1563; citation_doi=10.2967/jnumed.112.119131; citation_id=CR13
citation_journal_title=Med Phys.; citation_title=Quantitative Monte Carlo-based holmium-166 SPECT reconstruction; citation_author=M Elschot, MLJ Smits, JFW Nijsen; citation_volume=40; citation_issue=11; citation_publication_date=2013; citation_pages=112502; citation_doi=10.1118/1.4823788; citation_id=CR14
citation_journal_title=IEEE Trans Nucl Sci.; citation_title=A method for attenuation correction in radionuclide computed tomography; citation_author=LT Chang; citation_volume=25; citation_issue=1; citation_publication_date=1978; citation_pages=638-643; citation_doi=10.1109/TNS.1978.4329385; citation_id=CR15
citation_journal_title=J Nucl Med.; citation_title=Improved SPECT quantification using compensation for scattered photons; citation_author=RJ Jaszczak, KL Greer, CE Floyd, CC Harris, RE Coleman; citation_volume=25; citation_issue=8; citation_publication_date=1984; citation_pages=893-900; citation_id=CR16
citation_journal_title=IEEE Signal Processing Magazine; citation_title=Convolutional Neural Networks for Inverse Problems in Imaging: A Review; citation_author=Michael T. McCann, Kyong Hwan Jin, Michael Unser; citation_volume=34; citation_issue=6; citation_publication_date=2017; citation_pages=85-95; citation_doi=10.1109/MSP.2017.2739299; citation_id=CR17
Ronneberger O, Fischer P, Brox T. U-net: Convolutional networks for biomedical image segmentation. MICCAI. 2015:234–41.
Kingma DP, Ba JL. ADAM: a method for stochastic optimization: ICLR; 2015.
Abadi M, Barham P, Chen J, et al. TensorFlow: a system for large-scale machine learning. Oper Syst Des Implement. 2016:265–84.
Chollet, F. K. 2015.
https://github.com/fchollet/keras
. Accessed 1 Mar 2019.
citation_journal_title=J Nucl Med.; citation_title=Safety of a scout dose preceding hepatic radioembolization with 166Ho microspheres; citation_author=JF Prince, R Rooij, GH Bol, HWAM Jong, MAAJ Bosch, MGEH Lam; citation_volume=56; citation_issue=6; citation_publication_date=2015; citation_pages=817-823; citation_doi=10.2967/jnumed.115.155564; citation_id=CR22
citation_journal_title=J Nucl Med.; citation_title=Insights into the dose-response relationship of radioembolization with Resin 90Y-microspheres: a prospective cohort study in patients with colorectal cancer liver metastases; citation_author=AF Hover, CENM Rosenbaum, SG Elias, HWAM Jong, M Koopman, HM Verkooijen, A Alavi, MAAJ Bosch, MGEH Lam; citation_volume=57; citation_issue=7; citation_publication_date=2016; citation_pages=1014-1019; citation_doi=10.2967/jnumed.115.166942; citation_id=CR23
Xu J, Gong E, Pauly J, Zaharchuk G. 200x low-dose PET reconstruction using deep learning. 2017.
http://arxiv.org/abs/1712.04119
. Accessed 1 Mar 2019.
Jiao J, Ourselin S. Fast PET reconstruction using multi-scale fully convolutional neural networks. 2017.
http://arxiv.org/abs/1704.07244
. Accessed 1 Mar 2019.
citation_journal_title=EJNMMI Phys.; citation_title=Fast GPU-based Monte Carlo code for SPECT/CT reconstructions generates improved 177Lu images; citation_author=T Rydén, J Heydorn Lagerlöf, J Hemmingsson; citation_volume=5; citation_issue=1; citation_publication_date=2018; citation_pages=1; citation_doi=10.1186/s40658-017-0201-8; citation_id=CR26
Van der Velden S, Dietze MMA, Viergever MA, de Jong HWAM. Fast technetium-99m liver SPECT for evaluation of the pretreatment procedure for radioembolization dosimetry. Med Phys. 2018;46(1):345–55.