Tobias Rydén1, Martijn Van Essen2, Ida Marin1, Johanna Svensson3, Peter Bernhardt4,5. 1. Department of Medical Physics and Bioengineering, Sahlgrenska University Hospital, Gothenburg, Sweden. 2. Department of Clinical Physiology, Sahlgrenska University Hospital, Gothenburg, Sweden. 3. Department of Oncology, Institution of Clinical Science, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden; and. 4. Department of Medical Physics and Bioengineering, Sahlgrenska University Hospital, Gothenburg, Sweden peter.bernhardt@gu.se. 5. Department of Radiation Physics, Institution of Clinical Science, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden.
Abstract
The aims of this study were to decrease the 177Lu-SPECT acquisition time by reducing the number of projections and to circumvent image degradation by adding deep-learning-generated synthesized projections. Methods: We constructed a deep convolutional U-net-shaped neural network for generation of synthetic intermediate projections (CUSIPs). The number of SPECT investigations was 352 for training, 37 for validation, and 15 for testing. The input was every fourth projection of 120 acquired SPECT projections, that is, 30 projections. The output was 30 synthetic intermediate projections (SIPs) per CUSIP. SPECT images were reconstructed with 120 or 30 projections, or with 120 projections when 90 SIPs were generated from 30 projections (30-120SIPs), using 3 CUSIPs. The reconstructions were performed with 2 ordered-subset expectation maximization (OSEM) algorithms: attenuation-corrected (AC) OSEM, and attenuation, scatter, and collimator response-corrected (ASCC) OSEM. The quality of the SIPs and SPECT images was quantitatively evaluated with root-mean-square error, peak signal-to-noise ratio (PSNR), and structural similarity (SSIM) index metrics. From a Jaszczak SPECT phantom, the recovery and signal-to-noise ratio (SNR) were determined. In addition, an experienced observer qualitatively assessed the SPECT image quality of the test set. Kidney activity concentrations, as determined from the different SPECT images, were compared. Results: The generated SIPs had a mean SSIM value of 0.926 (SD, 0.061). For AC-OSEM, the reconstruction with 30-120SIPs had higher SSIM (0.993 vs. 0.989, P < 0.001) and PSNR (49.5 vs. 47.2, P < 0.001) values than the reconstruction with 30 projections. ASCC-OSEM had higher SSIM and PSNR values than AC-OSEM (P < 0.001). There was a minor loss in recovery for 30-120SIPs, but SNR was clearly improved compared with 30 projections. The observer assessed 27 of 30 images reconstructed with 30 projections as having unacceptable noise levels, whereas the corresponding values were 2 of 60 for 30-120SIPs and 120 projections. Image quality did not differ significantly between 30-120SIPs and 120 projections. The kidney activity concentration was similar between the different projection sets, excepting a minor reduction of 2.5% for ASCC-OSEM 30-120SIPs. Conclusion: Adopting SIPs for sparsely acquired projections considerably recovers image quality and could allow a reduced SPECT acquisition time in clinical dosimetry protocols.
The aims of this study were to decrease the 177Lu-SPECT acquisition time by reducing the number of projections and to circumvent image degradation by adding deep-learning-generated synthesized projections. Methods: We constructed a deep convolutional U-net-shaped neural network for generation of synthetic intermediate projections (CUSIPs). The number of SPECT investigations was 352 for training, 37 for validation, and 15 for testing. The input was every fourth projection of 120 acquired SPECT projections, that is, 30 projections. The output was 30 synthetic intermediate projections (SIPs) per CUSIP. SPECT images were reconstructed with 120 or 30 projections, or with 120 projections when 90 SIPs were generated from 30 projections (30-120SIPs), using 3 CUSIPs. The reconstructions were performed with 2 ordered-subset expectation maximization (OSEM) algorithms: attenuation-corrected (AC) OSEM, and attenuation, scatter, and collimator response-corrected (ASCC) OSEM. The quality of the SIPs and SPECT images was quantitatively evaluated with root-mean-square error, peak signal-to-noise ratio (PSNR), and structural similarity (SSIM) index metrics. From a Jaszczak SPECT phantom, the recovery and signal-to-noise ratio (SNR) were determined. In addition, an experienced observer qualitatively assessed the SPECT image quality of the test set. Kidney activity concentrations, as determined from the different SPECT images, were compared. Results: The generated SIPs had a mean SSIM value of 0.926 (SD, 0.061). For AC-OSEM, the reconstruction with 30-120SIPs had higher SSIM (0.993 vs. 0.989, P < 0.001) and PSNR (49.5 vs. 47.2, P < 0.001) values than the reconstruction with 30 projections. ASCC-OSEM had higher SSIM and PSNR values than AC-OSEM (P < 0.001). There was a minor loss in recovery for 30-120SIPs, but SNR was clearly improved compared with 30 projections. The observer assessed 27 of 30 images reconstructed with 30 projections as having unacceptable noise levels, whereas the corresponding values were 2 of 60 for 30-120SIPs and 120 projections. Image quality did not differ significantly between 30-120SIPs and 120 projections. The kidney activity concentration was similar between the different projection sets, excepting a minor reduction of 2.5% for ASCC-OSEM 30-120SIPs. Conclusion: Adopting SIPs for sparsely acquired projections considerably recovers image quality and could allow a reduced SPECT acquisition time in clinical dosimetry protocols.
Authors: Michael Ljungberg; Anna Celler; Mark W Konijnenberg; Keith F Eckerman; Yuni K Dewaraja; Katarina Sjögreen-Gleisner; Wesley E Bolch; A Bertrand Brill; Frederic Fahey; Darrell R Fisher; Robert Hobbs; Roger W Howell; Ruby F Meredith; George Sgouros; Pat Zanzonico; Klaus Bacher; Carlo Chiesa; Glenn Flux; Michael Lassmann; Lidia Strigari; Stephan Walrand Journal: J Nucl Med Date: 2015-10-15 Impact factor: 10.057
Authors: Andreas Delker; Wolfgang Peter Fendler; Clemens Kratochwil; Anika Brunegraf; Astrid Gosewisch; Franz Josef Gildehaus; Stefan Tritschler; Christian Georg Stief; Klaus Kopka; Uwe Haberkorn; Peter Bartenstein; Guido Böning Journal: Eur J Nucl Med Mol Imaging Date: 2015-08-29 Impact factor: 9.236
Authors: Michael Garkavij; Mattias Nickel; Katarina Sjögreen-Gleisner; Michael Ljungberg; Tomas Ohlsson; Karin Wingårdh; Sven-Erik Strand; Jan Tennvall Journal: Cancer Date: 2010-02-15 Impact factor: 6.860
Authors: Jonathan Strosberg; Ghassan El-Haddad; Edward Wolin; Andrew Hendifar; James Yao; Beth Chasen; Erik Mittra; Pamela L Kunz; Matthew H Kulke; Heather Jacene; David Bushnell; Thomas M O'Dorisio; Richard P Baum; Harshad R Kulkarni; Martyn Caplin; Rachida Lebtahi; Timothy Hobday; Ebrahim Delpassand; Eric Van Cutsem; Al Benson; Rajaventhan Srirajaskanthan; Marianne Pavel; Jaime Mora; Jordan Berlin; Enrique Grande; Nicholas Reed; Ettore Seregni; Kjell Öberg; Maribel Lopera Sierra; Paola Santoro; Thomas Thevenet; Jack L Erion; Philippe Ruszniewski; Dik Kwekkeboom; Eric Krenning Journal: N Engl J Med Date: 2017-01-12 Impact factor: 91.245
Authors: John Violet; Price Jackson; Justin Ferdinandus; Shahneen Sandhu; Tim Akhurst; Amir Iravani; Grace Kong; Aravind Ravi Kumar; Sue Ping Thang; Peter Eu; Mark Scalzo; Declan Murphy; Scott Williams; Rodney J Hicks; Michael S Hofman Journal: J Nucl Med Date: 2018-10-05 Impact factor: 10.057
Authors: Karl D Spuhler; John Gardus; Yi Gao; Christine DeLorenzo; Ramin Parsey; Chuan Huang Journal: J Nucl Med Date: 2018-08-30 Impact factor: 10.057
Authors: Anna Sundlöv; Katarina Sjögreen-Gleisner; Johanna Svensson; Michael Ljungberg; Tomas Olsson; Peter Bernhardt; Jan Tennvall Journal: Eur J Nucl Med Mol Imaging Date: 2017-03-22 Impact factor: 9.236
Authors: Linn Hagmarker; Johanna Svensson; Tobias Rydén; Martijn van Essen; Anna Sundlöv; Katarina Sjögreen Gleisner; Peter Gjertsson; Peter Bernhardt Journal: J Nucl Med Date: 2019-03-22 Impact factor: 10.057
Authors: Katarina Sjögreen Gleisner; Nicolas Chouin; Pablo Minguez Gabina; Francesco Cicone; Silvano Gnesin; Caroline Stokke; Mark Konijnenberg; Marta Cremonesi; Frederik A Verburg; Peter Bernhardt; Uta Eberlein; Jonathan Gear Journal: Eur J Nucl Med Mol Imaging Date: 2022-03-14 Impact factor: 10.057