Artifacts correction algorithm for iodine-131 SPECT planar imaging

doi:10.16511/j.cnki.qhdxxb.2022.26.060

Download: PDF(5468 KB) HTML
Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks

Abstract [Objective] Iodine-131 SPECT (single photon emission computed tomography) planar imaging has been widely used in the clinical diagnosis and treatment evaluation of thyroid cancer. Because of the high-energy emissions of iodine-131, the photons have a high probability of penetrating the collimator septa of SPECT, causing "spoke" artifacts in the final result. The "spoke" artifacts make it difficult to distinguish accurate concentration regions of iodine-131, and they may obscure lower uptake regions nearby, such as metastatic spread to lymph nodes. In this paper, a deconvolution method based on the point spread function was combined with a priori regularization to suppress the spoke artifacts and to improve the diagnostic accuracy in clinical studies.[Methods] This study is based on the NET632 SPECT system with the corresponding high-energy general purpose collimator. The collected data follow the Poisson distribution, and they consist of two parts, a forward projection of the true activity distribution and the scatter data. The forward projection progress can be well modeled using a shift-invariant PSF (point spread function). An objective function is built based on the aforementioned approximation, and a priori function is introduced to regularize the reconstruction. A monotonic and convergent algorithm is derived to iteratively solve the objective function. In contrast, the conventional deconvolution method regularizes the solution using a total variation term, and the objective function is optimized based on the "one-step-late" algorithm; thus, nonnegativity and convergence can not be guaranteed. The triple-energy window method is employed to estimate scattering data, and PSFs of different sizes are generated based on Monte Carlo simulations. Simulated NEMA torso phantom data are reconstructed with different parameters to validate the monotonicity and convergence of the proposed method. Moreover, the dataset is also used to evaluate the effects of PSF size and regularization strength on reconstruction images. Normalized spoke counts and background noise are calculated for quantitative comparison. Simulation data are also used to compare the reconstruction performance of the proposed method and the conventional deconvolution method. With the optimized parameters determined by simulation data, the proposed method is further validated by clinical point data and volunteer data.[Results] With different reconstruction parameters, the objective function value increased monotonically, and the image differences between two adjacent iterations rapidly reduced to a value close to zero. The simulation study also demonstrated that a 127?27 PSF size could provide performance similar to a 255?55 PSF size, which was significantly better than 63?3 and 31?1 PSF sizes. A study on different regularization strengths suggested an optimized regularization parameter, 0.01. When the PSF size and regularization parameter were set as 127?27 and 0.01, the mean spoke counts could be reduced to 4% of the original value with a low background noise level. A comparison study based on simulation data showed the superiority of the proposed method over the conventional deconvolution method. The clinical point data and volunteer data also validated the performance of the proposed method, and the mean spoke counts could be reduced to 35% and 28% of the original values, respectively.[Conclusions] The proposed method suppresses the spoke artifacts in iodine-131 imaging using a PSF that models the physical response, and it also introduces a priori regularization to suppress noise amplified by deconvolution. The derived algorithm can guarantee the monotonicity and convergence of the iterative reconstruction. Studies on simulation data and clinical data have demonstrated that the proposed method can achieve the desired performance and is expected to improve the diagnostic accuracy in clinical studies.

Keywords iodine-131 single photon emission computed tomography "spoke" artifacts deconvolution a priori regularization

Issue Date: 23 April 2023

	Service

	E-mail this article
	E-mail Alert
	RSS
	Articles by authors

	CHENG Li
	LIU Fan
	GAO Lilei
	LIU Hui
	LIU Yaqiang

Cite this article:

CHENG Li,LIU Fan,GAO Lilei, et al. Artifacts correction algorithm for iodine-131 SPECT planar imaging[J]. Journal of Tsinghua University(Science and Technology), 2023, 63(5): 802-810.

URL:

http://jst.tsinghuajournals.com/EN/10.16511/j.cnki.qhdxxb.2022.26.060 OR http://jst.tsinghuajournals.com/EN/Y2023/V63/I5/802

[1] AHMADZADEHFAR H, BIERSACK H J, HERRMANN K. Clinical applications of SPECT-CT[M]. Berlin:Springer, 2014.
[2] 金永杰, 马天予. 核医学仪器与方法[M]. 哈尔滨:哈尔滨工程大学出版社, 2010. JIN Y J, MA T Y. Nuclear medicine instruments and methods[M]. Harbin:Harbin Engineering University Press, 2010. (in Chinese)
[3] AVRAM A M. Radioiodine scintigraphy with SPECT/CT:An important diagnostic tool for thyroid cancer staging and risk stratification[J]. Journal of Nuclear Medicine Technology, 2014, 42(3):170-180.
[4] AHMED N, NIYAZ K, BORAKATI A, et al. Hybrid SPECT/CT imaging in the management of differentiated thyroid carcinoma[J]. Asian Pacific Journal of Cancer Prevention:APJCP, 2018, 19(2):303-308.
[5] BARRACK F, SCUFFHAM J, MCQUAID S. Septal penetration correction in ¹³¹I imaging following thyroid cancer treatment[J]. Physics in Medicine & Biology, 2018, 63(7):075012.
[6] DEY N, BLANC-FERAUD L, ZIMMER C, et al. Richardson-Lucy algorithm with total variation regularization for 3D confocal microscope deconvolution[J]. Microscopy Research and Technique, 2006, 69(4):260-266.
[7] LOENING A M, GAMBHIR S S. AMIDE:A free software tool for multimodality medical image analysis[J]. Molecular Imaging, 2003, 2(3):131-137.
[8] ZHAO W R, WANG X Q, LIANG Y K, et al. Initial evaluation of three novel quantiative imaging utilities from a newly developed SPECT camera in China[J]. The Journal of Nuclear Medicine, 2018, 59(S1):219.
[9] LANGE K, HUNTER D R, YANG I. Optimization transfer using surrogate objective functions[J]. Journal of Computational and Graphical Statistics, 2000, 9(1):1-20.
[10] WANG G B, QI J Y. Penalized likelihood PET image reconstruction using patch-based edge-preserving regularization[J]. IEEE Transactions on Medical Imaging, 2012, 31(12):2194-2204.
[11] HUTTON B F, BUVAT I, BEEKMAN F J. Review and current status of SPECT scatter correction[J]. Physics in Medicine & Biology, 2011, 56(14):R85-R112.
[12] DEWARAJA Y K, LJUNGBERG M, KORAL K F. Characterization of scatter and penetration using Monte Carlo simulation in ¹³¹I imaging[J]. The Journal of Nuclear Medicine, 2000, 41(1):123-130.
[13] JAN S, SANTIN G, STRUL D, et al. GATE:A simulation toolkit for PET and SPECT[J]. Physics in Medicine & Biology, 2004, 49(19):4543-4561.