Deep Learning-based PET Image Denoising and Reconstruction: a Review

Overview

Journal Radiol Phys Technol

Specialties Biophysics
Biotechnology
Radiology

Date 2024 Feb 6

PMID 38319563

Authors

Fumio Hashimoto

Yuya Onishi

Kibo Ote

Hideaki Tashima

Andrew J Reader

Taiga Yamaya

Affiliations

Soon will be listed here.

Abstract

This review focuses on positron emission tomography (PET) imaging algorithms and traces the evolution of PET image reconstruction methods. First, we provide an overview of conventional PET image reconstruction methods from filtered backprojection through to recent iterative PET image reconstruction algorithms, and then review deep learning methods for PET data up to the latest innovations within three main categories. The first category involves post-processing methods for PET image denoising. The second category comprises direct image reconstruction methods that learn mappings from sinograms to the reconstructed images in an end-to-end manner. The third category comprises iterative reconstruction methods that combine conventional iterative image reconstruction with neural-network enhancement. We discuss future perspectives on PET imaging and deep learning technology.

Citing Articles

Arrival time mapping with O-gas PET for cerebrovascular steno-occlusive diseases: a comparative study with CT perfusion.

Ibaraki M, Shinohara Y, Watanabe A, Sato K, Ohmura T, Yamamoto H EJNMMI Res. 2025; 15(1):20.

PMID: 40055252 PMC: 11889313. DOI: 10.1186/s13550-025-01209-7.

Deep learning image enhancement algorithms in PET/CT imaging: a phantom and sarcoma patient radiomic evaluation.

Bonney L, Kalisvaart G, van Velden F, Bradley K, Hassan A, Grootjans W Eur J Nucl Med Mol Imaging. 2025; .

PMID: 40014074 DOI: 10.1007/s00259-025-07149-7.

Fundamental study on improving the quality of X-ray fluorescence computed tomography images by applying deep image prior to projection images as a pre-denoising method.

Kusakari S, Sato K, Tsushima Y, Matsuoka M, Sasaya T, Sunaguchi N Int J Comput Assist Radiol Surg. 2024; .

PMID: 39739291 DOI: 10.1007/s11548-024-03307-8.

The impact of long axial field of view (LAFOV) PET on oncologic imaging.

Cook G, Alberts I, Wagner T, Fischer B, Nazir M, Lilburn D Eur J Radiol. 2024; 183:111873.

PMID: 39647272 PMC: 11904125. DOI: 10.1016/j.ejrad.2024.111873.

Whole-body PET image denoising for reduced acquisition time.

Kruzhilov I, Kudin S, Vetoshkin L, Sokolova E, Kokh V Front Med (Lausanne). 2024; 11:1415058.

PMID: 39403284 PMC: 11471667. DOI: 10.3389/fmed.2024.1415058.

References

Obi T, Matej S, Lewitt R, Herman G . 2.5-D simultaneous multislice reconstruction by series expansion methods from Fourier-rebinned PET data. IEEE Trans Med Imaging. 2000; 19(5):474-84. DOI: 10.1109/42.870257. View

Wang Y, Zhou L, Yu B, Wang L, Zu C, Lalush D . 3D Auto-Context-Based Locality Adaptive Multi-Modality GANs for PET Synthesis. IEEE Trans Med Imaging. 2018; 38(6):1328-1339. PMC: 6541547. DOI: 10.1109/TMI.2018.2884053. View

Gong K, Kim K, Cui J, Wu D, Li Q . The Evolution of Image Reconstruction in PET: From Filtered Back-Projection to Artificial Intelligence. PET Clin. 2021; 16(4):533-542. DOI: 10.1016/j.cpet.2021.06.004. View

Hashimoto F, Ote K . ReconU-Net: a direct PET image reconstruction using U-Net architecture with back projection-induced skip connection. Phys Med Biol. 2024; 69(10). DOI: 10.1088/1361-6560/ad40f6. View

Sano A, Nishio T, Masuda T, Karasawa K . Denoising PET images for proton therapy using a residual U-net. Biomed Phys Eng Express. 2021; 7(2). DOI: 10.1088/2057-1976/abe33c. View

Li Y, Hu J, Sari H, Xue S, Ma R, Kandarpa S . A deep neural network for parametric image reconstruction on a large axial field-of-view PET. Eur J Nucl Med Mol Imaging. 2022; 50(3):701-714. DOI: 10.1007/s00259-022-06003-4. View

Haggstrom I, Schmidtlein C, Campanella G, Fuchs T . DeepPET: A deep encoder-decoder network for directly solving the PET image reconstruction inverse problem. Med Image Anal. 2019; 54:253-262. PMC: 6537887. DOI: 10.1016/j.media.2019.03.013. View

Wilson D, Tsui B, Barrett H . Noise properties of the EM algorithm: II. Monte Carlo simulations. Phys Med Biol. 1994; 39(5):847-71. DOI: 10.1088/0031-9155/39/5/005. View

Gong K, Catana C, Qi J, Li Q . Direct Reconstruction of Linear Parametric Images From Dynamic PET Using Nonlocal Deep Image Prior. IEEE Trans Med Imaging. 2021; 41(3):680-689. PMC: 8956450. DOI: 10.1109/TMI.2021.3120913. View

10.

Qi J, Leahy R, Cherry S, Chatziioannou A, Farquhar T . High-resolution 3D Bayesian image reconstruction using the microPET small-animal scanner. Phys Med Biol. 1998; 43(4):1001-13. DOI: 10.1088/0031-9155/43/4/027. View

11.

Reader A, Schramm G . Artificial Intelligence for PET Image Reconstruction. J Nucl Med. 2021; 62(10):1330-1333. DOI: 10.2967/jnumed.121.262303. View

12.

Suzuki K . Overview of deep learning in medical imaging. Radiol Phys Technol. 2017; 10(3):257-273. DOI: 10.1007/s12194-017-0406-5. View

13.

Wang B, Liu H . FBP-Net for direct reconstruction of dynamic PET images. Phys Med Biol. 2020; 65(23). DOI: 10.1088/1361-6560/abc09d. View

14.

Liu C, Qi J . Higher SNR PET image prediction using a deep learning model and MRI image. Phys Med Biol. 2019; 64(11):115004. PMC: 7413624. DOI: 10.1088/1361-6560/ab0dc0. View

15.

Onishi Y, Hashimoto F, Ote K, Ota R . Unbiased TOF estimation using leading-edge discriminator and convolutional neural network trained by single-source-position waveforms. Phys Med Biol. 2022; 67(4). DOI: 10.1088/1361-6560/ac508f. View

16.

Mehranian A, Wollenweber S, Walker M, Bradley K, Fielding P, Su K . Image enhancement of whole-body oncology [F]-FDG PET scans using deep neural networks to reduce noise. Eur J Nucl Med Mol Imaging. 2021; 49(2):539-549. PMC: 8803788. DOI: 10.1007/s00259-021-05478-x. View

17.

Kamasak M, Bouman C, Morris E, Sauer K . Direct reconstruction of kinetic parameter images from dynamic PET data. IEEE Trans Med Imaging. 2005; 24(5):636-50. DOI: 10.1109/TMI.2005.845317. View

18.

Gong K, Eric Berg , Cherry S, Qi J . Machine Learning in PET: from Photon Detection to Quantitative Image Reconstruction. Proc IEEE Inst Electr Electron Eng. 2023; 108(1):51-68. PMC: 10691821. DOI: 10.1109/JPROC.2019.2936809. View

19.

Fu Y, Dong S, Niu M, Xue L, Guo H, Huang Y . AIGAN: Attention-encoding Integrated Generative Adversarial Network for the reconstruction of low-dose CT and low-dose PET images. Med Image Anal. 2023; 86:102787. DOI: 10.1016/j.media.2023.102787. View

20.

Hashimoto F, Ohba H, Ote K, Kakimoto A, Tsukada H, Ouchi Y . 4D deep image prior: dynamic PET image denoising using an unsupervised four-dimensional branch convolutional neural network. Phys Med Biol. 2020; 66(1):015006. DOI: 10.1088/1361-6560/abcd1a. View