Skip to main content
Key Specialties
Cardiology Diabetology Endocrinology Gastroenterology Geriatrics and Gerontology Gynecology and Obstetrics Hematology Infectious Disease Internal Medicine Nephrology Neurology Oncology Pediatrics Respiratory Medicine Rheumatology Surgery
Congress Coverage
2024 ACC Congress 2023 ESMO Congress 2023 EASD Congress 2023 ERS Congress 2023 ESC Congress 2023 EHA Congress 2023 ASCO Annual Meeting
Clinical Collections
Advances in Alzheimer’s

ACC 2024 Congress

Catch up on all the top stories and latest evidence in cardiology research with our ACC 2024 Congress hub page.
ADVANCED SEARCH
Log in
Top
EJNMMI Research
Published in: EJNMMI Research 1/2020

Open Access 01-12-2020 | Positron Emission Tomography | Original research

Convolutional neural networks for improving image quality with noisy PET data

Authors: Josh Schaefferkoetter, Jianhua Yan, Claudia Ortega, Andrew Sertic, Eli Lechtman, Yael Eshet, Ur Metser, Patrick Veit-Haibach

Published in: EJNMMI Research | Issue 1/2020

Login to get access

Abstract

Goal

PET is a relatively noisy process compared to other imaging modalities, and sparsity of acquisition data leads to noise in the images. Recent work has focused on machine learning techniques to improve PET images, and this study investigates a deep learning approach to improve the quality of reconstructed image volumes through denoising by a 3D convolution neural network. Potential improvements were evaluated within a clinical context by physician performance in a reading task.

Methods

A wide range of controlled noise levels was emulated from a set of chest PET data in patients with lung cancer, and a convolutional neural network was trained to denoise the reconstructed images using the full-count reconstructions as the ground truth. The benefits, over conventional Gaussian smoothing, were quantified across all noise levels by observer performance in an image ranking and lesion detection task.

Results

The CNN-denoised images were generally ranked by the physicians equal to or better than the Gaussian-smoothed images for all count levels, with the largest effects observed in the lowest-count image sets. For the CNN-denoised images, overall lesion contrast recovery was 60% and 90% at the 1 and 20 million count levels, respectively. Notwithstanding the reduced lesion contrast recovery in noisy data, the CNN-denoised images also yielded better lesion detectability in low count levels. For example, at 1 million true counts, the average true positive detection rate was around 40% for the CNN-denoised images and 30% for the smoothed images.

Conclusion

Significant improvements were found for CNN-denoising for very noisy images, and to some degree for all noise levels. The technique presented here offered however limited benefit for detection performance for images at the count levels routinely encountered in the clinic.
Literature
1.
Schaefferkoetter JD, et al. Initial assessment of image quality for low-dose PET: evaluation of lesion detectability. Physics in Medicine and Biology. 2015;60(14):5543.PubMedCrossRef
2.
3.
Townsend D. Physical principles and technology of clinical PET imaging. Annal Acad Med Singapore. 2004;33(2):133–45.
4.
Nuyts, J., et al. Maximum-likelihood reconstruction based on a modified Poisson distribution to reduce bias in PET. in Nuclear Science Symposium and Medical Imaging Conference (NSS/MIC), IEEE 2011. Valencia, Spain.
5.
Hong I, et al. Complementary frame reconstruction: a low-biased dynamic PET technique for low count density data in projection space. Physics in medicine and biology. 2014;59(18):5441.PubMedCrossRef
6.
Conti M, et al. First experimental results of time-of-flight reconstruction on an LSO PET scanner. Physics in Medicine and Biology. 2005;50:4507–26.PubMedCrossRef
7.
Panin VY, et al. Fully 3-D PET reconstruction with system matrix derived from point source measurements. Med Imag IEEE Transact. 2006;25(7):907–21.CrossRef
8.
Ren S, et al. Faster r-cnn: towards real-time object detection with region proposal networks. In: Advances in neural information processing systems; 2015.
9.
Dai J, et al. R-fcn: Object detection via region-based fully convolutional networks. In: Advances in neural information processing systems; 2016.
10.
Krizhevsky A, Sutskever I, Hinton GE. Imagenet classification with deep convolutional neural networks. In: Advances in neural information processing systems; 2012.
11.
CireşAn D, et al. Multi-column deep neural network for traffic sign classification. Neural networks. 2012;32:333–8.PubMedCrossRef
12.
Long J, Shelhamer E, Darrell T. Fully convolutional networks for semantic segmentation. in Proceedings of the IEEE conference on computer vision and pattern recognition; 2015.
13.
Noh H, Hong S, Han B. Learning deconvolution network for semantic segmentation. in Proceedings of the IEEE international conference on computer vision; 2015.
14.
Hannun, A., et al., Deep speech: scaling up end-to-end speech recognition. arXiv preprint arXiv:1412.5567, 2014.
15.
Schwenck J, et al. Comparison of 68Ga-labelled PSMA-11 and 11C-choline in the detection of prostate cancer metastases by PET/CT. European journal of nuclear medicine and molecular imaging. 2017;44(1):92–101.PubMedCrossRef
16.
Ledig, C., et al. Photo-realistic single image super-resolution using a generative adversarial network. in Proceedings of the IEEE conference on computer vision and pattern recognition. 2017.
17.
Ronneberger, O., P. Fischer, and T. Brox. U-net: Convolutional networks for biomedical image segmentation. in International Conference on Medical image computing and computer-assisted intervention. 2015. Springer.
18.
Milletari, F., N. Navab, and S.-A. Ahmadi. V-net: Fully convolutional neural networks for volumetric medical image segmentation. in 2016 Fourth International Conference on 3D Vision (3DV). 2016. IEEE.
19.
Wang, S., et al. Accelerating magnetic resonance imaging via deep learning. in 2016 IEEE 13th International Symposium on Biomedical Imaging (ISBI). 2016. IEEE.
20.
Kang E, Min J, Ye JC. A deep convolutional neural network using directional wavelets for low-dose X-ray CT reconstruction. Medical physics. 2017;44(10):e360–75.PubMedCrossRef
21.
22.
Cireşan DC, et al. Mitosis detection in breast cancer histology images with deep neural networks. in International Conference on Medical Image Computing and Computer-assisted Intervention. Springer: Berlin, Heidelberg. 2013;411–18.
23.
Liu C-C, Qi J. Higher SNR PET image prediction using a deep learning model and MRI image. Physics in Medicine & Biology. 2019;64(11):115004.CrossRef
24.
Lu W, et al. An investigation of quantitative accuracy for deep learning based denoising in oncological PET. Physics in Medicine & Biology. 2019;64(16):165019.CrossRef
25.
Cui J, et al. PET image denoising using unsupervised deep learning. European journal of nuclear medicine and molecular imaging. 2019;46(13):2780–9.PubMedCrossRef
26.
Gong K, et al. Pet image denoising using a deep neural network through fine tuning. IEEE Transact Radiation Plasma Med Sci. 2018;3(2):153–61.CrossRef
27.
Fu J, et al. Deep learning approaches using 2D and 3D convolutional neural networks for generating male pelvic synthetic computed tomography from magnetic resonance imaging. Medical Physics. 2019;46(9):3788–98.PubMedCrossRef
28.
Ulyanov D, Vedaldi A, Lempitsky V. Instance normalization: the missing ingredient for fast stylization. arXiv preprint arXiv:1607.08022; 2016.
29.
Schaefferkoetter J, et al. Low dose positron emission tomography emulation from decimated high statistics: a clinical validation study. Medical Physics. 2019;46(6):2638–45.
30.
Sah B-R, et al. Feasibility of 18F-FDG dose reductions in breast cancer PET/MRI. J Nucl Med. 2018;59(12):1817–22.PubMedCrossRef
31.
Schaefferkoetter JD, et al. Quantitative accuracy and lesion detectability of low-dose 18F-FDG PET for lung cancer screening. J Nucl Med. 2017;58(3):399–405.PubMedCrossRef
32.
Swensson R. Unified measurement of observer performance in detecting and localizing target objects on images. Med Phys. 1996;23:1709–25.PubMedCrossRef
33.
Hashimoto F, et al. Dynamic PET image denoising using deep convolutional neural networks without prior training datasets. IEEE Access. 2019;7:96594–603.CrossRef
34.
Johnson J, Alahi A, Fei-Fei L. Perceptual losses for real-time style transfer and super-resolution. in European conference on computer vision. Springer: Cham. 2016;694–711.
35.
Yang, Q., et al., CT image denoising with perceptive deep neural networks. arXiv preprint arXiv:1702.07019, 2017.
36.
Schaefferkoetter J, et al. An investigation of a protocol for PET/CT screening of patients at risk of lung cancer. J Nucl Med. 2018;59(supplement 1):1354.
37.
Leynes AP, et al. Zero-echo-time and Dixon deep pseudo-CT (ZeDD CT): direct generation of pseudo-CT images for pelvic PET/MRI attenuation correction using deep convolutional neural networks with multiparametric MRI. Journal of Nuclear Medicine. 2018;59(5):852–8.PubMedPubMedCentralCrossRef
38.
Ye JC, Han Y, Cha E. Deep convolutional framelets: a general deep learning framework for inverse problems. SIAM Journal on Imaging Sciences. 2018;11(2):991–1048.CrossRef
39.
Dabov K, et al. Image denoising with block-matching and 3D filtering. in Image processing: algorithms and systems, neural networks, and machine Learning. International Society for Optics and Photonics. 2006;6064:606414.
40.
Arabi H, Zaidi H. Spatially guided nonlocal mean approach for denoising of PET images. Medical Physics. 2020;47(4):1656–69.PubMedCrossRef
Metadata
Title
Convolutional neural networks for improving image quality with noisy PET data
Authors
Josh Schaefferkoetter
Jianhua Yan
Claudia Ortega
Andrew Sertic
Eli Lechtman
Yael Eshet
Ur Metser
Patrick Veit-Haibach
Publication date
01-12-2020
Publisher
Springer Berlin Heidelberg
Published in
EJNMMI Research / Issue 1/2020
Electronic ISSN: 2191-219X
DOI
https://doi.org/10.1186/s13550-020-00695-1

Other articles of this Issue 1/2020

EJNMMI Research 1/2020 Go to the issue

Original research

Evaluation of camera-based freehand SPECT in preoperative sentinel lymph node mapping for melanoma patients

Original research

Development and validation of a prognostic model incorporating [18F]FDG PET/CT radiomics for patients with minor salivary gland carcinoma

Original research

Application of the PET ligand [11C]ORM-13070 to examine receptor occupancy by the α2C-adrenoceptor antagonist ORM-12741: translational validation of target engagement in rat and human brain

Original research

Brown adipose tissue uptake of triglyceride-rich lipoprotein-derived fatty acids in diabetic or obese mice under different temperature conditions

Original research

A preclinical PET dual-tracer imaging protocol for ER and HER2 phenotyping in breast cancer xenografts

Original research

FDG PET versus CT radiomics to predict outcome in malignant pleural mesothelioma patients