Advertisement

Journal of Digital Imaging

, Volume 32, Issue 5, pp 773–778 | Cite as

Full-Dose PET Image Estimation from Low-Dose PET Image Using Deep Learning: a Pilot Study

  • Sydney KaplanEmail author
  • Yang-Ming Zhu
Article

Abstract

Positron emission tomography (PET) imaging is an effective tool used in determining disease stage and lesion malignancy; however, radiation exposure to patients and technicians during PET scans continues to draw concern. One way to minimize radiation exposure is to reduce the dose of radioactive tracer administered in order to obtain the scan. Yet, low-dose images are inherently noisy and have poor image quality making them difficult to read. This paper proposes the use of a deep learning model that takes specific image features into account in the loss function to denoise low-dose PET image slices and estimate their full-dose image quality equivalent. Testing on low-dose image slices indicates a significant improvement in image quality that is comparable to the ground truth full–dose image slices. Additionally, this approach can lower the cost of conducting a PET scan since less radioactive material is required per scan, which may promote the usage of PET scans for medical diagnosis.

Keywords

Deep learning Denoising Image estimation Low-dose PET 

Notes

Acknowledgements

The authors wish to thank many colleagues at Philips, particularly Steve Cochoff and Andriy Andreyev, for discussion and support during this study. We are also grateful to anonymous reviewers whose comments and suggestions greatly improve the paper.

References

  1. 1.
    Juweid ME, Stroobants S, Hoekstra OS, Mottaghy FM, Dietlein M, Guermazi A, Wiseman GA, Kostakoglu L, Scheidhauer K, Buck A, Naumann R, Spaepen K, Hicks RJ, Weber WA, Reske SN, Schwaiger M, Schwartz LH, Zijlstra JM, Siegel BA, Cheson BD, Imaging Subcommittee of International Harmonization Project in Lymphoma: Use of positron emission tomography for response assessment of lymphoma: Consensus of the Imaging Subcommittee of International Harmonization Project in Lymphoma. Journal of Clinical Oncology 25(5):571–578, Feb. 2007CrossRefGoogle Scholar
  2. 2.
    Avril NE, Weber WA: Monitoring response to treatment in patients utilizing PET. Radiologic Clinics of North America 43(1):189–204, Jan. 2005CrossRefGoogle Scholar
  3. 3.
    Fletcher JW, Djulbegovic B, Soares HP, Siegel BA, Lowe VJ, Lyman GH, Coleman RE, Wahl R, Paschold JC, Avril N, Einhorn LH, Suh WW, Samson D, Delbeke D, Gorman M, Shields AF: Recommendations of the use of 18F-FDG PET in oncology. The Journal of Nuclear Medicine 49(3):480–508, Mar. 2008CrossRefGoogle Scholar
  4. 4.
    Kinahan PE, Fletcher JW: PET/CT standardized uptake values (SUVs) in clinical practice and assessing response to therapy. Semin Ultrasound CT MR 31(6):496–505, Dec, 2010CrossRefGoogle Scholar
  5. 5.
    Huang B, Wai-Ming Law M, Khong P: Whole-body PET/CT scanning: estimation of radiation dose and cancer risk. Medical Physics 251(1):166–174, Apr. 2009Google Scholar
  6. 6.
    Xiang L, Qiao Y, Nie D, An L, Lin W, Wang Q, Shen D: Deep auto-context convolutional neural networks for standard-dose PET image estimation from low-dose PET/MRI. Neurocomputing 267(1):406–416, Jun. 2017CrossRefGoogle Scholar
  7. 7.
    Q. Yang, G. Wang, P. Yan, and M. K. Kalra, “CT image denoising with perceptive deep neural networks,” in The 14th International Meeting on Fully Three-Dimensional Image Reconstruction in Radiology and Nuclear Medicine, Xián China, 2017, pp. 858–863.Google Scholar
  8. 8.
    Wolterink J, Leiner T, Viergever MA, Išgum I: Generative adversarial networks for noise reduction in low-dose CT. IEEE Transactions of Medical Imaging 36(12):2536–2545, Dec. 2017CrossRefGoogle Scholar
  9. 9.
    Yang W et al.: Improving low-dose CT image using residual convolutional network. IEEE Special Section on Advanced Signal Processing Methods in Medical Imaging 5(1):24698–24705, Oct. 2017Google Scholar
  10. 10.
    Chen H et al.: Low-dose CT denoising with convolutional neural network. In: IEEE 14th International Symposium on Biomedical Imaging. Australia: Melbourne, p. 2017Google Scholar
  11. 11.
    K. Suzuki et al, “Neural network convolution (NNC) for converting ultra-low-dose to ‘virtual’ high-dose CT images,” in Machine Learning in Medical Imaging, Quebec City, Canada, 2017, pp. 334–343.Google Scholar
  12. 12.
    Jifara W et al.: Medical image denoising using convolutional neural network: a residual learning approach. The Journal of Supercomputing., 2017.  https://doi.org/10.1007/s11227-017-2080-0
  13. 13.
    J. Xu, E. Gong, J. Pauly, G. Zaharchuk, 200x low-dose PET reconstruction using deep learning, https://arxiv.org/abs/1712.04119 (last accessed Oct. 23, 2018).

Copyright information

© Society for Imaging Informatics in Medicine 2018

Authors and Affiliations

  1. 1.Philips HealthcareHighland HeightsUSA
  2. 2.Department of NeurologyWashington University School of MedicineSt. LouisUSA
  3. 3.Siemens HealthineersFlandersUSA

Personalised recommendations