Skip to main content
Log in

Deep learning–based denoising algorithm in comparison to iterative reconstruction and filtered back projection: a 12-reader phantom study

European Radiology Aims and scope Submit manuscript

Cite this article

Abstract

Objectives

(1) To compare low-contrast detectability of a deep learning–based denoising algorithm (DLA) with ADMIRE and FBP, and (2) to compare image quality parameters of DLA with those of reconstruction methods from two different CT vendors (ADMIRE, IMR, and FBP).

Materials and methods

Using abdominal CT images of 100 patients reconstructed via ADMIRE and FBP, we trained DLA by feeding FBP images as input and ADMIRE images as the ground truth. To measure the low-contrast detectability, the randomized repeat scans of Catphan® phantom were performed under various conditions of radiation exposures. Twelve radiologists evaluated the presence/absence of a target on a five-point confidence scale. The multi-reader multi-case area under the receiver operating characteristic curve (AUC) was calculated, and non-inferiority tests were performed. Using American College of Radiology CT accreditation phantom, contrast-to-noise ratio, target transfer function, noise magnitude, and detectability index (d’) of DLA, ADMIRE, IMR, and FBPs were computed.

Results

The AUC of DLA in low-contrast detectability was non-inferior to that of ADMIRE (p < .001) and superior to that of FBP (p < .001). DLA improved the image quality in terms of all physical measurements compared to FBPs from both CT vendors and showed profiles of physical measurements similar to those of ADMIRE.

Conclusions

The low-contrast detectability of the proposed deep learning–based denoising algorithm was non-inferior to that of ADMIRE and superior to that of FBP. The DLA could successfully improve image quality compared with FBP while showing the similar physical profiles of ADMIRE.

Key Points

• Low-contrast detectability in the images denoised using the deep learning algorithm was non-inferior to that in the images reconstructed using standard algorithms.

• The proposed deep learning algorithm showed similar profiles of physical measurements to advanced iterative reconstruction algorithm (ADMIRE).

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6

Abbreviations

ADMIRE:

Advanced modeled iterative reconstruction

AUC:

Area under the receiver operating characteristic curve

CNR:

Contrast-to-noise ratio

CT:

Computed tomography

DICOM:

Digital imaging and communications in medicine

DLA:

Deep learning–based denoising algorithm

FBP:

Filtered back projection

GUI:

Graphical user interface

HU:

Hounsfield unit

IMR :

Iterative model reconstruction

IR:

Iterative reconstruction

MBIR:

Model-based iterative reconstruction

NPS:

Noise power spectrum

TTF:

Target transfer function

References

  1. Papanicolas I, Woskie LR, Jha AK (2018) Health care spending in the United States and other high-income countries. JAMA 319:1024–1039

    Article  Google Scholar 

  2. Brenner DJ, Hall EJ (2007) Computed tomography–an increasing source of radiation exposure. N Engl J Med 357:2277–2284

    Article  CAS  Google Scholar 

  3. Smith-Bindman R, Lipson J, Marcus R et al (2009) Radiation dose associated with common computed tomography examinations and the associated lifetime attributable risk of cancer. Arch Intern Med 169:2078–2086

    Article  Google Scholar 

  4. Solomon J, Mileto A, Ramirez-Giraldo JC, Samei E (2015) Diagnostic performance of an advanced modeled iterative reconstruction algorithm for low-contrast detectability with a third-generation dual-source multidetector CT scanner: potential for radiation dose reduction in a multireader study. Radiology 275:735–745

    Article  Google Scholar 

  5. Volders D, Bols A, Haspeslagh M, Coenegrachts K (2013) Model-based iterative reconstruction and adaptive statistical iterative reconstruction techniques in abdominal CT: comparison of image quality in the detection of colorectal liver metastases. Radiology 269:469–474

    Article  Google Scholar 

  6. Chang W, Lee JM, Lee K et al (2013) Assessment of a model-based, iterative reconstruction algorithm (MBIR) regarding image quality and dose reduction in liver computed tomography. Invest Radiol 48:598–606

    Article  CAS  Google Scholar 

  7. Fontarensky M, Alfidja A, Perignon R et al (2015) Reduced radiation dose with model-based iterative reconstruction versus standard dose with adaptive statistical iterative reconstruction in abdominal CT for diagnosis of acute renal colic. Radiology 276:156–166

    Article  Google Scholar 

  8. Akagi M, Nakamura Y, Higaki T et al (2019) Deep learning reconstruction improves image quality of abdominal ultra-high-resolution CT. Eur Radiol 29:6163–6171

    Article  Google Scholar 

  9. Tatsugami F, Higaki T, Nakamura Y et al (2019) Deep learning-based image restoration algorithm for coronary CT angiography. Eur Radiol 29:5322–5329

  10. Nakamura Y, Higaki T, Tatsugami F et al (2019) Deep learning–based CT image reconstruction: initial evaluation targeting hypovascular hepatic metastases. Radiol Artif Intell 1:e180011

    Article  Google Scholar 

  11. Kang E, Chang W, Yoo J, Ye JC (2018) Deep convolutional framelet denosing for low-dose CT via wavelet residual network. IEEE Trans Med Imaging 37:1358–1369

    Article  Google Scholar 

  12. Shin YJ, Chang W, Ye JC et al (2020) Low-dose abdominal CT using a deep learning-based denoising algorithm: a comparison with CT reconstructed with filtered back projection or iterative reconstruction algorithm. Korean J Radiol 21:356–364

    Article  Google Scholar 

  13. Du W, Chen H, Wu Z, Sun H, Liao P, Zhang Y (2017) Stacked competitive networks for noise reduction in low-dose CT. PLoS One 12:e0190069

    Article  Google Scholar 

  14. Chen H, Zhang Y, Zhang W et al (2017) Low-dose CT via convolutional neural network. Biomed Opt Express 8:679–694

    Article  Google Scholar 

  15. Greffier J, Hamard A, Pereira F et al (2020) Image quality and dose reduction opportunity of deep learning image reconstruction algorithm for CT: a phantom study. Eur Radiol 30:3951–3959

    Article  Google Scholar 

  16. Rensink RA (2002) Change detection. Annu Rev Psychol 53:245–277

    Article  Google Scholar 

  17. Ro T, Russell C, Lavie N (2001) Changing faces: a detection advantage in the flicker paradigm. Psychol Sci 12:94–99

    Article  CAS  Google Scholar 

  18. Kim B, Lee H, Kim KJ et al (2011) Comparison of three image comparison methods for the visual assessment of the image fidelity of compressed computed tomography images. Med Phys 38:836–844

    Article  Google Scholar 

  19. Hillis SL, Obuchowski NA, Berbaum KS (2011) Power estimation for multireader ROC methods an updated and unified approach. Acad Radiol 18:129–142

    Article  Google Scholar 

  20. Chen W, Petrick NA, Sahiner B (2012) Hypothesis testing in noninferiority and equivalence MRMC ROC studies. Acad Radiol 19:1158–1165

    Article  Google Scholar 

Download references

Acknowledgements

The authors sincerely thank Jaeseok Bae, MD, PhD, and Sunkyung Jeon, MD, PhD, from Seoul National University Hospital; Jeong-sub Lee, MD, PhD, from Jeju National University Hospital; Jungheum Cho, MD, Jieun Park, MD, Yongju Kim, MD, Junbum Yeo, MD, and Sunyoung Park, MD, from Seoul National University Bundang Hospital for their participation in the human observer study. The authors thank Heejin Lee from Seoul National University Bundang Hospital for her assistance with physical measurement. The authors also thank Jong-June Jeon, PhD, from the Department of Statistics of University of Seoul and Sunkyu Choi, MS, from the Medical Research Collaborating Center (MRCC) of Seoul National University Bundang Hospital for their assistance with sample size calculations and statistical analyses.

Funding

This study was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) grant funded by the Korea government (*MSIT) (grant number NRF-2018R1C1B6007999), by the Seoul National University Bundang Hospital Research Fund (grant number 16-2018-004), by the National Research Foundation of Korea (NRF) grant funded by the Korea government (*MSIT) (grant number NRF-2016R1A2B3008104), and by a KAIST grant funded by Ministry of Science and ICT (grant number N11180149).

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Won Chang.

Ethics declarations

Guarantor

The scientific guarantor of this publication is Won Chang.

Conflict of interest

The authors of this manuscript declare no relationships with any companies whose products or services may be related to the subject matter of the article.

Statistics and biometry

Jong-June Jeon, PhD, and Sunkyu Choi, MS, kindly provided statistical advice for this manuscript.

Informed consent

Written informed consent was waived by the Institutional Review Board.

Ethical approval

Institutional Review Board approval was obtained

Methodology

• Not applicable

• Experimental

• Multicenter study

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

ESM 1

(DOCX 1811 kb)

Rights and permissions

Reprints and Permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Kim, Y., Oh, D.Y., Chang, W. et al. Deep learning–based denoising algorithm in comparison to iterative reconstruction and filtered back projection: a 12-reader phantom study. Eur Radiol 31, 8755–8764 (2021). https://doi.org/10.1007/s00330-021-07810-3

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00330-021-07810-3

Keywords

Navigation