Skip to main content
SpringerLink
Log in

Precise whole liver automatic segmentation and quantification of PDFF and R2* on MR images

  • Imaging Informatics and Artificial Intelligence
  • Published:
European Radiology Aims and scope Submit manuscript

Abstract

Objective

To automate the segmentation of whole liver parenchyma on multi-echo chemical shift encoded (MECSE) MR examinations using convolutional neural networks (CNNs) to seamlessly quantify precise organ-related imaging biomarkers such as the fat fraction and iron load.

Methods

A retrospective multicenter collection of 183 MECSE liver MR examinations was conducted. An encoder-decoder CNN was trained (107 studies) following a 5-fold cross-validation strategy to improve the model performance and ensure lack of overfitting. Proton density fat fraction (PDFF) and R2* were quantified on both manual and CNN segmentation masks. Different metrics were used to evaluate the CNN performance over both unseen internal (46 studies) and external (29 studies) validation datasets to analyze reproducibility.

Results

The internal test showed excellent results for the automatic segmentation with a dice coefficient (DC) of 0.93 ± 0.03 and high correlation between the quantification done with the predicted mask and the manual segmentation (rPDFF = 1 and rR2* = 1; p values < 0.001). The external validation was also excellent with a different vendor but the same magnetic field strength, proving the generalization of the model to other manufacturers with DC of 0.94 ± 0.02. Results were lower for the 1.5-T MR same vendor scanner with DC of 0.87 ± 0.06. Both external validations showed high correlation in the quantification (rPDFF = 1 and rR2* = 1; p values < 0.001). In both internal and external validation datasets, the relative error for the PDFF and R2* quantification was below 4% and 1% respectively.

Conclusion

Liver parenchyma can be accurately segmented with CNN in a vendor-neutral virtual approach, allowing to obtain reproducible automatic whole organ virtual biopsies.

Key points

• Whole liver parenchyma can be automatically segmented using convolutional neural networks.

• Deep learning allows the creation of automatic pipelines for the precise quantification of liver-related imaging biomarkers such as PDFF and R2*.

• MR “virtual biopsy” can become a fast and automatic procedure for the assessment of chronic diffuse liver diseases in clinical practice.

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

Access this article

Log in via an institution

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

Similar content being viewed by others

Abbreviations

AI:

Artificial intelligence

ASSD:

Average symmetric surface distance

CNN:

Convolutional neural network

DC:

Dice coefficient

FDR:

False discovery rate

MECSE:

Multi-echo chemical shift encoded

MSD:

Maximum surface distance

NAFLD:

Non-alcoholic fatty liver disease

NASH:

Non-alcoholic steatohepatitis

PDFF:

Proton density fat fraction

RVD:

Relative volume difference

VOE:

Volumetric overlap error

References

  1. Donato H, França M, Candelária I, Caseiro-Alves F (2017) Liver MRI: from basic protocol to advanced techniques. Eur J Radiol 93:30–39

    Article  Google Scholar 

  2. Farrell GC, Larter CZ (2006) Nonalcoholic fatty liver disease: from steatosis to cirrhosis. Hepatology. 43(2 Suppl 2):S99–S112

    Article  CAS  Google Scholar 

  3. Kinner S, Reeder SB, Yokoo T (2016) Quantitative imaging biomarkers of NAFLD. Dig Dis Sci 61(5):1337–1347

    Article  CAS  Google Scholar 

  4. França M, Alberich-Bayarri Á, Martí-Bonmatí L et al (2017) Accurate simultaneous quantification of liver steatosis and iron overload in diffuse liver diseases with MRI. Abdom Radiol (NY) 42(5):1434–1443

    Article  Google Scholar 

  5. Kühn JP, Hernando D, Muñoz del Rio A et al (2012) Effect of multipeak spectral modeling of fat for liver iron and fat quantification: correlation of biopsy with MR imaging results. Radiology. 265(1):133–142

    Article  Google Scholar 

  6. Rostoker G, Laroudie M, Blanc R et al (2017) Signal-intensity-ratio MRI accurately estimates hepatic iron load in hemodialysis patients. Heliyon. 3(1):e00226

    Article  Google Scholar 

  7. Ratziu V, Charlotte F, Heurtier A et al (2005) Sampling variability of liver biopsy in nonalcoholic fatty liver disease. Gastroenterology. 128(7):1898–1906

    Article  Google Scholar 

  8. Butensky E, Fischer R, Hudes M et al (2005) Variability in hepatic iron concentration in percutaneous needle biopsy specimens from patients with transfusional hemosiderosis. Am J Clin Pathol 123(1):146–152

    Article  Google Scholar 

  9. Deugnier Y, Turlin B (2007) Pathology of hepatic iron overload. World J Gastroenterol 13(35):4755–4760

    Article  CAS  Google Scholar 

  10. Martí-Bonmatí L, Alberich-Bayarri A, Sánchez-González J (2012) Overload hepatitides: quanti-qualitative analysis. Abdom Imaging 37(2):180–187

    Article  Google Scholar 

  11. McCarville MB, Hillenbrand CM, Loeffler RB et al (2010) Comparison of whole liver and small region-of-interest measurements of MRI liver R2* in children with iron overload. Pediatr Radiol 40(8):1360–1367

    Article  Google Scholar 

  12. Campo CA, Hernando D, Schubert T, Bookwalter CA, Pay AJV, Reeder SB (2017) Standardized approach for ROI-based measurements of proton density fat fraction and R2* in the liver. AJR Am J Roentgenol 209(3):592–603

    Article  Google Scholar 

  13. Esfandiarkhani M, Foruzan AH (2017) A generalized active shape model for segmentation of liver in low-contrast CT volumes. Comput Biol Med 82:59–70

    Article  Google Scholar 

  14. Yan Z, Zhang S, Tan C et al (2015) Atlas-based liver segmentation and hepatic fat-fraction assessment for clinical trials. Comput Med Imaging Graph 41:80–92

    Article  Google Scholar 

  15. Dura E, Domingo J, Göçeri E, Martí-Bonmatí L (2017) A method for liver segmentation in perfusion MR images using probabilistic atlases and viscous reconstruction. Pattern Anal Appl 21(4):1083–1095

    Article  Google Scholar 

  16. Xu Y, Xu C, Kuang X et al (2016) 3D-SIFT-Flow for atlas-based CT liver image segmentation. Med Phys 43(5):2229

    Article  Google Scholar 

  17. Yuan Z, Wang Y, Yang J, Liu Y (2010). A novel automatic liver segmentation technique for MR images. 2010 3rd International Congress on Image and Signal Processing, Yantai. 1282–1286

  18. Göçeri E (2016) Fully automated liver segmentation using Sobolev gradient-based level set evolution. Int J Numer Method Biomed Eng 32(11)

  19. Yang D, Xu D, Zhou SK et al (2017). Automatic liver segmentation using an adversarial image-to-image network. Medical Image Computing and Computer-Assisted. MICCAI 2017. Lecture Notes in Computer Science, vol 10435. Springer, Cham

  20. Han X (2017). Automatic liver lesion segmentation using a deep convolutional neural network method. arXiv: 1704.07239

  21. Qin W, Wu J, Han F et al (2018) Superpixel-based and boundary-sensitive convolutional neural network for automated liver segmentation. Phys Med Biol 63(9):095017

    Article  Google Scholar 

  22. Christ PF, Ettlinger F, Grün F et al (2017). Automatic liver and tumor segmentation of CT and MRI volumes using cascaded fully convolutional neural networks. arXiv: 1702.05970

  23. Wang K, Mamidipalli A, Retson T et al (2019) Automated CT and MRI liver segmentation and biometry using a generalized convolutional neural network. Radiol Artif Intell 1(2):180022

    Article  Google Scholar 

  24. Lavdas I, Glocker B, Kamnitsas K et al (2017) Fully automatic, multiorgan segmentation in normal whole body magnetic resonance imaging (MRI), using classification forests (CFs), convolutional neural networks (CNNs), and a multi-atlas (MA) approach. Med Phys 44(10):5210–5220

    Article  Google Scholar 

  25. Jansen MJA, Kuijf HJ, Niekel M et al (2019) Liver segmentation and metastases detection in MR images using convolutional neural networks. J Med Imaging (Bellingham) 6(4):044003

    Google Scholar 

  26. Montagnon E, Cerny M, Cadrin-Chênevert A et al (2020) Deep learning workflow in radiology: a primer. Insights Imaging 11(1)

  27. Ronneberger O, Fischer P, Brox T (2015) U-Net: Convolutional networks for biomedical image segmentation. Medical Image Computing and Computer-Assisted. MICCAI 2015. Lecture Notes in Computer Science, vol 9351. Springer, Cham

  28. Wang L, Yang Y, Min R, Chakradhar S (2017) Accelerating deep neural network training with inconsistent stochastic gradient descent. Neural Netw 93:219–229

    Article  Google Scholar 

  29. Lee C, Xie S, Gallagher P, Zhang Z, Tu Z (2015). Deeply-supervised nets. Artif Intell Stat 562–570

  30. Dou Q, Chen H, Jin Y, Yu L, Qin J, Heng P (2016) 3D deeply supervised network for automatic liver segmentation from CT volumes. Medical Image Computing and Computer-Assisted. MICCAI 2016. Springer, Cham

  31. Sudre CH, Li W, Vercauteren T, Ourselin S, Cardoso MJ (2017). Generalised Dice overlap as a deep learning loss function for highly unbalanced segmentations. Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support. DLMIA 2017, ML-CDS 2017. Lecture Notes in Computer Science, vol 10553. Springer, Cham

  32. Kingma DP, Ba J (2015) Adam: a method for stochastic optimization. international conference on learning representations. 1–13. arXiv: 1412.6980

  33. Stocker D, Bashir MR, Kannengiesser SAR, Reiner CS (2018) Accuracy of automated liver contouring, fat fraction, and R2* measurement on gradient multiecho magnetic resonance images. J Comput Assist Tomogr 42(5):697–706

    Article  Google Scholar 

  34. Fezza SA, Bakhti Y, Hamidouche W, Deforges O (2019) Perceptual evaluation of adversarial attacks for CNN-based image classification. 2019 Eleventh International Conference on Quality of Multimedia Experience (QoMEX). 1–6

  35. Chen L, Bentley P, Mori K, Misawa K, Fujiwara M, Rueckert D (2019) Intelligent image synthesis to attack a segmentation CNN using adversarial learning. Simulation and Synthesis in Medical Imaging. SASHIMI 2019. Lecture Notes in Computer Science, vol 11827. Springer, Cham

  36. Yan W, Huang L, Xia L et al (2020) MRI manufacturer shift and adaptation: increasing the generalizability of deep learning segmentation for MR images acquired with different scanners. Radiol Artif Intell 2(4):e190195

    Article  Google Scholar 

Download references

Acknowledgements

DMA is recipient of a Río Hortega award (CM19/00212), Instituto de Salud Carlos III.

Funding

This study was partially funded by the Spanish Ministry of Science and innovation, Instituto de Salud Carlos III (PI19/0380) and GILEAD Sciences (Grant Number: GLD19/00050). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Author information

Authors and Affiliations

  1. Quantitative Imaging Biomarkers in Medicine, QUIBIM S.L, Aragon Avenue, 30, 13th floor, Office J, 46021, Valencia, Spain

    Ana Jimenez-Pastor, Angel Alberich-Bayarri & Rafael Lopez-Gonzalez

  2. Digestive Disease Department, Clinic University Hospital, INCLIVA Health Research Institute, Valencia, Spain

    David Marti-Aguado

  3. Radiology Department, Centro Hospitalar Universitário do Porto (CHUP), Porto, Portugal

    Manuela França

  4. Laboratorio Clinico Diagnomed, Concepción, Chile

    Rodrigo San Martin Bachmann

  5. ARGUS Diagnostico Medico, Buenos Aires, Argentina

    Juan Mazzucco

  6. Biomedical Imaging Research Group (GIBI230-PREBI) at La Fe Health Research Institute, and Imaging La Fe node at Distributed Network for Biomedical Imaging (ReDIB) Unique Scientific and Technical Infrastructures (ICTS), Valencia, Spain

    Luis Marti-Bonmati

  7. Radiology Department, La Fe University and Polytechnic Hospital, Valencia, Spain

    Luis Marti-Bonmati

Authors
  1. Ana Jimenez-Pastor
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Angel Alberich-Bayarri
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Rafael Lopez-Gonzalez
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. David Marti-Aguado
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Manuela França
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Rodrigo San Martin Bachmann
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Juan Mazzucco
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Luis Marti-Bonmati
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Ana Jimenez-Pastor.

Ethics declarations

Guarantor

The scientific guarantor of this publication is Luis Marti-Bonmati.

Conflict of interest

The authors of this manuscript declare relationships with the following companies: QUIBIM SL.

Statistics and biometry

No complex statistical methods were necessary for this paper.

Informed consent

Written informed consent was waived by the Institutional Review Board.

Ethical approval

Institutional Review Board approval was obtained.

Methodology

• Retrospective

• Observational

• Multicenter study

Additional information

Publisher’s note

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

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Jimenez-Pastor, A., Alberich-Bayarri, A., Lopez-Gonzalez, R. et al. Precise whole liver automatic segmentation and quantification of PDFF and R2* on MR images. Eur Radiol 31, 7876–7887 (2021). https://doi.org/10.1007/s00330-021-07838-5

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00330-021-07838-5

Keywords

Search

Navigation