Skip to main content
SpringerLink
Log in

An end-end deep learning framework for lesion segmentation on multi-contrast MR images—an exploratory study in a rat model of traumatic brain injury

  • Original Article
  • Published:
Medical & Biological Engineering & Computing Aims and scope Submit manuscript

Abstract

Traumatic brain injury (TBI) engenders traumatic necrosis and penumbra—areas of secondary neural injury which are crucial targets for therapeutic interventions. Segmenting manually areas of ongoing changes like necrosis, edema, hematoma, and inflammation is tedious, error-prone, and biased. Using the multi-parametric MR data from a rodent model study, we demonstrate the effectiveness of an end-end deep learning global-attention-based UNet (GA-UNet) framework for automatic segmentation and quantification of TBI lesions. Longitudinal MR scans (2 h, 1, 3, 7, 14, 30, and 60 days) were performed on eight Sprague–Dawley rats after controlled cortical injury was performed. TBI lesion and sub-regions segmentation was performed using 3D-UNet and GA-UNet. Dice statistics (DSI) and Hausdorff distance were calculated to assess the performance. MR scan variations-based (bias, noise, blur, ghosting) data augmentation was performed to develop a robust model.

Training/validation median DSI for U-Net was 0.9368 with T2w and MPRAGE inputs, whereas GA-UNet had 0.9537 for the same. Testing accuracies were higher for GA-UNet than U-Net with a DSI of 0.8232 for the T2w-MPRAGE inputs.

Longitudinally, necrosis remained constant while oligemia and penumbra decreased, and edema appearing around day 3 which increased with time. GA-UNet shows promise for multi-contrast MR image-based segmentation/quantification of TBI in large cohort studies.

Graphical Abstract

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
Fig. 7
Fig. 8
Fig. 9
Fig. 10

Similar content being viewed by others

Data availability

The authors confirm that all data underlying the findings are fully available without restriction. All the data which is used in the manuscript will be shared based on the request. For any request for data and DL models, please contact Bhanu Prakash KN—Principal Investigator—Clinical Data Analytics & Radiomics, Bioinformatics Institute, A*STAR.

References

  1. Centers for Disease Control and Prevention. (2015). Report to congress on traumatic brain injury in the United States: epidemiology and rehabilitation. National Center for Injury Prevention and Control; Division of Unintentional Injury Prevention. Atlanta, GA

  2. Marr, A.L., and Coronado, V.G. (Eds.). (2004). Central nervous system injury surveillance data submission standards—2002. Atlanta, GA: Centers for Disease Control and Prevention, National Center for Injury Prevention and Control.

  3. McKee AC, Daneshvar DH (2015) The neuropathology of traumatic brain injury. Handb Clin Neurol 127:45–66

    Article  PubMed  PubMed Central  Google Scholar 

  4. Unterberg AW, Stover J, Kress B, Kiening KL (2004) Edema and brain trauma. Neuroscience 129:1021–1029

    Article  CAS  PubMed  Google Scholar 

  5. Schweitzer AD, Niogi SN, Whitlow CJ, Tsiouris AJ (2019) Traumatic brain injury imaging patterns and complications. Radiograph: Rev Pub Radiol Soc North Am Inc 39(6):1571–1595

    Article  Google Scholar 

  6. Christine Turtzo L, Budde MD, Gold EM, Lewis BK, Janes L, Yarnell A, Grunberg NE, Watson W, Frank JA (2013) The evolution of traumatic brain injury in a rat focal contusion model. NMR Biomed. 26(4):468–479. https://doi.org/10.1002/nbm.2886

    Article  PubMed  Google Scholar 

  7. Stoffel M, Rinecker M, Graf R, Baethmann A, Plesnila N (2002) Nitric oxide in the penumbra of focal cortical necrosis in rats. Neurosci Lett 324:201–204

    Article  CAS  PubMed  Google Scholar 

  8. Ren H, Lu H (2019) Dynamic features of brain edema in rat models of traumatic brain injury. NeuroReport. 30(9):605–611. https://doi.org/10.1097/WNR.0000000000001213

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  9. Ledig C, Heckemann RA, Hammers A, López JC, Newcombe VF, Makropoulos A, Lötjönen J, Menon DK, Rueckert D (2015) Robust whole-brain segmentation: application to traumatic brain injury. Med Image Anal 21(1):40–58

    Article  PubMed  Google Scholar 

  10. Irimia A, Chambers MC, Alger JR, Filippou M, Prastawa M, Wang B, Hovda DA, Gerig G, Toga AW, Kikinis R, Vespa PM, Van Horn JD (2011) Comparison of acute and chronic traumatic brain injury using semi-automatic multimodal segmentation of MR volumes. J Neurotrauma 28(11):2287–2306

    Article  PubMed  PubMed Central  Google Scholar 

  11. Smith CM (2017). Chapter 8 - Neurotrauma. Handbook of clinical neurology, 145.

  12. Immonen R, Heikkinen T, Tähtivaara L, Nurmi A, Stenius T-K, Puoliväli J, Gröhn O (2010) Cerebral blood volume alterations in the perilesional areas in the rat brain after traumatic brain injury—comparison with behavioral outcome. J Cereb Blood Flow Metab 30(7):1318–1328. https://doi.org/10.1038/jcbfm.2010.15

    Article  PubMed  PubMed Central  Google Scholar 

  13. Soni N, Mohamed AZ, Kurniawan ND, Borges K, Nasrallah F (2019) Diffusion magnetic resonance imaging unveils the spatiotemporal microstructural gray matter changes following injury in the rodent brain. J Neurotrauma 36:1306–1317. https://doi.org/10.1089/neu.2018.5972

    Article  PubMed  Google Scholar 

  14. Long JA, Watts LT, Chemello J, Huang S, Shen Q, Duong TQ (2015) Multiparametric and longitudinal MRI characterization of mild traumatic brain injury in rats. J Neurotrauma 32:598–607. https://doi.org/10.1089/neu.2014.3563

    Article  PubMed  PubMed Central  Google Scholar 

  15. Guenette JP, Stern YorghosTripodis RA et al (2018) Automated versus manual segmentation of brain region volumes in former football players. NeruImage Clinical 18:888–896

    Article  Google Scholar 

  16. Akkus Z, Galimzianova A, Hoogi A, Rubin DL & Erickson BJ (2017). Deep learning for brain MRI segmentation: state of the art and future directions. Journal of Digital Imaging.

  17. Kamnitsas K, Ledig C, Newcombe V, Simpson JP, Kane AD, Menon DK, Rueckert D, Glocker B (2015) Segmentation of traumatic brain injuries with convolutional neural networks

    Google Scholar 

  18. Çiçek Ö, Abdulkadir A, Lienkamp SS, Brox T, Ronneberger O (2016) 3D U-Net: learning dense volumetric segmentation from sparse annotation. MICCAI

    Google Scholar 

  19. Kamnitsas K, Ledig C, Newcombe VF, Simpson JP, Kane AD, Menon DK, Rueckert D, Glocker B (2017) Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation. Med Image Anal 36:61–78

    Article  PubMed  Google Scholar 

  20. Kamnitsas K, Ferrante E, Parisot S, Ledig C, Nori AV, Criminisi A, Rueckert D & Glocker B (2016). DeepMedic for brain tumor segmentation. BrainLes@MICCAI.

  21. Kayalibay B, Jensen G & Smagt PV (2017). CNN-based segmentation of medical imaging data. ArXiv, abs/1701.03056..

  22. Wang G, Song T, Dong Q, Cui M, Huang N, Zhang S (2020) Automatic ischemic stroke lesion segmentation from computed tomography perfusion images by image synthesis and attention-based deep neural networks. Med Image Anal 65:101787

    Article  PubMed  Google Scholar 

  23. Tureckova Alzbeta and A. Rodríguez-Sánchez.(2018). ISLES Challenge: U-shaped convolution neural network with dilated convolution for 3D stroke lesion segmentation. BrainLes@MICCAI.

  24. Cornelio LK, Castillo MA, Naval P (2018) U-ISLES: ischemic stroke lesion segmentation using U-Net. IntelliSys

    Google Scholar 

  25. Cai Y, Wu S, Zhao W, Li Z, Wu Z & Ji S (2018). Concussion classification via deep learning using whole-brain white matter fiber strains. PLoS ONE, 13.

  26. Azad MM, Ganapathy A, Vadlamudi S, Paruchuri H (2021) Medical diagnosis using deep learning techniques: a research survey

    Google Scholar 

  27. Gravesteijn B, Nieboer D, Ercole A, Lingsma HF, Nelson D, Calster BV & Steyerberg E (2020). Machine learning algorithms performed no better than regression models for prognostication in traumatic brain injury. J Clin Epidemiol.

  28. Gan C, Sun D, Qin K, Zhao H & Xiao F.(2020). Improved traumatic brain injury classification approach based on deep learning. 2020 9th International Conference on Bioinformatics and Biomedical Science

  29. Raj R, Luostarinen T, Pursiainen E, Posti J, Takala R, Bendel S, Konttila T & Korja M (2019). Machine learning-based dynamic mortality prediction after traumatic brain injury. Scientific Reports, 9..

  30. Cenek M, Hu M, York G, Dahl S (2018) Survey of image processing techniques for brain pathology diagnosis: challenges and opportunities. Frontiers in Robotics and A I:5

    Google Scholar 

  31. Monteiro M, Newcombe V, Mathieu F, Adatia K, Kamnitsas K, Ferrante E, Das T, Whitehouse D, Rueckert D, Menon D, Glocker B (2020) Multiclass semantic segmentation and quantification of traumatic brain injury lesions on head CT using deep learning: an algorithm development and multicentre validation study. The Lancet Digital health 2(6):e314–e322

    Article  PubMed  Google Scholar 

  32. Phaphuangwittayakul A, Guo Y, Ying F, Dawod AY, Angkurawaranon S, Angkurawaranon C (2022) An optimal deep learning framework for multi-type hemorrhagic lesions detection and quantification in head CT images for traumatic brain injury. Applied Intelligence (Dordrecht, Netherlands) 52:7320–7338

    PubMed  Google Scholar 

  33. Adil SM, Elahi C, Patel DN, Seas A, Warman PI, Fuller AT, Haglund MM, Dunn TW (2022Aug) Deep learning to predict traumatic brain injury outcomes in the low-resource setting. World Neurosurg 164:e8–e16. https://doi.org/10.1016/j.wneu.2022.02.097

    Article  PubMed  Google Scholar 

  34. Farzaneh N, Williamson CA, Gryak J & Najarian K (2021). A hierarchical expert-guided machine learning framework for clinical decision support systems: an application to traumatic brain injury prognostication. NPJ Digital Medicine, 4.

  35. Ronneberger O, Fischer P & Brox T (2015). U-Net: convolutional networks for biomedical image segmentation. ArXiv, abs/1505.04597.

  36. Oktay O, Schlemper J, Folgoc LL, Lee MJ, Heinrich M, Misawa K, Mori K, McDonagh SG, Hammerla N, Kainz B, Glocker B & Rueckert D. (2018). Attention U-Net: learning where to look for the pancreas. ArXiv, abs/1804.03999.

  37. Roy S, Knutsen A, Korotco A, Bosomtwi A, Dardzinski B, Butman JA, DzungPham L (2018) A deep learning framework for brain extraction in humans and animals with traumatic brain injury 2018, IEEE 15th International Symposium on Biomedical Imaging (ISBI 2018). D.C., USA, Washington

    Google Scholar 

  38. Feo R, Giove F (2019) Towards an efficient segmentation of small rodents brain a short critical review. J Neurosci Methods 323:82–89

    Article  PubMed  Google Scholar 

  39. Mohamed AZ, Cumming P & Nasrallah FA (2021). Traumatic brain injury augurs ill for prolonged deficits in the brain’s structural and functional integrity following controlled cortical impact injury. Sci Rep 11.

  40. Smith SM (2002) Fast robust automated brain extraction. Hum Brain Mapp 17:143–155. https://doi.org/10.1002/hbm.10062

    Article  PubMed  PubMed Central  Google Scholar 

  41. Zheng W, Chee M, Zagorodnov V (2009) Improvement of brain segmentation accuracy by optimizing non-uniformity correction using N3. Neuroimage 48:73–83

    Article  PubMed  Google Scholar 

  42. Jenkinson M, Bannister P, Brady M, Smith S (2002) Improved optimization for the robust and accurate linear registration and motion correction of brain images. Neuroimage 17:825–841. https://doi.org/10.1016/S1053-8119(02)91132-8

    Article  PubMed  Google Scholar 

  43. Pérez-García F, Sparks R & Ourselin S (2020). TorchIO: a Python library for efficient loading, preprocessing, augmentation and patch-based sampling of medical images in deep learning. ArXiv, abs/2003.04696.

  44. Muhammed Kürşad Uçar, Majid Nour, Hatem Sindi,Kemal Polat(2020). The effect of training and testing process on machine learning in biomedical datasets. Mathematical Problems in Engineering. 2020 2836236

  45. Zenoda website - https://zenodo.org

  46. Ertürk A, Mentz S, Stout EE, Hedehus M, Dominguez SL, Neumaier L, Krammer F, Llovera G, Srinivasan K, Hansen DV, Liesz A, Scearce-Levie K, Sheng M (2016) Interfering with the chronic immune response rescues chronic degeneration after traumatic brain injury. J Neurosci 36:9962–9975

    Article  PubMed  PubMed Central  Google Scholar 

  47. Turtzo LC, Budde MD, Gold EM, Lewis BK, Janes L, Yarnell AM, Grunberg NE, Watson W & Frank JA (2013). The evolution of traumatic brain injury in a rat focal contusion model. NMR in Biomedicine, 26.

  48. Sergey L, Szegedy C (2015, March 2). Batch normalization: accelerating deep network training by reducing internal covariate shift, from https://arxiv.org/abs/1502.03167.

  49. Nair V, Hinton G (2010). Rectified linear units improve restricted boltzmann machines, from https://www.cs.toronto.edu/~fritz/absps/reluICML.pdf

  50. Chen L, Zhu Y, Papandreou G, Schroff F & Adam H (2018). Encoder-decoder with atrous separable convolution for semantic image segmentation. ECCV.

  51. Islam M, Vibashan V, Jose VJ, Wijethilake N, Utkarsh U & Ren H (2019). Brain tumor segmentation and survival prediction using 3D attention UNet. BrainLes@MICCAI

  52. Huang H, Lin L, Tong R, Hu H, Zhang Q, Iwamoto Y, Han X, Chen Y & Wu J (2020). UNet 3+: a full-scale connected UNet for medical image segmentation. ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 1055–1059.

  53. Loshchilov, Ilya, and Frank Hutter. (2017). Decoupled weight decay regularization.ICLR.

  54. Srivastava N, Hinton GE, Krizhevsky A, Sutskever I, Salakhutdinov R (2014) Dropout: a simple way to prevent neural networks from overfitting. J Mach Learn Res 15:1929–1958

    Google Scholar 

  55. Huttenlocher DP, Klanderman GA, Rucklidge WJ (Sept. 1993) Comparing images using the Hausdorff distance. IEEE Trans Pattern Anal Mach Intell 15(9):850–863. https://doi.org/10.1109/34.232073

    Article  Google Scholar 

  56. React JavaScript webpage https://reactjs.org/

  57. Nickerson C (1997) A note on A concordance correlation coefficient to evaluate reproducibility. Biometrics 53(4):1503–1507

    Article  Google Scholar 

  58. Onyszchuk G, Al-Hafez B, He YY, Bilgen M, Berman NE, Brooks WM (2007) A mouse model of sensorimotor controlled cortical impact: characterization using longitudinal magnetic resonance imaging, behavioral assessments, and histology. J Neurosci Methods. 160(2):187–96

    Article  PubMed  Google Scholar 

  59. Onyszchuk G, He YY, Berman NE, Brooks WM (2008) Detrimental effects of aging on outcome from traumatic brain injury: a behavioral, magnetic resonance imaging, and histological study in mice. J Neurotrauma 25(2):153–71

    Article  PubMed  Google Scholar 

  60. Hernandez-Ontiveros DG, Tajiri N, Acosta SA, Giunta BN, Tan J & Borlongan CV (2013). Microglia activation as a biomarker for traumatic brain injury. Frontiers in Neurology, 4.

  61. Roy Snehashis, Butman John, Chan Leighton, Pham Dzung. (2018). TBI contusion segmentation from MRI using convolutional neural networks. 158–162. https://doi.org/10.1109/ISBI.2018.8363545.

Download references

Acknowledgements

The authors thank Agency for Science, Technology, and Research (A*STAR), the Singapore Bioimaging Consortium (SBIC), the Institute of Bioengineering and Bioimaging, the Bioinformatics Institute, and the Queensland Brain Institute (QBI), Australia for supporting this study.

Funding

This research was sponsored by Motor Accident Insurance Commission (MAIC), The Queensland Government, Australia (Grant: 2014000857) at Queensland Brain Institute.

Author information

Author notes

  1. Authors Arvind CS  and Abdalla Mohammed have equally contributed to the work.

Authors and Affiliations

  1. Clinical Data Analytics & Radiomics, Cellular Image Informatics, Bioinformatics Institute, A*STAR, 30 Biopolis St Matrix, Singapore, 138671, Singapore

    Bhanu Prakash KN & Arvind CS

  2. Cellular Image Informatics, Bioinformatics Institute, A*STAR Horizontal Technology Centers, Singapore, Singapore

    Bhanu Prakash KN

  3. Queensland Brain Institute, The University of Queensland, Building 79, Upland Road, Saint Lucia, Brisbane, QLD, 4072, Australia

    Abdalla Mohammed, Xuan Vinh To, Hussein Srour & Fatima Nasrallah

  4. Signal and Image Processing Group, Laboratory of Molecular Imaging, Singapore Bioimaging Consortium, A*STAR, 02-02 Helios 11, Biopolis Way, Singapore, 138667, Singapore

    Krishna Kanth Chitta

Authors
  1. Bhanu Prakash KN
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Arvind CS
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Abdalla Mohammed
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Krishna Kanth Chitta
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Xuan Vinh To
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Hussein Srour
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Fatima Nasrallah
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Bhanu Prakash KN.

Ethics declarations

Ethics approval

This study was approved by the Animal Research Ethics Committee (AEC) of the University of Queensland (IRB number: QBI/036/16/MAIC).

Conflict of interest

The authors declare no competing interests.

Additional information

Publisher's note

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

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (DOCX 37 KB)

Supplementary file2 (MP4 11632 KB)

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

KN, B.P., CS, A., Mohammed, A. et al. An end-end deep learning framework for lesion segmentation on multi-contrast MR images—an exploratory study in a rat model of traumatic brain injury. Med Biol Eng Comput 61, 847–865 (2023). https://doi.org/10.1007/s11517-022-02752-4

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11517-022-02752-4

Keywords

Search

Navigation