Skip to main content
SpringerLink
Log in

A novel methodology to train and deploy a machine learning model for personalized dose assessment in head CT

  • Computed Tomography
  • Published:
European Radiology Aims and scope Submit manuscript

Abstract

Objectives

To propose a machine learning–based methodology for the creation of radiation dose maps and the prediction of patient-specific organ/tissue doses associated with head CT examinations.

Methods

CT data were collected retrospectively for 343 patients who underwent standard head CT examinations. Patient-specific Monte Carlo (MC) simulations were performed to determine the radiation dose distribution to patients’ organs/tissues. The collected CT images and the MC–produced dose maps were processed and used for the training of the deep neural network (DNN) model. For the training and validation processes, data from 231 and 112 head CT examinations, respectively, were used. Furthermore, a software tool was developed to produce dose maps from head CT images using the trained DNN model and to automatically calculate the dose to the brain and cranial bones.

Results

The mean (range) percentage differences between the doses predicted from the DNN model and those provided by MC simulations for the brain, eye lenses, and cranial bones were 4.5% (0–17.7%), 5.7% (0.2–19.0%), and 5.2% (0.1–18.9%), respectively. The graphical user interface of the software offers a user-friendly way for radiation dose/risk assessment. The implementation of the DNN allowed for a 97% reduction in the computational time needed for the dose estimations.

Conclusions

A novel methodology that allows users to develop a DNN model for patient-specific CT dose prediction was developed and implemented. The approach demonstrated herein allows accurate and fast radiation dose estimation for the brain, eye lenses, and cranial bones of patients who undergo head CT examinations and can be used in everyday clinical practice.

Key Points

• The methodology presented herein allows fast and accurate radiation dose estimation for the brain, eye lenses, and cranial bones of patients who undergo head CT examinations and can be implemented in everyday clinical practice.

• The scripts developed in the current study will allow users to train models for the acquisition protocols of their CT scanners, generate dose maps, estimate the doses to the brain and cranial bones, and estimate the lifetime attributable risk of radiation-induced brain cancer.

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

Similar content being viewed by others

Abbreviations

CT:

Computed tomography

CTDI:

CT dose index

DNN:

Deep neural network

GUI:

Graphical user interface

HU:

Hounsfield unit

ICRP:

International Commission on Radiological Protection

LAR:

Lifetime attributable risk

MC:

Monte Carlo

ROI:

Region of interest

References

  1. Gaudreau K, Thome C, Weaver B, Boreham DR (2020) Cataract formation and low-dose radiation exposure from head computed tomography (CT) scans in Ontario, Canada, 1994-2015. Radiat Res 193:322–330

    Article  CAS  Google Scholar 

  2. Brinjikji W, Kallmes DF, Cloft HJ (2015) Rising utilization of CT in adult fall patients. AJR Am J Roentgenol 204:558–562

  3. Kim HG, Lee HJ, Lee SK, Kim HJ, Kim MJ (2017) Head CT: image quality improvement with ASIR-V using a reduced radiation dose protocol for children. Eur Radiol 27:3609–3617

    Article  Google Scholar 

  4. Investigative Report ICES (2007) Enhancing the effectiveness of health care for Ontarians through research Diagnostic Services in Ontario: Descriptive Analysis and Jurisdictional Review. ICES, Toronto, ON

    Google Scholar 

  5. Götz TI, Schmidkonz C, Chen S, Al-Baddai S, Kuwert T, Lang EW (2020) A deep learning approach to radiation dose estimation. Phys Med Biol 65:035007

    Article  Google Scholar 

  6. Lee MS, Hwang D, Kim JH, Lee JS (2019) Deep-dose: a voxel dose estimation method using deep convolutional neural network for personalized internal dosimetry. Sci Rep 9(1):10308

    Article  Google Scholar 

  7. Maier J, Eulig E, Dorn S, Sawall S, Kachelrieß M (2018) Real-time patient-specific CT dose estimation using a deep convolutional neural network. IEEE Nuclear Science Symposium and Medical Imaging Conference Proceedings (NSS/MIC), pp 1–3

  8. Deak P, van Straten M, Shrimpton PC, Zankl M, Kalender WA (2008) Validation of a Monte Carlo tool for patient-specific dose simulations in multi-slice computed tomography. Eur Radiol 18:759–772

    Article  Google Scholar 

  9. Myronakis M, Perisinakis K, Tzedakis A, Gourtsoyianni S, Damilakis J (2009) Evaluation of a patient-specific Monte Carlo software for CT dosimetry. Radiat Prot Dosimetry 133:248–255

    Article  CAS  Google Scholar 

  10. Damilakis J, Perisinakis K, Tzedakis A, Papadakis AE, Karantanas A (2010) Radiation dose to the conceptus from multidetector CT during early gestation: a method that allows for variations in maternal body size and conceptus position. Radiology 257:483–489

    Article  Google Scholar 

  11. Lloyd SP (1982) Least squares quantization in PCM. IEEE Trans Inf Theory 28:129–137

    Article  Google Scholar 

  12. Erisoglu M, Calis N, Sakallioglu S (2011) A new algorithm for initial cluster centers in k-means algorithm. Pattern Recognit Lett 32:1701–1705

    Article  Google Scholar 

  13. Pedregosa F, Varoquaux G, Gramfort A et al (2011) Scikit-learn: machine learning in Python. J Mach Learn Res 12:2825–2830

    Google Scholar 

  14. Harris CR, Millman KJ, van der Walt SJ et al (2020) Array programming with NumPy. Nature 585:357–362

    Article  CAS  Google Scholar 

  15. Kingma DP, Ba J (2014) Adam: a method for stochastic optimization. Available at: http://arxiv.org/abs/1412.6980.

  16. Srivastava N, Hinton G, 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 

  17. Li L, Jamieson K, DeSalvo G, Rostamizadeh A, Talwalkar A (2018) Hyperband: a novel bandit-based approach to hyperparameter optimization. J Mach Learn Res 18:1–52

    Google Scholar 

  18. Agarap AF (2018) Deep learning using rectified linear units (relu). ArXiv Preprint ArXiv:1803.08375.

  19. National Research Council (2006) Health risks from exposure to low levels of ionizing radiation: BEIR VII Phase 2. The National Academies Press, Washington, DC

    Google Scholar 

  20. Van Rossum G (2020) The Python library reference, release 3.8.2. Python Software Foundation 2020

  21. Schindelin J, Arganda-Carreras I, Frise E et al (2012) Fiji: an open-source platform for biological-image analysis. Nat Methods 9:676–682

    Article  CAS  Google Scholar 

  22. Damilakis J (2021) CT dosimetry: what has been achieved and what remains to be done. Invest Radiol 56:62–68

    Article  Google Scholar 

  23. ICRP (2007) The 2007 Recommendations of the International Commission on Radiological Protection. CRP publication 103. Ann ICRP 37:9–34

    Google Scholar 

  24. Harrison JD, Balonov M, Bochud F et al (2021) ICRP Publication 147: use of dose quantities in radiological protection. Ann ICRP 50:9–82

    Article  CAS  Google Scholar 

  25. Ria F, Bergantin A, Vai A et al (2017) Awareness of medical radiation exposure among patients: a patient survey as a first step for effective communication of ionizing radiation risks. Phys Med 43:57–62

    Article  CAS  Google Scholar 

  26. Sin HK, Wong CS, Huang B, Yiu KL, Wong WL, Chu YCT (2013) Assessing local patients’ knowledge and awareness of radiation dose and risks associated with medical imaging: a questionnaire study. J Med Imaging Radiat Oncol 57:38–44

    Article  Google Scholar 

  27. Schuster AL, Forman HP, Strassle PD, Meyer LT, Connelly SV, Lee CI (2018) Awareness of radiation risks from CT scans among patients and providers and obstacles for informed decision-making. Emerg Radiol 25:41–49

    Article  Google Scholar 

  28. Lee CI, Haims AH, Monico EP, Brink JA, Forman HP (2004) Diagnostic CT scans: assessment of patient, physician, and radiologist awareness of radiation dose and possible risks. Radiology 231:393–398

    Article  Google Scholar 

  29. Peng Z, Fang X, Yan P et al (2020) A method of rapid quantification of patient-specific organ doses for CT using deep-learning-based multi-organ segmentation and GPU-accelerated Monte Carlo dose computing. Med Phys 47(6):2526–2536

    Article  CAS  Google Scholar 

  30. Kearney V, Chan JW, Haaf S, Descovich M, Solberg TD (2018) DoseNet: a volumetric dose prediction algorithm using 3D fully-convolutional neural networks. Phys Med Biol 63(23):235022

    Article  Google Scholar 

  31. European Commission (2014) Council Directive 2013/59/Euratom of 5 December 2013. Off J Eur Union. https://doi.org/10.3000/19770677.L_2013.124.eng

Download references

Funding

This study has received funding from the research project “Patient dosimetry in computed tomography, interventional radiology and mammography” (KA10697).

Author information

Authors and Affiliations

  1. Department of Medical Physics, School of Medicine, University of Crete, P.O. Box 2208, 71003, Heraklion, Crete, Greece

    Eleftherios Tzanis & John Damilakis

Authors
  1. Eleftherios Tzanis
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. John Damilakis
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to John Damilakis.

Ethics declarations

Guarantor

The scientific guarantor of this publication is Prof. John Damilakis.

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

No complex statistical methods were necessary for this paper.

Informed consent

Written informed consent was not required for this study.

Ethical approval

Institutional review board approval was obtained.

Methodology

• Retrospective

• Experimental

• Performed at one institution

Additional information

Publisher’s note

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

Supplementary Information

ESM 1

(RAR 927 kb)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Tzanis, E., Damilakis, J. A novel methodology to train and deploy a machine learning model for personalized dose assessment in head CT. Eur Radiol 32, 6418–6426 (2022). https://doi.org/10.1007/s00330-022-08756-w

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00330-022-08756-w

Keywords

Search

Navigation