Advertisement

European Radiology

, Volume 29, Issue 5, pp 2724–2726 | Cite as

Use of artificial neural networks to predict anterior communicating artery aneurysm rupture: possible methodological considerations

  • Guido Adriaan de Jong
  • René AquariusEmail author
Letter to the Editor
  • 61 Downloads

Key Points

  • Use of algorithms to generate synthetic cases might result in a misrepresentation of the entire population.

  • Training an artificial neural network with a mix of real and synthetic data might lead to non-realistic prediction precision.

Notes

Acknowledgements

We would like to acknowledge professor Ronald Bartels for his constructive contributions to the presented work.

Funding

The authors state that this work has not received any funding.

Compliance with ethical standards

Guarantor

The scientific guarantor of this publication is Ronald Bartels (head of the neurosurgery department).

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 because data collection from human subjects was not needed.

Ethical approval

Institutional Review Board approval was not required because synthetic data was generated for this study.

Methodology

• Performed at one institution

References

  1. 1.
    Liu J, Chen Y, Lan L et al (2018) Prediction of rupture risk in anterior communicating artery aneurysms with a feed-forward artificial neural network. Eur Radiol.  https://doi.org/10.1007/s00330-017-5300-3
  2. 2.
    He H, Bai Y, Garcia EA, Li S (2008) ADASYN: adaptive synthetic sampling approach for imbalanced learning. IEEE International Joint Conference on Neural Networks, pp 1322–1328Google Scholar
  3. 3.
    Tang B, He H (2015) KernelADASYN: kernel based adaptive synthetic data generation for imbalanced learning. IEEE Congress on Evolutionary Computation, pp 664–671Google Scholar
  4. 4.
    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 15:1929–1958Google Scholar
  5. 5.
    Boughorbel S, Jarray F, El-Anbari M (2017) Optimal classifier for imbalanced data using Matthews correlation coefficient metric. PLoS One 12:1–17CrossRefGoogle Scholar

Copyright information

© European Society of Radiology 2018

Authors and Affiliations

  1. 1.Radboud University Medical Center, Department of NeurosurgeryRadboud University Medical CenterNijmegenThe Netherlands

Personalised recommendations