Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Viewpoint
  • Published:

Artificial neural networks: has the time come for their use in prostate cancer patients?

Nature Clinical Practice Urology volume 2, pages 262–263 (2005)Cite this article

Artificial neural networks (ANNs) are multivariate computational models based on human synaptic physiology that have the ability to learn from experience. Applications have been developed to support clinical decision-making in various fields. An ANN model must first be 'trained' by using cases with known outcomes. It will then adjust its weighting of various input variables over time to optimize the output data. A key advantage of ANNs compared with conventional methods like logistic regression is their ability to resolve complex nonlinear relations among variables, without the need for any prior assumptions about these relations. However, contradictory results have been reported for the performance of ANNs, indicating that it is not always possible to confirm a genuine advantage to their use compared with logistic regression.1,2 In a comparison of 28 studies, an ANN model performed as well as logistic regression in 50% of all cases, and ANNs performed as well as logistic regression in 7 of the 8 largest studies, with a sample size of >5000 patients.1

Nevertheless, we believe that the time to apply ANN tools to improve the care of prostate cancer patients is long overdue. It is well recognized that multifactorial analysis integrating clinical expertise and patient values is required for a reliable diagnosis in patients with prostate cancer, and is a precondition of optimal treatment. Snow et al.3 introduced two ANN models for the diagnosis and prognosis of prostate cancer in 1994. Since then, several other approaches have been published. ANN architecture varies, with different input variables, hidden layers, and usually one output layer. There is also great variation in the percentage of training (60–80%) and testing (20–30%) data involved, and the need for final validation. Discussion of the use of different computational methods (e.g. Gaussian, sinusoid, and sigmoid), model solutions (e.g. classification and clustering), and accuracy testing methods is beyond the scope of this article.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

References

  1. Sargent DJ (2001) Comparison of artificial neural networks with other statistical approaches: results from medical data sets. Cancer 91 (Suppl): 1636–1642

    Article  CAS  Google Scholar 

  2. Schwarzer G and Schumacher M (2002) Artificial neural networks for diagnosis and prognosis in prostate cancer. Semin Urol Oncol 20: 89–95

    Article  Google Scholar 

  3. Snow PB et al. (1994) Artificial neural networks in the diagnosis and prognosis of prostate cancer: a pilot study. J Urol 152 (Pt 2): 1923–1926

    Article  CAS  Google Scholar 

  4. Thompson IM et al. (2004) Prevalence of prostate cancer among men with a prostate-specific antigen level ≥4.0 ng per milliliter. N Engl J Med 350: 2239–2246

    Article  CAS  Google Scholar 

  5. Finne P et al. (2000) Predicting the outcome of prostate biopsy in screen-positive men by a multilayer perceptron network. Urology 56: 418–422

    Article  CAS  Google Scholar 

  6. Stephan C et al. (2002) Multicenter evaluation of an artificial neural network to increase the prostate cancer detection rate and reduce unnecessary biopsies. Clin Chem 48: 1279–1287

    CAS  PubMed  Google Scholar 

  7. Partin AW et al. (1997) Combination of prostate-specific antigen, clinical stage, and Gleason score to predict pathological stage of localized prostate cancer. A multi-institutional update. JAMA 277: 1445–1451

    Article  CAS  Google Scholar 

  8. Han M et al. (2001) Evaluation of artificial neural networks for the prediction of pathologic stage in prostate carcinoma. Cancer 91 (Suppl): 1661–1666

    Article  CAS  Google Scholar 

  9. Anagnostou T et al. (2003) Artificial neural networks for decision-making in urologic oncology. Eur Urol 43: 596–603

    Article  Google Scholar 

  10. Veltri RW et al. (2002) Comparison of logistic regression and neural net modeling for prediction of prostate cancer pathologic stage. Clin Chem 48: 1828–1834

    CAS  PubMed  Google Scholar 

  11. Khan J et al. (2001) Classification and diagnostic prediction of cancers using gene expression profiling and artificial neural networks. Nat Med 7: 673–679

    Article  CAS  Google Scholar 

  12. Naguib RN et al. (1998) Neural network analysis of combined conventional and experimental prognostic markers in prostate cancer: a pilot study. Br J Cancer 78: 246–250

    Article  CAS  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. Biomarkers of Prostate Cancer Group at the University Hospital Charité, Humboldt University, Berlin

    Carsten Stephan, Henning Cammann & Klaus Jung

  2. Department of Urology,

    Carsten Stephan, Henning Cammann & Klaus Jung

  3. Research Division,

    Carsten Stephan, Henning Cammann & Klaus Jung

  4. Institute for Medical Biometry, Berlin

    Carsten Stephan, Henning Cammann & Klaus Jung

Authors
  1. Carsten Stephan
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Henning Cammann
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Klaus Jung
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Carsten Stephan.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Supplementary information

Rights and permissions

Reprints and permissions

About this article

Cite this article

Stephan, C., Cammann, H. & Jung, K. Artificial neural networks: has the time come for their use in prostate cancer patients?. Nat Rev Urol 2, 262–263 (2005). https://doi.org/10.1038/ncpuro0207

Download citation

  • Received:

  • Accepted:

  • Issue Date:

  • DOI: https://doi.org/10.1038/ncpuro0207

This article is cited by

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing