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:

How to judge a tumor marker

Nature Clinical Practice Oncology volume 2, pages 482–483 (2005)Cite this article

Evaluating the PROGNOSTIC VALUE of tumor markers is an important part of oncology research. This journal recently published a set of guidelines1 meant to encourage more clarity and comprehensiveness in the reporting of such studies, and we applaud the effort. Points 16 and 17 of these guidelines, however, appear to assume that the new marker's prognostic value will be inferred from its coefficient in a MULTIVARIATE REGRESSION MODEL.1 This is certainly the approach taken in many prognostic studies, including some of the ones underlying the ASCO recommendation that estrogen and progesterone receptors can be used as markers for breast cancer.2 Unfortunately this approach addresses only indirectly the central research question of a prognostic study: does the new marker enhance our ability to predict a clinically important outcome?

Suppose we are evaluating a new blood marker intended to help predict whether a prostatectomy patient will experience a recurrence during the first year after surgery. We measure this marker in a sample of patients immediately after surgery and follow them for 1 year. To assess the predictive value of the marker, we fit a logistic regression MODEL to the data; the dependent variable is recurrence, and the independent variables include both the new marker and the prognostic factors currently in use. From this model, we estimate the ODDS RATIO associated with the marker to be 2.37, with a 95% CI of 1.55–2.80 and a P value of 0.0002. Under the traditional approach, the results would be taken as evidence of strong prognostic value. But what do they really tell us?

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. McShane LM et al. (2005) Reporting recommendations for tumor marker prognostic studies (REMARK). Nat Clin Pract Oncol 2: 416–422

    CAS  PubMed  Google Scholar 

  2. Bast RC et al. (2001) 2000 Update of recommendations for the use of tumor markers in breast and colorectal cancer: clinical practice guidelines of the American Society of Clinical Oncology. J Clin Oncol 19: 1865–1878

    Article  PubMed  Google Scholar 

  3. Chatfield C (1995) Model uncertainty, data mining and statistical inference. J R Stat Soc [Ser A] 158: 419–466

    Article  Google Scholar 

  4. Kattan MW (2003) Judging new markers by their ability to improve predictive accuracy. J Natl Cancer Inst 95: 634–635

    Article  CAS  PubMed  Google Scholar 

  5. Harrell FE et al. (1996) Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med 15: 361–387

    Article  PubMed  Google Scholar 

  6. Kattan MW and Gonen M Prediction from multivariable regression models. Urol Oncol, in press

  7. Breiman L (2001) Statistical modeling: the two cultures. Stat Sci 16: 199–231

    Article  Google Scholar 

  8. Efron B and Tibshirani RJ (1998) An introduction to the bootstrap. NY: Chapman & Hall/CRC

    Google Scholar 

  9. Ehrenberg ASC and Bound JA (1993) Predictability and prediction. J R Stat Soc [Ser A] 156: 167–206

    Article  Google Scholar 

  10. Hayes DF et al. (1996) Tumor Marker Utility Grading System: a framework to evaluate clinical utility of tumor markers. J Natl Cancer Inst 88: 1456–1466

    Article  CAS  PubMed  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. biostatistician in the Department of Quantitative Health Sciences at the Cleveland Clinic Foundation,

    Ethan M Katz MS

  2. the Chairman of the Department of Quantitative Health Sciences at the Cleveland Clinic Foundation,

    Michael W Kattan PhD

Authors
  1. Ethan M Katz MS
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Michael W Kattan PhD
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Ethan M Katz MS.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Related links

Related links

Related links in Nature Research

Reporting recommendations for tumor marker prognostic studies (REMARK)

Rights and permissions

Reprints and permissions

About this article

Cite this article

Katz, E., Kattan, M. How to judge a tumor marker. Nat Rev Clin Oncol 2, 482–483 (2005). https://doi.org/10.1038/ncponc0318

Download citation

  • Received:

  • Accepted:

  • Issue Date:

  • DOI: https://doi.org/10.1038/ncponc0318

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