Keeping data continuous when analyzing the prognostic impact of a tumor marker: an example with cathepsin D in breast cancer

Breast Cancer Res Treat. 2003 Nov;82(1):47-59. doi: 10.1023/B:BREA.0000003919.75055.e8.

Abstract

The prognostic value of cathepsin D has been recently recognized, but as many quantitative tumor markers, its clinical use remains unclear partly because of methodological issues in defining cut-off values. Guidelines have been proposed for analyzing quantitative prognostic factors, underlining the need for keeping data continuous, instead of categorizing them. Flexible approaches, parametric and non-parametric, have been proposed in order to improve the knowledge of the functional form relating a continuous factor to the risk. We studied the prognostic value of cathepsin D in a retrospective hospital cohort of 771 patients with breast cancer, and focused our overall survival analysis, based on the Cox regression, on two flexible approaches: smoothing splines and fractional polynomials. We also determined a cut-off value from the maximum likelihood estimate of a threshold model. These different approaches complemented each other for (1) identifying the functional form relating cathepsin D to the risk, and obtaining a cut-off value and (2) optimizing the adjustment for complex covariate like age at diagnosis in the final multivariate Cox model. We found a significant increase in the death rate, reaching 70% with a doubling of the level of cathepsin D, after the threshold of 37.5 pmol mg(-1). The proper prognostic impact of this marker could be confirmed and a methodology providing appropriate ways to use markers in clinical practice was proposed.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Aged
  • Biomarkers, Tumor / analysis*
  • Breast Neoplasms / metabolism*
  • Breast Neoplasms / mortality*
  • Breast Neoplasms / pathology
  • Cathepsin D / analysis*
  • Cohort Studies
  • Female
  • Humans
  • Lymphatic Metastasis
  • Middle Aged
  • Multivariate Analysis
  • Prognosis
  • Receptors, Estrogen / analysis
  • Receptors, Progesterone / analysis
  • Retrospective Studies
  • Survival Analysis

Substances

  • Biomarkers, Tumor
  • Receptors, Estrogen
  • Receptors, Progesterone
  • Cathepsin D