Send to

Choose Destination
See comment in PubMed Commons below
Clin Breast Cancer. 2006 Dec;7(5):372-9.

Gene expression profile assays as predictors of recurrence-free survival in early-stage breast cancer: a metaanalysis.

Author information

University of Rochester School of Medicine and Dentistry, Rochester, NY 14642, USA.



Several investigators have reported efforts to define gene expression signatures based on prediction of survival in early-stage breast cancer. The analysis reported here reviews test performance characteristics of reported gene expression signatures in women with breast cancer.


All published reports of gene expression profiling in breast cancer were sought through an extensive search of the published literature. Seventeen cohorts of patients with primary breast cancer were identified reporting on the relationship between a gene expression signature and recurrence-free survival. Several measures of test performance were evaluated including sensitivity, specificity, likelihood ratio, predictive value, and the diagnostic odds ratio as an overall measure of test performance.


Reported series included 2908 patients ranging from 20 to 668 per study. Seven cohorts were evaluated using cross validation techniques, and 10 were studied in independent cohorts. Overall, 52.6% of patients were classified as high risk and 20.5% experienced disease recurrence. The false negative rate was > 20% in 8 studies (47%) and false positive rate was > 50% in 6 (35%). The number of genes in the assay correlated with assay sensitivity (rsp= 0.537; P = 0.032), the positive predictive value (rsp = 0.501; P = 0.048), and the diagnostic odds ratio (rsp = 0.532; P = 0.041).


Gene expression profiles based on microarray analysis show early promise for predicting survival in patients with breast cancer. However, the use of these assays in therapeutic decision-making must consider the limitations of assay test performance and the specific patient population being evaluated.

[Indexed for MEDLINE]
PubMed Commons home

PubMed Commons

How to join PubMed Commons

    Supplemental Content

    Full text links

    Icon for Elsevier Science
    Loading ...
    Support Center