ABSTRACT
Recent reports suggest that immunohistochemistry (IHC) markers can be used to give prognostic information in breast cancer that is similar to that contained in the Genomic Health Inc. OncoDX recurrence score (Onco-RS). Our own previous work confirmed that an IHC model based on estrogen receptor (ER), progesterone receptor (PR), and Ki67 predicts 62% of the Onco-RS variability. Other markers used in the Onco-RS include proteins thought to increase tumoral invasive potential, and one such marker is matrix metalloproteinase-11, also called stromelysin 3 (ST3). The goal of this study is to examine the additional value of including ST3 in an IHC-based model that also includes ER, PR, and Ki67 in predicting the Onco-RS, as compared with an IHC model based only on ER, PR, and Ki67 (IHC-RS). The patient population consists of a retrospectively identified cohort of 91 women with ER-positive, HER2neu-negative breast cancer who completed OncoDX testing. Using stepwise multiple regression incorporating Ki67 percentage and semiquantitative ER, PR, and ST3 scores, the ST3 score was not statistically significant.
Subject(s)
Breast Neoplasms/diagnosis , Breast Neoplasms/pathology , Immunohistochemistry/standards , Matrix Metalloproteinase 11/metabolism , Adult , Aged , Biomarkers, Tumor/metabolism , Cohort Studies , Female , Humans , Immunohistochemistry/methods , Ki-67 Antigen/metabolism , Middle Aged , Neoplasm Recurrence, Local , Predictive Value of Tests , Prognosis , Receptor, ErbB-2/metabolism , Receptors, Estrogen/metabolism , Receptors, Progesterone/metabolism , Research Design/standards , Retrospective StudiesABSTRACT
Recent reports suggest that immunohistochemistry (IHC) markers can be used to give prognostic information in breast cancer that is similar to that contained in the Genomic Health Inc. OncoDx recurrence score (Onco-RS). The goal of this study is to examine the potential prognostic value of a score derived from results of a simple set of IHC tests in the prediction of the Onco-RS. A score (IHC-RS) was derived to predict the Onco-RS using IHC-based quantitative and semiquantitative results from a subset of markers selected from those used in the generation of an Onco-RS score. The patient population consists of a retrospectively identified cohort of 158 women with ER-positive, HER2neu-negative breast cancer who completed OncoDx testing. A predictive model was developed to generate the IHC-RS using stepwise multiple regression incorporating Ki67 percentage and semiquantitative ER and PR scores. Using only these 3 IHC markers, the IHC-RS predicted 62% of the Onco-RS variability (adjusted R=0.624, P=0.004). In addition, analysis of outliers in the correlation between the IHC-RS and the Onco-RS reveals the possibility of sampling error as a drawback of the Onco-RS. This is contrasted against potential interlab and intralab variability and preanalytic issues that may negatively impact the implementation of an IHC-RS.
