The NIH Consensus Development Conference in November 2000 concluded that adjuvant hormonal therapy should be offered to all patients with tumors expressing estrogen receptor (ER) and/or progesterone receptor (PR), assessed by immunohistochemistry. However, numerous studies have shown the response rate in this group will be, at best, just over 50%. Furthermore, Tamoxifen, the most commonly used hormonal therapy is known to cause a number of side affects, some of which are life threatening. Thus there is a need to determine which breast cancer patients will respond to tamoxifen, or newer hormonal therapies, in a manner more specific than the nearly 20 year old method of subjective estimation of estrogen or progesterone receptor expression. The objective of this proposal is to use in situ protein expression analysis to define patients that will NOT respond to hormonal therapy. We propose two specific aims: 1. To use AQUA-based analysis of protein expression in a retrospective series to develop a diagnostic cocktail and cut-point that defines hormonal therapy non-responders. 2. To complete a small pseudo-prospective study to assess breast cancer core needle biopsies to determine the predictive value of the optimized expression cocktail for selection of the patients that will NOT respond to hormonal therapy. To achieve these aims, we will construct a richly annotated tissue microarray from tumors from a series of patients who have received hormonal therapy from 3-5 years ago. We will then analyze this array, using AQUA-based quantitative analysis, for protein expression of ER, PR and a series of related molecules. AQUA is a novel set of algorithms developed in our lab that uses molecular (rather than morphologic) compartmentalization to measure protein expression. We have shown that it's accuracy is comparable to ELISA assays but without loss of spatial information which occurs in biochemical tissue extraction. We will analyze the AQUA scores using a novel bio-informatics model developed in our lab that allows optimal cut-point determination integrating outcome data. When this is done for each protein in the series, we will define the combination of expression levels that best predicts non-response to hormonal therapy. Finally, we will use the AQUA software to analyze the newly discovered marker set on recuts of core needle biopsies from a pseudo-prospective series of patients with known response to hormonal therapy toward the goal of predicting therapeutic response on actual biopsy specimens.
Showing the most recent 10 out of 23 publications
