Date on Master's Thesis/Doctoral Dissertation

5-2016

Document Type

Master's Thesis

Degree Name

M.S.

Department

Bioinformatics and Biostatistics

Degree Program

Biostatistics, MS

Committee Chair

Wittliff, James L.

Committee Co-Chair (if applicable)

Brock, Guy N.

Committee Member

Lorenz, Douglas J.

Committee Member

Kerber, Richard A.

Author's Keywords

breast cancer; recurrence; peptide hormones; survival; LASSO; laser capture micro-dissection

Abstract

Certain hormones and/or receptors influencing normal cellular pathways were detected in breast cancers. The hypothesis is that gene subsets predict risk of breast carcinoma recurrence in patients with primary disease. Gene expression of 55 hormones and 73 receptors were determined by microarray with LCM-procured carcinoma cells of 247 de-identified biopsies. Univariate and multivariate Cox regressions were determined using expression levels of each hormone/receptor gene, individually or as a pair. Significant genes derived for each subset were analyzed to predict risk of cancer recurrence with 1000 LASSO training/test sets. A 14-gene molecular signature was identified for predicting clinical outcome without regard to estrogen or progestin receptor status of biopsies. A three-gene signature was derived for ER+ cancers while a 9-gene signature was deciphered for ER- cancers. Molecular signatures derived were compared with results in public databases. Collectively, results suggest gene subsets in primary breast cancer have been identified that predict recurrence.

Download

Share

COinS