Estrogen receptor-positive breast cancers from African American women uniquely identified using Principal Components Analysis. Florin M. Selaru, Jing Yin, Yan Xu, Yuriko Mori, Suna Wang, Andreea Olaru, Elena Deacu, Fumiaki Sato, John M. Abraham, David Shibata, Stephen J. Meltzer, Division of Gastroenterology, Department of Medicine and Greenebaum Cancer Center, University of Maryland, Baltimore, Baltimore MD Aim: To identify the most important clinical features influencing global gene expression in breast cancers. Materials and Methods: In order to identify these features in an unbiased fashion, we applied Principal Components Analysis (PCA) to cDNA microarray data from 36 breast cancers. This bioinformatics method identifies levels of meaning within complex data. These levels of meaning can in turn be correlated with clinical features, identifying those exerting a major influence on global gene expression. Results: 35 principal components (PC) were found and ranked in order of their impact on gene expression. These PCs were then correlated with clinicopathological features. PC 7 identified a unique subgroup consisting of estrogen receptor (ER)-positive (+) African-American (AA) patients. This group exhibited global expression profiles significantly different from both ER (-) AA women and Caucasian women (regardless of ER status). Thus, PC 7 placed ER (+) AA breast cancers into one group and all of the remaining tumors into the second group, regardless of race or ER status (p=0.0002, one-way ANOVA). Furthermore, additional components also correlated significantly with other clinicopathologic features. These components and the features with which they correlated were: PC 1, with year of surgical procedure (p=2*10^-7); PC 4, with lymph node (LN) metastasis (p = 0.04); PC 5, with race per se (Caucasian vs. African-American (p = 0.002); PC 6, with histologic type (strictly intraductal vs. lobular or mixed, p=0.01); PC 9, with age (less vs. more than 50 years old, p = 0.003); PC 10, with tumor stage (stage 2 vs. stage 3, p=0.007); PC 12, with Elston's grade (stages 1, 2 and 3, p=0.04); PC 23, with tumor size (less vs. more than 5 cm, p = 0.009); and PC 25, with progesterone receptor (PR) status (p = 0.03). All p values were calculated using the one-way analysis of variance (ANOVA) test. Conclusions: Our findings provide, for the first time, a global gene expression basis for the existence of estrogen receptor-positive African-American breast tumors as a biologically distinct cancer subgroup. Moreover, this study demonstrates the power of PCA to detect gene expression bases for other clinical features, such as age, race per se, tumor size, PR status, LN metastasis, or other as-yet uncharacterized tumor parameters. Thus, the application of PCA to cDNA microarray data offers the potential to identify and explain important biologic aspects of malignancy. AACR Abstract Number: LB-130, 2003 Remember we are NOT Doctors and have NO medical training. This site is like an Encylopedia - there are many pages, many links on many topics. Support our work with any size DONATION - see left side of any page - for how to donate. You can help raise awareness of CAM.