PDF - 160KB, Supplementary Figure 1 -Histograms of BUM model to the P-values from the ANOVA test of race/ethnicity by breast cancer subtype in 177 proteins. Supplementary Figure 2 - Histograms of BUM model to the P-values from the ANOVA test of race/ethnicity by breast cancer subtype in 16,438 probe sets. Supplementary Figure 3 - Unsupervised clustering of the significant 384 probe sets with FDR <0.05 among the HER2-positive patients show differences according to race (A). Validation cohort of 59 different HER2-positive tumors across the significant 384 probes showed no pattern in the race comparisons (B). The row side bar color displays the race (White: red; Hispanic: blue; Black: black; Others: green)
ARTICLE ABSTRACT
Background: Differences in gene or protein expression patterns between breast cancers according to race/ethnicity and cancer subtype.Methods: Transcriptional profiling was performed using Affymetrix HG-U133A platform in 376 patients and reverse phase protein array analysis (RPPA) was done for 177 proteins in 255 patients from a separate cohort. Unsupervised clustering was conducted, as well as supervised comparison by race and tumor subtype. Standard statistical methods, BRB-Array tools, and Ingenuity Pathways software packages were used to analyze the data.Results: Median age was 50 years in both the cohorts. In the RPPA cohort, 54.5% of the tumors were hormone receptor–positive (HR-positive), 20.7% HER2-positive, and 24.71% triple-negative (TNBC). One hundred and forty-seven (57.6%), 47 (18.43%), and 46 (18.1%) of the patients were White, Hispanic, and Black, respectively. Unsupervised hierarchical clustering of the protein expression data showed no distinct clusters by race (P values were 0.492, 0.489, and 0.494 for the HR-positive, HER2-positive, and TNBC tumors respectively). In the gene expression cohort, 54.2% of the tumors were HR-positive, 16.5% HER2-positive, and 29.3% TNBC. Two hundred and sixteen (57.5%), 111 (29.52%), and 32 (8.52%) patients were White, Hispanic, and Black, respectively. No probe set with a false discovery rate (FDR) of <0.05 showed an association with race by breast cancer subtype; similar results were obtained using pathway and gene set enrichment analysis methods.Conclusions: We did not detect a significant variation in RNA or protein expression comparing different race/ethnicity groups of women with breast cancer.Impact: More research on the complex network of factors that result in outcomes differences among race/ethnicities is needed. Cancer Epidemiol Biomarkers Prev; 23(2); 316–23. ©2013 AACR.