Serum Protein Signatures for Breast Cancer Detection in Treatment-Naïve African American Women Using Integrated Proteomics and Pattern Analysis

Breast cancer is the leading cause of cancer-related mortality in African American (AA) women. In this study we evaluated the serum proteomic profile of AA women with breast cancer using an integrated proteomic framework with multivariate pattern analysis. Using 2D-DIGE, thousands of serum protein spots were detected across 33 gels; 46 spots met criteria for presence, statistical significance, and differential expression. Proteins from the spots were identified by MALDI-TOF/TOF and matched in curated databases, highlighting serum biomarkers including ceruloplasmin, alpha-2-macroglobulin, complement component C3 and C6, alpha-1-antitrypsin, alpha-1B-glycoprotein, alpha-2-HS-glycoprotein and haptoglobin-related protein. LC-MS/MS analysis revealed 163 differentiating peptides after imputing and filtering 286 peptides. These were evaluated using cumulative distribution function (CDF) analysis, a nonparametric method suited for limited sample sizes. Peptide patterns were explored with Random Forest, showing concordance with CDF. The model achieved an AUC of 0.85 at the peptide level. This workflow identified differentiating proteins (CERU, A2MG, CO3, VTDB, HEMO, APOB, APOA4, CFAH, CO4A, AACT, K1C10, ITIH2, ITIH4), highlighting CERU, A2MG, and CO3 with overexpression and reproducible identification across platforms. We present an integrated, non-invasive serum protein biomarker signature panel specific to AA women, through reproducible proteomic sensor framework to support early detection and breast cancer prevention.