Mammography remains the diagnostic test of choice for breast cancer, but 20% of cancers still go undetected. Many serum biomarkers have been reported for breast cancer but none have proven to represent effective diagnostic strategies. ProteinChip mass spectrometry is an innovative technology that searches the proteome for differentially expressed proteins, allowing for the creation of a panel or profile of biomarkers. The objective of this study was to construct unique cancer-associated serum profiles that, combined with a classification algorithm, would enhance the detection of breast cancer Pretreatment serum samples from 134 female patients (45 with cancer, 42 with benign disease, 47 normal) were procured prospectively following institutional review board-approved protocols. Proteins were denatured, applied onto ProteinChip affinity surfaces, and subjected to surface enhanced laser desorption/ionization (SELDI) time-of-flight mass spectrometry. The SELDI output was analyzed using Biomarker Pattern Software to develop a classification tree based on group-specific protein profiles. The cross-validation analysis of cancer versus normal revealed sensitivity and specificity rates of 80% and 79%, and for cancer versus benign disease, 78% and 83%, respectively. When 2 different chip surfaces were combined the sensitivity and specificity increased to 90% and 93%, respectively. The sensitivity and specificity of this technique are comparable to those of mammography and, if confirmed in a larger study, this technique could provide the means toward development of a simple blood test to aid in the early detection of breast cancer. The combination of SELDI ProteinChip mass spectrometry and a classification- and regression-tree algorithm has the potential to use serum protein expression profiles for detection and diagnosis of breast cancer.