Expression profiling using proteomic techniques has a great potential to identify new biomarkers that might help to better diagnose and treat diseases such as breast cancer, which is one of the leading causes of cancer death in women. Surface-enhanced laser desorption ionization mass spectrometry (SELDI-MS) combines chromatographic separation of peptides and proteins with mass spectrometry and is a fast, user-friendly tool to analyze protein and peptide profiles. SELDI-MS was employed for a comparative analysis of lobular invasive versus ductal invasive breast tumors to find differentially expressed proteins and peptides, and to validate this technique for biomarker identification using complex samples such as tissue. After optimization of sample preparation using HMEC and MCF-7 cell lines, 20 breast tumors were analyzed, and about 550 mass signals corresponding to an estimated 140 native peptides and proteins were detected in each tumor. Only 14% of the mass signals were present in more than six tumors of one subgroup or in more than 12 tumors of both groups showing a great overall heterogeneity of the peptide and protein profiles obtained. Peptide mass signals specific for each of the analyzed groups were identified. In addition, we detected peptides from laser-microdissected ductal invasive and intraductal tumor parts corresponding to peptides present in whole tumors. The low amount of identified peptides and proteins and the observed heterogeneity suggest that SELDI-MS is not well suited for biomarker identification of and profiling experiments on complex samples such as tumor tissue.