Breast cancer is a heterogenous disease and it is of importance to select patients with regard to different prognosis and treatment sensitivity to individualize treatment regimes. In this study we successfully adapted a protein extraction protocol from mRNA extracted tumor samples enabling two-dimensional gel electrophoresis (2-DE) analysis of samples previously analyzed by cDNA microarray. The aim was to find candidate proteins that distinguish breast cancer patients with or without recurrences after adjuvant CMF (cyclophosphamide, methotrexate and 5-FU) treatment within four years to follow-up. We identified several proteins distinguishing the recurrence group from the non-recurrence group, especially in the ER and PgR positive subgroup (n = 7). The induced proteins were involved in translation/folding, iron ion binding, and protease inhibition, whereas proteins involved in signaling, ubiquitination, and splicing were decreased in expression. These results show that it is possible to use 2-DE to separate high abundant proteins in breast cancer tissue and to find discriminating proteins to identify patients with different prognosis after adjuvant CMF treatment.