Prostate cancer is the most common solid organ malignancy affecting men in the United States and Western Europe. Currently, the main diagnostic tools used to look for evidence of prostate cancer include physical examination using digital rectal exam (DRE), serum concentrations of prostate specific antigen (PSA) and biopsy. However, due to the low specificity of PSA in differentiating prostate cancer from other benign conditions, many patients undergo overtreatment for their disease. There is an urgent need for additional markers to improve the diagnostic accuracy for early stages of prostate cancer. Proteomic analysis of serum has the potential to identify such markers. An initial discovery study has been completed using 12 serum samples from patients with different grades of prostate cancer (Gleason score 5 and 7) undergoing radical prostatectomy. Serum samples were subjected to immunoaffinity depletion and protein expression analysis using 2D-DIGE. Image analysis isolated 63 spots that displayed differential expression between the Gleason score 5 and 7 cohorts (p < 0.05), 13 of which were identified as statistically significant using two independent image analysis packages. Identification of differentially expressed spots was carried out using LC-MS/MS. Because of their functional relevance and potential significance with regards to prostate cancer progression, two of these proteins, pigment epithelium-derived factor (PEDF) and zinc-alpha2-glycoprotein (ZAG), have undergone extensive validation in serum and tissue samples from the original cohort and also from a larger independent cohort of patients. These results have indicated that PEDF is a more accurate predictor of early stage prostate cancer. We are confident that proteomics-based approaches have the potential to provide more insight into the underlying molecular mechanisms of the disease and also hold great promise for biomarker discovery in prostate cancer.