46 Background: Numerous molecular pathways have been implicated in aggressive prostate cancer. However, it has been difficult to gain an overview of the dominant pathways involved. By analyzing multiple prostate cancer gene signatures of poor prognosis through a protein interaction network, we are able to in an unbiased manner prioritize shared key biological mechanisms. Methods: We evaluate prostate gene signatures through a previously published Single Protein Analysis of Network (SPAN) methodology to develop a prostate cancer network signature. We assess this signature using a Gene Ontology based information-theoretic metric as well as assess its prognostic ability in an independent clinical dataset. We then examine the interconnectivity of the network signature to develop a phenotype-pathway map that provides an executive summary of critical mechanisms involved in advanced prostate cancer. Results: Of the 13 prostate cancer signatures that were evaluated, nine gene signatures plus a previously unpublished gene signature met criteria for evaluation. SPAN analysis of these protein interactions resulted in a 42-gene network signature. This network signature had the highest information-theoretic score based on Gene Ontology annotation similarity. Further, the signature identified a statistically significant survival difference (p=0.009) in a cohort of men with prostate cancer in an independent dataset. Additional analysis of the signature generated a phenotype-pathway map that not only recapitulated the centrality of PIK3/NF-κB signaling, but also highlighted less well-established pathways such as the JAK2 kinase activation cascade. Conclusions: SPAN analysis provides a robust means of abstracting disparate prostate cancer gene expression signatures into clinically useful, prioritized pathways that may ultimately serve as attractive, therapeutic targets. No significant financial relationships to disclose.