A meta-analysis of recent data from the literature underscores the considerable body of present knowledge concerning prostate carcinogenesis, in part due to the numerous molecular biology tools now at our disposal. As concerns early events, much interest is being paid to modifications in the expression of GSTP1 and NKX3.1 occurring in totipotent stem cell populations. The discovery of fusion genes implicating TMPRSS2 and ERG (and, on rare occasions, other ETS family transcription factors) constitutes a major advance. Under physiological androgenic stimulation, the presence of these fusion genes leads to overexpression of genes involved in cell growth and differentiation. Concomitantly, alterations in numerous signalling pathways (growth factors, Wnt-beta catenine, PI3K/Akt) are responsible for the onset of an aggressive tumor phenotype. Hormono-independence is currently explained by an amplification of, or mutations in, androgenic receptors. These are facilitated by genomic instabilities linked to alterations in proteins which regulate gene expression, such as EZH2, and by the influence of the tumor microenvironment. Disturbances in the interactions between tumor cells and the microenvironment contribute to local extension of the tumor. Changes in the expression of E-cadherin are responsible for modifications in cell adhesion to the extracellular matrix. The expression of metalloproteases and of angiogenic factors favors tumor dissemination. Finally, the bone tropism in prostate metastases is probably linked to osteomimetic properties of prostate tumor cells which are capable of expressing certain proteins involved in bone remodelling, such as Runx-2, BSP (bone sialoprotein) and BMP (bone morphogenetic protein). Numerous studies remain to be carried out in order to correlate the identified genetic profiles and molecular anomalies with tumor prognosis. Nevertheless, the possibility of decrypting these anomalies for use in therapeutic applications is encouraging.