Abstract: Prostate cancer (PC) is prevalent in both indolent and lethal forms. Although many PC patients diagnosed today have organ-confined disease curable by prostatectomy or radiation therapy, 20-30% of PCs will relapse to lethal disease within 5-years of treatment. Relatively little attention has been paid to distinguishing the molecular characteristics of proven lethal metastatic PC from non-lethal cancers. A better understanding of the origins and evolution of lethal cancers should allow screening and treatment to be better tailored to the needs of each patient. To this end, we performed an integrative molecular profiling of a lethal PC from one patient (A21). High-coverage whole genome sequence, transcriptome sequence, and methylation analysis was performed on 9 anatomically separate metastases obtained by autopsy, and targeted sequencing was performed in multiple cancerous and noncancerous foci within the radical prostatectomy specimen removed 5 years prior to death. Molecular results were analyzed in relation to detailed clinical data. Integrated whole genome sequence analysis revealed convergent evolution of AR gene amplification events, and inception of p.L702H mutation in the AR present only in liver metastases. In addition, the analysis showed parallel increases in AR regulated transcripts in the liver metastases, suggesting a dominant effect by the mutation. Mutation of PI3/PI4 kinase member PIK3CG was found in all metastases but in no primary tumor foci studied. The study demonstrated the power of an integrated approach to interrogate clonal evolution of cancer suggesting that such studies are valuable on the individual level and could lead the way towards personalized treatment. In the individual studied, cessation of corticosteroid treatment and/or therapeutic manipulation of PI3K/AKT/mTOR activity theoretically could have provided a personalized benefit. These findings suggest that similar integrated analysis in large cohorts of patients with metastatic cancer could accelerate progress in establishing effective personalized cancer medicine.