raise serious questions concerning the efficacy and safety of these VEGF antagonists. Overall survival (OS) improvement is the fundamental principle of any new therapy. Instead of improving, could anti-angiogenic agents, specifically bevaci-zumab, actually reduce OS? Although this provocative finding requires confirmation, a potential decrease in OS by administra-tion of the widely used drug bevacizumab in a metastatic setting has a central role in treatment decision making. Why is pro-gression-free survival (PFS) selected as a primary end point in several clinical trials, less appropriate than OS? Is the identifica-tion of new biomarkers for patient selec-tion essential to prove the clinical utility of anti-angiogenic therapy only in specific subgroups of patients and not generally in all patients? Over the last decade, molecularly tar-geted therapy has developed rapidly. Adding targeted agents to traditional sur-gery, radiotherapy and chemotherapy rep-resents the major hope in the war against solid cancers. The concept is scientifically rational. Targeting signaling pathways that are deregulated and responsible for can-cer progression and metastasis can lead to effective cancer treatment. However, most Phase III trials without marker-based selection of patients were negative. This treatment failure raises the sugges-tion of a change in drug development and clinical trial design towards personalized genotype–phenotype predictions