This study is a meta-epidemiological study which aim is to quantify the clinical benefit of the biomarker-based strategy compared to conventional strategy across all drugs with such a strategy.

Personalized medicine consist to differentially treat patients based on their individual characteristics (mainly genetics). It is of the most promising area of cancer research and cancer care. The label of more than 140 FDA- approved drugs mention a biomarker, the majority being indicated in oncology. However, it has also been suggested that the hopes of personalized medicine were not matched by evidence. Indeed, there is a threat that genetic biomarkers are used without evidence that this use translates in improved outcomes for patients, and the use of biomarkers is in need for thorough validation.

In a previous work the investigators showed that the mention of a pharmacogenomic biomarker in a drug label can have different meanings depending on the drug, and that oncology had higher proportion of required or recommended genetic testing compared to other therapeutic areas. Hence, the investigators will include only the drug-biomarker pairs with (i) required or recommended genetic testing, or (ii) with biomarker-based indication and (iii) with at least one indication in oncology.

Inclusion Criteria:

• phase II or III randomized clinical trials
• including patients with one of the drug-biomarker-indication triplet studied
• evaluating treatment effect in biomarker-defined strata (either in biomarker-positive patients, in biomarker-negative patients or in both strata).
• reporting OS or PFS or DFS

Exclusion Criteria:

• None

This is a systematic review. No patient will be directly included in this study, but the investigators will include clinical trials.