In Reply We thank Drs Braillon and Menkes for their letter. Indeed, in our investigation we found that the majority of new cancer drug approvals are made on the basis of a surrogate.1 Surrogate end points were used both in cases of accelerated (or provisional) approvals, as well as in cases of traditional (or regular) approvals. The use of surrogates in the context of traditional approvals is particularly troubling because, practically speaking, marketing authorization is unlikely to be revoked if efficacy on the patient centered outcomes (ie, survival or quality of life) is subsequently not demonstrated. We believe that all cancer drugs approved on the basis of a surrogate should have postmarketing commitments to establish efficacy against patient-centered outcomes. For this reason, we would advise caution with the use of surrogates to support traditional approval. For instance, everolimus was approved for metastatic breast cancer in combination with exemestane based on a prolongation of progression-free survival, but it later became clear the drug does not improve overall survival.2 Yet, because it received traditional approval, it appears unlikely that approval will be revoked. The repeated failure of everolimus to improve survival for at least 3 cancer indications is the subject of a recent piece of investigative journalism.3