The majority of drugs approved for the treatment of malignant disease are traditional cytotoxic agents that, in many cases, have been in use for decades. In the recent past, we have seen the approval of the so-called targeted agents and with this have emerge concepts such as biomarker and optimal biological dose, but which came first and what is actually behind the paradigm shift that is all too evident in modern oncology drug development? Following critical examination of the issue, it can be argued that many, if not all, cytotoxic chemotherapies are targeted, for example, methotrexate, the folate pathway and the use of neutropenia as a biomarker although not titled as such, in the development of these agents. The most obvious change is the toxicity profile of the new molecular entities and, therefore, the recognition that driving a dose towards a maximally tolerated dose for use in later phase development is inappropriate. Again, it can be argued that this should never have been appropriate, however, the tools necessary to determine the underlying pharmacokinetic/pharmacodynamic (PK/PD) relationships were lacking. We believe this has brought about the most dramatic change in how oncology drugs are developed, rather than the classification as cytotoxic or targeted. We will argue that it is not practicable to develop a targeted agent using the traditional paradigm, but equally, the same would now be true for cytotoxics.