||During the past few years, debates have frequently erupted in oncology journals regarding the question of whether to prolong or end treatment. These debates have been informed by developments from both within and outside the field. Within Bioethics, some writers have put forward a number of principles for judging the legitimacy of medical interventions, notably that of patient autonomy. Broad social and political developments have also profoundly affected medical practices at the end of life. Though therapeutic options have evolved, whether to stop or to pursue treatment in the face of certain death has been a central issue in medical oncology since the early 1950s. A critical appraisal of the history of this issue can help us to better understand the tangled relationship(s) between innovation, “cure,” death, and the symptoms and subjective experiences of sufferers. This paper addresses an aspect of this complex problem, namely how limits are established regarding both treatment and therapeutic innovation near the end of life. Utilizing a grounded theory and situational analysis approach it traces how the issues at stake were defined and the ways in which the dilemma was progressively transformed as a result of the combined effects of a proliferating number of stakeholders, molecules, instruments, and techniques. It discusses three different moments, as they epitomize how the links between chemotherapy and palliation were construed through the evolving forms of clinical research and innovative therapies.