This article examines the relationship between diagnosis and therapy, focusing on the case of leukemia and cancer chemotherapy in the 1960s. This case, I argue, reinforces the need to study diagnoses from a social-science perspective, because the persistent controversy around leukemia classification was resolved by institutional restructuring introduced through clinical experimentation, rather than by techno-scientific advances. In an attempt to prove that chemical cancer therapy was possible, oncologists replaced the question ‘Is this drug working?’ with the question ‘How can we make this drug work?’ To create the conditions and criteria under which drugs could work, oncologists undertook the reclassification of cancers and patients, producing a new diagnostic style that reversed the roles of diagnosis and therapy. Experts gained and secured the power to classify not by solving existing problems, but by redefining what counts as a problem and what qualifies as a solution. Similarly, therapies can become transformative not only when they ‘work’, but when they work just well enough to mobilize resources and support. Theorizing these displacements, I develop the concept of ‘adequate trials’ in order to capture modes of innovation in which a deep commitment to give new technologies a ‘fair chance’ to succeed (i.e. an ‘adequate trial’) leads experts to redefine the tasks and goals of their field. To further our theoretical understanding of how rigid drug testing becomes malleable and conducive to normative change, I analyze the organizational, scientific, and jurisdictional conditions that gave rise to oncologists’ practical orientations.