Novel treatments for cancer are tested initially in phase 1 trials enrolling patients with advanced disease who have exhausted standard treatment options. Although these trials are designed to evaluate safety and to define dosing for future efficacy trials, most patients volunteer with the hope of obtaining medical benefit. Do phase 1 oncology trials promote a 'therapeutic misconception' among eligible patients about the personal meaning of trial participation, or do they offer them a reasonable prospect of direct medical benefit as compared with available alternatives? Recent evidence on outcomes of phase 1 oncology trials is examined systematically, with the aim of accurately assessing the prospect of direct medical benefit for participants and drawing implications for informed consent. We argue that, in view of important uncertainties, aggregate data from phase 1 trials relating to the surrogate outcomes of tumor shrinkage and stable disease do not permit any definitive estimate of a 'clinical benefit rate.' Nevertheless, these trials do offer participants a prospect of direct medical benefit. As a result, accurately informed patients may reasonably decide to enroll in phase 1 oncology trials in hopes of obtaining benefit, after considering the anticipated risks and available clinical alternatives. Motivation to enroll in these studies to receive personal benefit does not, in itself, compromise informed consent.