Clinical T rial D esigns f or t he E arly C linical D evelopment of T herapeutic C ancer V accines

There are major differences between ther- apeutic tumor vaccines and chemotherapeutic agents that have important implications for the design of early clinical trials. Many vaccines are inherently safe and do not require phase I dose finding trials. Patients with advanced cancers and compromised immune systems are not good candidates for assessing either the toxicity or efficacy of therapeutic cancer vaccines. The rapid pace of development of new vaccine candidates and the variety of possible adjuvants and modifications in method of administration makes it important to use efficient designs for clinical screening and evaluation of vaccine regimens. We review the potential advantages of a wide range of clinical trial designs for the develop- ment of tumor vaccines. We address the role of immu- nological endpoints in early clinical trials of tumor vac- cines, investigate the design implications of attempting to use disease stabilization as an end point and discuss the difficulties of reliably utilizing historical control data. Several conclusions for expediting the clinical development of effective cancer vaccines are proposed. J Clin Oncol 19:1848-1854. © 2001 by American Society of Clinical Oncology. D EVELOPMENT OF therapeutic cancer vaccines is a major area of oncologic research today. Whereas the important principles for the design of phase III trials apply to both tumor vaccines and chemotherapeutic drugs, there are major differences between these two classes of thera- peutics that have important implications for early clinical development. First, the phase I concept of dose escalation to find a maximum-tolerated dose does not apply to most vaccines. Most vaccines are incapable of causing immediate serious or life-threatening toxicities at doses feasible to manufacture. Second, neither toxicity nor efficacy can be assessed in patients with advanced malignant disease asso- ciated with a blunted immune response because both toxic- ity and efficacy depend on the immune response. Conse- quently, issues of patient selection and end point definition require reconsideration for the early phases of vaccine development. Third, vaccination strategies often combine multiple agents, each of which needs to be optimized, such as adjuvants, cytokines, or costimulatory molecules. Al- though some of these combination regimens can be opti- mized in preclinical models, there remain comparisons that need to be performed in human patients. Thus, trial designs that allow the rapid screening of multiple variations on a regimen would greatly facilitate the development of cancer vaccines. The purpose of this article is to address these and other issues and to provide approaches that facilitate the clinical development of therapeutic cancer vaccines. Early clinical development of tumor vaccines can be time consuming, costly, and ineffective. Much time can be wasted on unnecessary phase I trials or on phase II trials that ask phase III questions. We hope that this article helps investigators to design vaccine trials that are efficient, in terms of duration and numbers of patients treated, and yield valid and useful information to facilitate further vaccine development.