Interactive Clinical Trial Design

Abstract

Clinical trials have been traditionally carried out in a "trial and error" fashion, which is highly inefficient in measures of human and animal suffering, cost and time to market of the newly discovered compounds. Currently, pharmaceutical companies investigate various methods for increasing their productivity in drug development, in order to compensate for increasing costs and to avoid regulatory fiascos. The major drawback of the traditional system lies in the lack of a priori guidance about the potentially successful treatments. The need for a change in paradigm of clinical trial design has been reiterated, and it has been suggested that the new paradigm should be based on formal methods for predicting disease progression under specific treatment regimens of given drugs or drug combinations. A complex set of mathematical models, denoted Virtual Patient Model, retrieving the dynamics of key biological, pathological and pharmacological processes in the body of a patient undergoing anti cancer drug treatment has been developed. The Virtual Patient Model has been employed for studying improved regimens for cytotoxic and cytostatic mono- and combination drug regimens and for selecting optimal personalized treatments. By simulating in the Virtual Patient Model in a population of patients, one can conduct Virtual Clinical Trials recreating and improving drug development. To this end a collection of Virtual Patients Models is created (denoted Synthetic Human Population). Each Virtual Patient in the population is represented by a set of parameters for the Virtual Patient Model. The inclusive set of parameters represents the distributions of disease, physiological and PK/PD parameters in the population. The Virtual Clinical Trials can be employed in drug development in conjunction with an elaborate algorithm, Interactive Clinical Trial Design (ICTD), which provides a method for a step-by-step process of model prediction and in-vivo verification. The user can employ the ICTD for fine-tuning and testing the drug/disease/population models interactively with the "real" clinical trials, so that relatively early during development the Virtual Patient Model can be employed for checking the most appropriate treatment schedules for the drug, or for making an early "No-Go" decision. The ICTD algorithm is expected to replace the current drug trial-and-error policy by a new policy of clinical trials, which will be based upon a gradual improvement and zeroing-in on the best prediction-directed treatment schedules. © 2012 John Wiley & Sons, Ltd.