Failures of Phase III programs after successful Phase II programs is probably the worst outcome of a clinical development program, as it failed in the most costly way. Nevertheless, these failures occur not infrequently. In psychiatry, highly publicized Phase II success stories ended in discontinuations of development programs, such as the NK1-antagonist program in depression several years ago. More recently, other examples have emerged. Some skip the Phase II process altogether with designs, which are supposed to provide “pivotal” data for regulatory purposes in large Phase III-like studies, which are just labeled as Phase II. These failures do not come out of the blue. Sometimes it is important to go back to basics and consider the purpose of Phase II trials.
What is the purpose of a Phase II trial?
The purpose of Phase II trials, besides gaining insights into the safety of a compound, is broadly exploratory, i.e. to generate data, which help with the design of the pivotal Phase III program. In a therapeutic area, a reasonably performed Phase II study can provide insights into clinical and biological patient characteristics, which match the properties of the drug under study. With an increased interest in personalized medicine, these boundaries between patient populations have to be understood in order to be successful. This approach is in direct contradiction to the urge to generate a “pivotal” Phase II outcome.
How are these missed objective trials performed?
By enrolling a large number of very well controlled, homogeneous patient populations into the trial, which reduces variability, these trials have a chance to have a “positive” outcome. However, they missed their main objective, which is to understand the strength and limits of a compound in a heterogeneous population. What is lost in these trials is the required variability, which is necessary in order to define the most suitable population. The rational way for the development path is to have wider spectrum of patients in Phase II without any expectation of an overall significant differentiation in a rather heterogeneous group. The goal is to define where the signal is and to design the Phase III program around this signal. Related to the misunderstanding of the purpose, Phase II is the disappointment in a lack of significant effect in these more exploratory trials, which may lead to the discontinuation of a development program.
However, we must control for some of the variability in these trials. To just compare one variable (active) group with another variable (placebo or active comparator) group would leave too much room for speculation. Therefore, we argue for an approach in which patients are stratified according to mechanism-related characteristics. These can be clinical features, and from my perspective preferable, specific biomarkers. Two main biomarker approaches are relevant for neuroscience: 1. Predictive markers, as determined at baseline; 2. Early markers of target engagement. Besides potential predictive genetic markers, which generally have only a low impact on the clinical change with a given treatment, functional markers work for both purposes. With the advent of easy-to-handle instruments to measure for example EEG-related markers, like evoked potentials and sleep-EEG markers; ECG related markers, like heart rate variability and neuroendocrine markers from saliva new opportunities emerged. A recent ISCTM Symposium highlighted several examples from the neuroscience field.
For advice on the importance of Phase II design or to speak with a Covance representative, please contact us.