Controversies in the Design of Phase II Clinical Trials

Lesley Seymour, MD, PhD, Co-Director of the Investigational New Drug Program Director of the Audit and Monitoring Group of the NCIC Clinical Trials Group, Kingston, Ontario, Canada

H&O What are the challenges of current phase II trials, and why is there a need to update the traditional goals of phase II testing?

LS The major limitation of traditional phase II trial designs is that they were developed for drugs that result in rapid tumor shrinkage (objective response). Drugs that had high rates of response, higher than that of known active agents, were usually proven to be effective (ie, prolong survival) in later phase III studies, whereas agents with low response rates were considered inactive. Newer molecularly targeted agents, although they can result in survival benefit for patients, often have relatively low response rates. Teasing out what is an interesting (but low) response rate, and what response rate should stop further clinical development, is thus difficult. Larger studies, especially randomized ones, allow more robust decisions to be made, but are more costly in terms of patients required, resources, and time.

Numerous phase II clinical trials are testing combinations of new drugs with standard treatments. In this type of situation, it is often difficult to determine whether there is any incremental benefit of adding a new drug, especially when it is not expected to substantially increase the overall response rate. In part because of this lack of clarity, many phase III studies are proving to be negative.

H&O What should be the ultimate goals of phase II trials?

LS The goal of phase II trials is to identify promising agents for further study. We need more efficient trial designs, especially when we have a plethora of new drugs that require testing. Investigators need to carefully consider the most appropriate endpoint for the trial, as well as the use of randomization, while also ensuring efficiency. Adaptive statistical designs should be considered in order to achieve this.

Another goal is to figure out a way to run a robust, well-powered, randomized study that will result in accurate data but with minimum cost and resource usage. Currently, this is where much of the discussions of adaptive designs and novel methodologies are emanating from. As a result, investigators are torn between developing a drug that does not work and patients’ unmet need for new effective drugs.

H&O What are some alternate endpoints that are being explored?

LS Endpoints such as looking at response as a continuous variable, progression-free survival, patient-reported outcomes (PROs), imaging endpoints, and biomarker-based endpoints are of interest and are actively being explored and validated. PROs are relevant secondary endpoints, particularly in diseases such as prostate or pancreatic cancer, especially with larger and/or randomized trials. Prognostic/predictive markers are also being investigated in phase II trials. The conundrum is whether accrual should be limited to a certain subset of patients defined by a biomarker. Although there are some examples where the biomarker is clearly validated, and the inclusion of a biomarker-selected group is appropriate, this is usually not the case. The inclusion of unselected patients in phase II trials is then preferable, but it is recommended that the study be adequately powered to answer the question in the cohort of patients with the putative biomarker.

H&O What are the downsides to including imaging and biomarker-based endpoints and PROs in phase II trials?

LS Cost, as well as increased resources and potential delays in accrual and conduct of the trial, are always of concern when trials become more complex. Ensuring that the design and endpoints are carefully chosen and justified, and that all endpoints are validated and qualified before their inclusion, minimizes these risks.

H&O What is the Clinical Trial Design Task Force and what is its role?

LS The Clinical Trial Design Taskforce of the Investigational Drug Steering Committee (IDSC) of the National Cancer Institute Cancer Therapy and Evaluation Program advises the IDSC on trial design of early clinical trials. As part of this mandate, 2 workshops and resulting recommendations have been coordinated, one addressing phase I design and the other phase II design. The consensus recommendations are presented in Tables 1–3.

H&O Some argue that phase II trials should be eliminated. What is your stance?

LS Our goal should be to get the best drug in class licensed and available to patients, and for this it is necessary to have some kind of screening mechanism given the large number of new agents. Phase II trials are screening trials to identify promising drugs and to weed out ineffective therapies. It is not feasible or reasonable to conduct phase III trials for all new agents. However, timelines can be optimized by using adaptive designs or seamless phase I/II or phase II/III studies. Although some investigators have suggested that phase II studies are of limited value and serve only to delay definitive trials, I believe they have an important role in early drug development, performing a screening function and allowing better definition of toxicity, dose, and schedule.

Controversies in the Design of Phase II Clinical Trials

Suggested Readings