The Phase III Trial in the Era of Targeted Therapy: Unraveling the "Go or No Go" Decision

Reviewer: Ryan P. Smith, MD
The Abramson Cancer Center of the University of Pennsylvania

Author: Roberts Jr TG, Lynch Jr TJ, Chabner BA
Source: Journal of Clinical Oncology, 21(19), 2003, pg 3683-3695


  • In 1954, the first randomized trial in medical oncology was completed
  • Since then, many phase III randomized clinical trials have been completed. Many have been sufficiently powered to establish clinical benefit of incrementally improved advances
  • Randomized testing of all potentially useful novel therapies is not feasible because of cost and time constraints. In fact, despite the billions of dollars invested, only two to five oncology-related drugs are approved by the FDA each year
  • Therefore, there needs to be a consensus on how to prioritize novel targeted therapies for phase III testing
  • This is especially true for molecularly targeted therapies, whose effects are neither cytotoxic nor apoptotic, and therefore produce modest changes in tumor size that are difficult to detect by observing response rates alone
  • This report sought to provide a systematic overview of the characteristics of current phase III cancer treatment trials and offer a framework to assist in prioritizing agents for phase III testing

Materials and Methods

  • Two national databases, the Physician Data Query and the ClinicalTrials.gov database, were reviewed to obtain characteristics of ongoing phase III oncology trials
  • Data obtained included the studies' primary sponsor, treatment modality, disease setting, line of treatment, accrual goal, and planned accrual time
  • These were screened in particular for the study of molecularly targeted therapies


  • 409 active phase III trials were identified, 90% of which had accrual goals available for study and 70% of which had planned accrual periods available
  • Breast cancer (13%) and non small cell lung cancer (8%) were the two most well-represented diseases
  • Trial sponsorship strongly influenced planned accrual goals as the average planned accrual time of the industry-sponsored trials was 2.2 years, compared with 4.2 years for the NCI sponsored trials
  • The median accrual goal was 481 patients and was not different by sponsorship, though those focusing on molecularly targeted therapies had significantly smaller accrual goals (and shorter planned accrual periods)

Factors Stimulating Demand for Phase III Trials

There are an unprecedented number (1345) of antineoplastic drugs in current development. This is compared to the 85 drugs currently approved. Though there has been a mechanism for accelerated marketing approval of cancer drugs on the basis of surrogate endpoints, such as response rates in phase II trials, the FDA will continue to require randomized trials of novel agents in their granting of marketing approval.The reasons for this is first, accelerated approval is only available for an indication where the new therapy offers promise over the existing standard. Also, there is fear that accelerated approval will create incentive to perform large, single arm studies in refractory patient populations, in hopes of approval. Also, there is recognition that many drugs will have modest activity in phase II trials, needing phase III trials to demonstrate a superiority

Factors Limiting the Ability to Perform Phase III Trials

Three factors limit the performance of phase III trials: cost, developmental time, and risk. The cost of developing a new antineoplastic drug is estimated to be $802 million. Of this, the direct clinical expenditures cost is $175 million, of which phase III trials represent the majority (65% to 75%) of the cost. They also extend the time to marketing approval. Also, there is a high probability that the drug will fail to obtain approval. In the past 20 years, only 10% of all agents brought into clinical trials ultimately received FDA approval.

Critical Factors for Entering a Phase III Trial: A Conceptual Model

Using data collected from the review of the databases, the authors created a model for deciding whether or not to enter an agent into a phase III trial. The target validity is the first to consider. There is a risk in initiating phase III trials of targeted agents where the target has not been adequately validated. The agent also must have favorable pharmacokinetic properties, such as a high bioavailability, a long half-life, and shallow trough levels. The activity and toxicity in phase I and II trials should create a favorable therapeutic ratio. If this does not exist, it is unlikely that the drug will be at all clinically useful. Also in considering a drug for phase III trials, the disease targeted for treatment and the line of treatment should be considered. First line treatment is optimal, but it should be remembered that in going for this approval, randomization will be against the treatment with the most activity. Lastly, the potential for a market niche needs to be considered. In other words, a phase III trial is less likely to be initiated for a drug to treat an uncommon disease.

Authors' Conclusions

  • Phase III trials are by far the most expensive and time-consuming portion of the clinical testing of a drug
  • This study provides a systematic overview of characteristics of contemporary phase III trials and formalizes major criteria that sponsors use to make the go or no go decision.
  • Molecularly targeted therapies (or vaccine therapies) seldom identified patient populations that were likely to respond to the agents being studied by the end of phase II, therefore requiring phase III trials. This should be improved upon to save time, money , and improve on efficacy of research

Scientific Implications

This report shows a sobering account of the expense and time required to gain approval for a novel therapy. It is because of these factors that novel drug prices, molecularly targeted therapies, and cancer vaccines are so expensive when they do gain approval. This also shows why there is relatively little research done on uncommon cancers. There is simply no market to cover the research deficit after approval is obtained. Though for safety reasons, much of this is needed, this article points out some areas where industry and governmental sponsors can improve. As the authors illustrate, there should be a stricter evaluation of patients and populations that may benefit from a novel therapy (be it cytotoxic agent, molecular agent, or vaccine), prior to initiating phase III trials. This can decrease time and cost, and increase efficacy, by treating a more homogeneous population.