Trial Investigators Discuss FDA Draft Guidance on Master Protocols

Targeted Therapies in OncologyNovember 2018
Volume 7
Issue 11

Complex drug trials that use a single protocol to test either 1 agent against multiple cancers or multiple agents against 1 cancer can make drug testing faster, cheaper, and more informative, particularly when those trials use the data to constantly improve protocols.

Complex drug trials that use a single protocol to test either 1 agent against multiple cancers or multiple agents against 1 cancer can make drug testing faster, cheaper, and more informative, particularly when those trials use the data to constantly improve protocols.

Those trials are rare, but the FDA hopes to make them more common with new recommendations for designing and conducting these types of studies. The draft documents do not relax existing trial-design standards. Instead, they are intended to guide investigators in planning trials that meet the standards of good trial design as laid out by the FDA. The guidance will be finalized after the agency reviews comments submitted in November.

“A master protocol can create challenges in trial conduct and analysis that, if not properly addressed, can increase risks to patients or delay the development of the drug,” said Patricia Keegan, MD, the director of the FDA’s Division of Oncology Products 2, Office of Hematology and Oncology Products in the Center for Drug Evaluation and Review, in an interview with Targeted Therapies in Oncology.

“The objective for publishing this draft guidance is to call attention to the potential challenges posed by these trial designs, to provide advice to industry and researchers on the information that sponsors should submit to the FDA to support the planned trial design, and [to advise] on how sponsors should interact with the FDA to facilitate efficient review,” Keegan added.

Adaptive trial protocols allow for prospectively planned modifications to 1 or more aspects of the trial based on data gathered from trial subjects from a prespecified analysis in the trial design. Such modifications may be able to shift trial patients away from experimental treatments that appear to be performing poorly, toward treatments that appear to be performing well, or they may be used to reduce the number of patients needed to detect efficacy in a certain subpopulation.1

Master protocol control trials consist of multiple substudies evaluating 1 or more investigational drugs, disease types, or both. Variable terms exist for different variations of the term master protocol, but all can streamline drug testing.2

Basket Trials

Basket trials test a single drug or drug combination against several tumor types. In most cases, a targeted therapy agent accounts for at least 1 part of the experimental regimen, and most trial patients have tumors that express an aberration consistent with the agent being used, regardless of where the tumor appears in the body.

Even though they express the same aberrations, tumors from different locations often respond differently to targeted agents and are said to constitute a different basket (FIGURE 1).3A prominent example of a basket trial is KEYNOTE-028, which is testing the PD-L1 inhibitor pembrolizumab (Keytruda) in more than 450 patients whose tumors express PD-L1.4

Most basket trials have relatively small numbers of patients in each basket. If a drug performs well against a specific tumor type, its manufacturers typically initiate a traditional phase III trial. A number of approvals for pembrolizumab stemmed from trials initiated by Merck because of the drug’s performance in an arm of the KEYNOTE-028 trial.

Umbrella Trials

Trials that test multiple agents against the same tumor type are considered umbrella trials. Investigators test each patient’s tumor for relevant mutations, then assign them to the most appropriate treatment based on the results. In many cases, umbrella trials come late enough in the testing process that patients with a particular mutation will be randomized to receive either an experimental treatment or the existing standard-of-care drug.

Subsequent patients within a given study expressing other mutations or aberrations will similarly be randomized to receive a therapy targeting abnormalities expressed by their tumors or the standard-of-care drug, and so on (FIGURE 2).3

For example, the FOCUS4 trial is enrolling patients with previously untreated metastatic colorectal cancer and providing them with 16 weeks of standard chemotherapy. Biomarkers are then used to assign them to 1 of 4 cohorts, where they are randomized to receive additional treatment with either a targeted agent or placebo.5Other examples of umbrella trials include the lung cancer trials ALCHEMIST and Lung-MAP.6,7

Platform Trials

Umbrella trials that use more sophisticated protocols based upon previous outcomes are often referred to as platform trials. The I-SPY 2 trial is testing neoadjuvant treatments for locally advanced breast cancer and is regarded as one of the pioneering models for platform trials.8

I-SPY investigator Hope S. Rugo, MD, the director of the Breast Oncology Clinical Trials Program at the Helen Diller Family Comprehensive Cancer Center at the University of California, San Francisco, boasted about the advantages of trials that can evaluate several agents all at once. “Even with the simplest of traditional trials, it often takes years of effort to get from initial idea to patient enrollment, mostly because of all the different organizations that need to approve every trial,” she said in an interview. “It’s very slow. It’s very expensive.”

I-SPY 2 uses a standard therapy control arm as well as an automated algorithm that updates randomization probabilities according to biomarker signature, MRI scans, and pathologic complete response results.9Each new patient’s cancer is evaluated and placed into 1 of 10 molecular subtypes before being electronically randomized to a given study arm.

The primary endpoint of pathologic complete response is assessed at the time of surgery. When efficacy is determined, results are then immediately fed back into the algorithm to modify future randomization.

When an experimental agent reaches a predetermined level of efficacy in 1 or more molecular subtypes, it “graduates” to a phase III trial. In 2014, neratinib (Nerlynx) advanced from I-SPY 2 to a phase III study testing its use in HER2-positive, hormone receptor (HR)—negative breast cancer.9 It later went on to receive FDA approval in July 2017 for the treatment of adult patients with early stage, HER2-overexpressed and amplified breast cancer following adjuvant trastuzumab (Herceptin).

Conversely, agents that fail to reach the efficiency threshold after a prespecified number of patients fail out of the trial.

As many as 5 agents can be tested at one time, and new agents can be plugged in when vacancies open up. This system has allowed I-SPY 2 to perform preliminary evaluations on 15 regimens over the past 8 years and “graduate” 6 of those agents on to phase III trials.7

“The more frequently we can combine multiple individual trials into single complex trials, the faster we will make progress against cancer and the further our research dollars will go,” Rugo said.

Understanding the Benefits

Investigators struggle to quantify the temporal and monetary savings that complex protocols can achieve over traditional trials. An analysis of the ongoing STAMPEDE prostate cancer trial, which simultaneously tests 5 experimental regimens against a single control arm, suggests that complex designs can significantly cut trial times and costs.

The original arms of the trial, which began in 2005, included men randomized to either androgen deprivation therapy (ADT) alone, ADT plus zoledronic acid, ADT plus docetaxel, ADT plus docetaxel and zoledronic acid, ADT plus celecoxib, or ADT plus zoledronic acid and celecoxib. Another trial arm added in 2010 tested abiraterone acetate (Zytiga), followed by new arms testing abiraterone plus enzalutamide (Xtandi), radiation therapy, and metformin.

The trial has already changed the standard of care for high-risk prostate cancer and is expected to have produced 8 major results by 2020. Those results are being produced at accelerated speeds as the trial becomes more efficient. STAMPEDE is on pace to produce 1 major result every 18 months from 2015 to 2020. Plans to test further treatments, specifically targeted therapies in those patients who would likely glean the greatest benefit, are in the works. Individual trials running parallel would likely have taken decades longer.10

The key aspects of the FDA guidance for Master Protocols has a number of specific recommendations:2

Protocols should reduce the necessary number of trial patients by using a single control arm, with control patients receiving the existing standard of care.

Protocols that employ combinations of 2 or more agents should provide strong reasons for their selection for the regimen. All data about safety, pharmacology, and efficacy, as well as a rationale for using the combination over a single agent and evidence of any synergy when used in combination, should be presented to the FDA.

Protocols designed to test targeted medications should explain which biomarkers they will use to select patients and develop well-defined criteria for marker positivity before trial initiation.

Protocols should lay out clear policies for ongoing evaluation of experimental treatment arms with specific criteria for declaring those treatments successful or unsuccessful, along with conditions for removing them from the trial. Specific policies for adding new experimental treatments to the trial should also be detailed.

Protocols should use independent data monitoring committees and, when sponsors believe 1 or more of the substudies in the trial could be used for marketing applications, independent radiologic review committees as well.

The draft guidelines also provide lengthier recommendations regarding patient safety, informed consent, trial ethics, institutional review, regulatory approvals, ongoing communication with the FDA, potential difficulties in conducting analyses from complex trials, and strategies for avoiding or overcoming those difficulties.

The draft guidance on adaptive trials is not entirely new; it’s an update of guidelines from 2010 that outline how to create rules for protocol modification.

“There’s a lot of enthusiasm in the research community about moving from traditional trials to these more sophisticated designs, and I think this FDA guidance will speed that transition,” said Ken Anderson, MD, the program director of the Jerome Lipper Multiple Myeloma Center and LeBow Institute for Myeloma Therapeutics at Harvard Medical School in Boston, Massachusetts. “Clinical researchers and industry can now be more confident in their ability to design a trial that will be useful in identifying efficacious drugs much more efficiently, ultimately shortening the time to a new drug approval.”

Anderson, who is also the Kraft Family Professor of Medicine, went on to say that these trials are efficient, saving both time and money. They are also more well adapted to precision medicine than traditional trial designs. “They are geared to treating smaller [subsets of] patients with specific tumor profiles,” he said, adding that this type of research “will more quickly identify treatments that are both effective and well tolerated.”

“Ultimately, they will help both clinical caregivers and patients alike choose medications that provide benefit to specific patients while avoiding ineffective therapies.


  1. US Department of Health and Human Services. Adaptive design for clinical trials of drugs and biologics guidance for industry. Published September 2018. Accessed October 31, 2018.
  2. US Department of Health and Human Services. Master protocols: efficient clinical design strategies to expedite development of oncology drugs and biologics guidance for industry. Published September 2018. Accessed October 31, 2018.
  3. Menis J, Hasan B, Besse B. New clinical research strategies in thoracic oncology: clinical trial design, adaptive, basket and umbrella trials, new end-points and new evaluations of response.Eur Respir Rev.2014;23(133):367-378. doi: 10.1183/09059180.00004214.
  4. New data from KEYNOTE-028, Merck’s trial evaluating KEYTRUDA (pembrolizumab) across a range of cancer types, presented at 2015 European Cancer Congress [news release]. Kenilworth, NJ: Merck & Co, Inc; September27, 2015. Accessed October 31, 2018.
  5. FOCUS4: A molecularly stratified trial programme in colorectal cancer. FOCUS4 Trial website. Accessed October 31, 2018.
  6. The ALCHEMIST Lung Cancer Trials. National Cancer Institute website. Updated July 24, 2017. Accessed October 31, 2018.
  7. Lung-MAP clinical trial. Lung-MAP website. Accessed October 31, 2018.
  8. The I-SPY 2 Trial. I-SPY website. Accessed October 31, 2018.
  9. Goodman A. I-SPY 2 trial ‘graduates’ Neratinib to phase III in HER2-positive, hormone receptor—negative breast cancer. ASCO Post website. Published May 15, 2014.
  10. Parmar MK, Sydes MR, Cafferty FH, et al. Testing many treatments within a single protocol over 10 years at MRC clinical trials unit at UCL: multi-arm, multi-stage platform, umbrella and basket protocols.Clin Trials.2017;14(5):451-461. doi: 10.1177/1740774517725697.
Related Videos
The Oncology Brothers with Joshua K. Sabari, MD, presenting slides
The Oncology Brothers with Joshua K. Sabari, MD, presenting slides
The Oncology Brothers with Joshua K. Sabari, MD, presenting slides
The Oncology Brothers with Joshua K. Sabari, MD, presenting slides
The Oncology Brothers with Joshua K. Sabari, MD, presenting slides
Related Content