Clinical Trials of New Cancer Drugs 'Frequently Flawed'

Pam Harrison

May 02, 2019

A significant proportion of new cancer drugs that were approved by the US Food and Drug Administration (FDA) during the past 5 years were approved on the basis of trials with suboptimal control arms, which likely biased the trials in favor of the experimental arm, a new study suggests.

"I think the problem is that the FDA sets a benchmark for a control arm that is often outdated when some trials are starting," lead author Talal Hilal, MD, Mayo Clinic, Phoenix, Arizona, told Medscape Medical News in an email.

"This becomes an easy target for companies to beat when they test their new drug against an older, outdated control," he added.

"But if we know that the control arm is suboptimal, then it is more likely for patients in that control arm to have disease progression or [any endpoint] event, and that will prove the superiority of the investigational agent faster, " Hilal said. This is "an outcome that any company would want, as their investigational agent [will then] beat the control arm as fast and as easily as possible," he added.

The study was published online May 2 in JAMA Oncology.

The study included 143 anticancer drug approvals by the FDA from January 2013 to July 2018; 47 approvals were excluded, owing to the fact that the approvals were based on single-arm studies.

The control arm of the study was considered to be suboptimal if restrictions were placed on the practitioner's choice of a control drug that excluded a recommended agent, or if the drug that was used as the control was specified but was not the recommended agent, or if the approach that was followed was not normally used in clinical practice.

A control arm was also considered suboptimal if an earlier randomized controlled trial (RCT) showed that the control agent used in the trial was inferior to a readily available alternative.

Analysis of all hematology/oncology drug approvals during the study interval showed that the FDA approved 96 drugs on the basis of results from 98 RCTs.

"Of these 96 anticancer drug approvals, 16 (17%) were based on trials with suboptimal controls," Hilal and colleagues report.

Interestingly, for the 16 drugs approved, only one drug trial was conducted in the United States; the other 15 were international trials, the investigators add.

For the majority (69%) of the trials, progression-free survival (PFS) was the primary endpoint. The overall response rate or other surrogate endpoints served as the primary endpoint in all but one trial. In that trial, overall survival was the primary outcome.

The investigators categorized how the control arm was suboptimal in each of the 16 trials, using their definition of "suboptimal."

In four of the suboptimal trials (25%), the investigators' choice of the active comparator treatment was limited; 63% of the trials used a control known to be inferior to other available agents, or the trials did not allow standard of care (SOC) combinations of agents to be used. In 13% of the suboptimal trials, the control arm had previously been shown to lack benefit associated with reexposure.

"When the threshold for determining the SOC therapy was extended to 2 years prior to start of accrual of an RCT of interest, 14 of the 16 clinical trials had control arms considered suboptimal," the investigators observe.

Hilal explained that determining the design of a clinical trial and how it is to be conducted is a long and arduous process. In some cases, the control arm is selected a year or longer before enrollment begins.

"When we set the date for 1 year prior to enrollment, the number was 16, but when we extended the period to 2 years prior to enrollment, the number was 14 — meaning that in two trials, the control arms were actually deemed SOC if you go back 2 years before the trial opened," Hilal noted.

"This reflects the reality that the benchmark the FDA sets for SOC can be outdated and does not always reflect the actual practice in the community," he emphasized.

Important Observations

"These findings raise important observations," the authors note.

First, for RCTs conducted in other countries, investigators may have no choice but to use a suboptimal control arm, because drugs that are considered SOC in the United States may not be available in other countries at study outset. (Hilal and colleagues point out that international RCT accrual only directly explained their choice for the control arm in three of the 16 trials.)

The investigators also note that most anticancer drugs that are approved by the FDA on the basis of results from RCTs receive regular or full approval.

This means that the FDA does not require any other RCTs to be conducted to confirm the clinical benefit observed in the original registration trial, even if the control arm was suboptimal, the authors point out.

The upshot of this is that when an experimental agent has not been proven superior to SOC treatments, "clinicians are potentially offering patients an agent that may be the equivalent or even inferior to established standard therapy," write the authors.

Added Hilal: "Oncologists need to examine the evidence for themselves when deciding on using a new approved agent and ask themselves, 'Is this drug proven to be more beneficial than the SOC that is commonly used? And if not, is it at least equivalent and less toxic than the SOC?'

"Moreover, if the new drug is more beneficial than SOC, what are the trade-offs for patients in terms of toxicity and cost, and what is the magnitude of benefit?" Hilal emphasized.

The study authors conclude that future regulatory trials should be designed in such a way as to ensure that any new anticancer agent is truly superior to what most clinicians would prescribe outside of a clinical trial setting.

Hilal has disclosed no relevant financial relationships. Coauthor Vimy Prasad, MD, has received grants from the Laura and John Arnold Foundation and royalties from his book, Ending Medical Reversal. He has also dreceived honoraria for grand rounds and lectures from several universities, medical centers, nonprofit groups, and professional societies.

JAMA Oncol. Published online May 2, 2019. Abstract

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.