One in Five Cancer Trials End Too Early, but GU Studies Not More Likely

The premature termination of clinical trials in the oncology setting is a problem that continues to vex researchers.

Matthew Galsky, MD

The premature termination of clinical trials in the oncology setting is a problem that continues to vex researchers, and results of a new study that involved more than 7500 adult cancer clinical trials suggest that the concern is justified, with as many as 20% of these studies failing to complete for reasons unconnected to either the efficacy of the intervention or side effects.

Clinical trials that terminate early—frequently because they don’t enroll enough patients—contribute little to the scientific knowledge base, waste limited financial resources, divert patients from other trials, and represent an “extreme example of inefficiency in the system,” noted Matthew D. Galsky, MD, associate professor of Medicine at Mount Sinai’s Icahn School of Medicine in New York City. Galsky, who also directs the Genitourinary Medical Oncology Program at the Tisch Cancer Institute, presented the study findings at a presscast held in advance of the 10th Annual 2014 Genitourinary (GU) Cancers Symposium.

Galsky said that the field of bladder cancer is full of examples of trials that terminated prematurely. A 2010 Institute of Medicine report previously found that about 40% of all clinical trials initiated by National Cancer Institute cooperative groups never finished, but the problem had not yet been comprehensively examined within the larger clinical trials universe, providing one of the rationales for the current study.

To better understand the scope of the challenge in general—and its impact on trials involving GU cancers in particular—Galsky and colleagues analyzed data from 7776 phase II-III interventional adult cancer clinical trials registered on between 2005-2011. GU tumor types represented in these trials included prostate (n = 491), kidney (142), bladder (75), and testicular (34).

Investigators used Kaplan-Meier and Cox regression analyses to determine the cumulative incidence of trials failing to complete over time and to identify variables associated with early termination.

Of the sample, 935 trials failed to complete enrollment for reasons unrelated to toxicity or efficacy of the intervention, and of these, 39% were due to poor accrual. Galsky pointed out that approximately 48,000 patients were enrolled in these prematurely ended trials between 2005 and 2012. Other factors related to failure to complete included logistical reasons, such as inadequate funding, cancellation of the trial by the sponsor, or the primary investigator leaving the institution that was conducting the study, Galsky explained.

“Importantly,” he continued, “about 18% of the trials that failed to complete did so for reasons related to the side effects of the experimental intervention or because the intervention lacked effectiveness.” Galsky noted that trials that ended for these two reasons were censored out of the larger analysis, “because these are actually good reasons for a trial to close and do contribute knowledge.”

The researchers also found that GU cancer trials were neither more nor less likely to complete compared with trials for other cancers.

Trials conducted at a single location, trials with industry sponsors, and trials with study locations only inside the United States were more likely to end prematurely, according to the researchers.

“These findings further underscore the clinical trial accrual problem we have in the United States,” said Galsky. “Not only does poor accrual lead to more expensive trials and to trials that generate answers much more slowly, but it also prevents many trials from generating any answers at all.”

Galsky stressed that the discussion is not directed at any particular stakeholder, nor is it a “finger-pointing exercise.”

“Rather, what we wanted to do was hold up a mirror to our activities as a cancer clinical trials community and really ask ourselves whether the system is optimized to bring better treatments to our patients as efficiently as possible. Clearly, there is some work to do.”

He concluded by noting that this analysis and others point to the need for better collaboration and communication within the system, as well as novel approaches to increase cancer clinical trial accrual, which has remained steady at 3%-5% of the adult cancer population for decades.

Stensland KD, McBride R, Wisnivesky JP, et al. Premature termination of genitourinary cancer clinical trials. Presented at: 2014 GU Cancers Symposium; January 30-February 1, 2014; San Francisco, CA. Abstract 288.


View more from the 2014 GU Cancers Symposium