Accelerated drug approval has not come with accelerated evidence

When it comes to serious conditions like cancer, bringing new therapies to market quickly is a priority. That’s exactly what the Food and Drug Administration’s “Accelerated Approval Program” is designed to do. Accelerated approval enables “earlier approval of drugs that treat serious conditions, and that fill an unmet medical need based on a surrogate endpoint.” Rather than prove a drug has a beneficial clinical outcome, such as improved survival or better quality of life, the drugmaker just has to prove that the drug improved a surrogate endpoint – a result that correlates with clinical benefit. After the drug is approved on a conditional basis, drug manufacturers are required to conduct a confirmatory trial showing that the drug actually provides a clinical benefit. 

While the accelerated approval program has led to a greater number of cancer drugs being approved in a short period of time, some oncologists and researchers argue that the process has merely created more profit for drug companies, without corresponding clinical benefit for cancer patients. For example, using surrogate endpoints shortens the time it takes to conduct a clinical trial and bring a drug to market, but these endpoints often fall short of predicting real clinical benefit.

Now, two studies in JAMA Internal Medicine provide more evidence to suggest that the FDA isn’t holding newly approved cancer drugs to high enough standards. Dr. Bishal Gyawali, Dr. Spencer Phillips Hey, and Dr. Aaron Kesselheim of the Program on Regulation, Therapeutics, and Law (PORTAL) at Harvard Medical School/Brigham and Women’s Hospital, analyzed the results of confirmatory trials for all cancer drugs granted accelerated approval from 1992 to 2017. 

If the accelerated approval program is working as intended, these confirmatory trials should show that approved drugs indeed offer a clinical benefit. However, that’s not what Gyawali et al.’s research found. Out of 93 cancer drug indications granted accelerated approval, only 19 (20%) had confirmatory trials that demonstrated a benefit in overall survival. For another 41% of approved indications, confirmatory trials showed improvement using a surrogate outcome, but not in overall survival. The rest of the approvals had not been confirmed yet because trials were delayed or in process. It appears that confirmatory trials are not working as a safeguard against drug manufacturers are using less rigorous evidence to get cancer drugs to market faster.  

Another study by Dr. Emerson Chen, Dr. Vikram Raghunathan, and Dr. Vinay Prasad at Oregon Health and Science University examines the use of response rate as a surrogate endpoint in cancer drug trials. Many cancer drugs are approved through the accelerated pathway based on response rate, which the authors define as “the percentage of patients whose tumors shrink beyond an arbitrary threshold.” Response rate is a surrogate endpoint; it correlates with overall survival, but does not have an inherent benefit to patients (hence “arbitrary threshold”). If drugs are being approved based on response rate, one would expect that the response rate would have to be pretty outstanding, especially if there is no control group in the trial. However, that’s not always the case, as Chen and colleagues found out.

Out of 85 cancer drug indicates that were approved from 2006 to 2018 based on response rate, the median response rate was 41%, meaning that, in the typical trial, fewer than half of patients saw their tumors shrink. One third of approved cancer drugs showed response rates less than 30%. Out of drugs that were approved through accelerated approval but had yet to conduct a confirmatory trial, the typical response rate was just 28%. 

What do these findings mean for patients? The lack of evidence of clinical benefit for new cancer drugs is frustrating for patients and families trying to make treatment decisions. Patient advocate Sally Schott wrote in the American Journal of Medicine that caring for her brother with advanced cancer was made more frustrating by the lack of controlled data for the treatments they gave him. “For my brother and me, we wanted more information, not just more choices,” she writes.

What we need is more randomized controlled trials of new drugs, that measure outcomes that patients care about, not just surrogate endpoints. Unfortunately, as long as the FDA continues to reward drug companies with accelerated approval for surrogate outcomes and low response rates, it is unlikely that manufacturers will conduct RCTs, which take longer and are more expensive.