Immuno-Oncology Resource Center

Immuno-Oncology Resource Center

Issue: May 25, 2021
Disclosures: Kim reports consultant roles with Benson and Novartis. Prasad reports funding from Arnold Ventures, LLC, and consultant roles with Century Health, Epicor and UnitedHealthcare. Ross reports research funding from Johnson & Johnson. Zettler reports employment with Regor Pharmaceuticals. Vandross reports no relevant financial disclosures.
May 25, 2021
15 min read

FDA’s accelerated approval program: Reevaluating the need for speed

Issue: May 25, 2021
Disclosures: Kim reports consultant roles with Benson and Novartis. Prasad reports funding from Arnold Ventures, LLC, and consultant roles with Century Health, Epicor and UnitedHealthcare. Ross reports research funding from Johnson & Johnson. Zettler reports employment with Regor Pharmaceuticals. Vandross reports no relevant financial disclosures.
You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact

Established in 1992 in response to the HIV epidemic, the FDA’s accelerated approval program offers a valuable opportunity to make potentially lifesaving therapeutics available to patients in a timely manner.

Accelerated approval is made possible through surrogate endpoints for which results generally are available more rapidly. To qualify as surrogates, these endpoints must be considered “reasonably likely” to predict or correlate with clinical benefit, and subsequent confirmatory trials must verify this benefit, supporting a conversion to full approval. If these confirmatory trials fail to show clinical benefit or are not conducted in a timely or proper manner, the marketing authorization for specific indications may be withdrawn.

The FDA’s reevaluation of the accelerated approval program is a positive for patients and providers, according to Marjorie Zettler, PhD, MPH.
The FDA’s reevaluation of the accelerated approval program is a positive for patients and providers, according to Marjorie Zettler, PhD, MPH. “Most oncologists in practice are not … tracking every drug to ensure that these confirmatory trials get done and show a clinical benefit,” she said. “They’re trusting that FDA is taking care of this.”
Source: Regor Pharmaceuticals.

Many of the surrogate endpoints upon which accelerated approvals are based have been very effective in predicting benefit and have made lifesaving drugs accessible to patients sooner.

However, some of the most common treatments to receive accelerated approval — anticancer agents — have failed to show clinical benefit in confirmatory trials. This has led to the voluntary withdrawal of oncology drug indications by AstraZeneca, Bristol Myers Squibb, Genentech and Merck.

The FDA responded with an industrywide reevaluation of the accelerated approval program that seeks to ensure the time saved in providing access to these cancer drugs is aligned with the kind of time patients truly value — quality time being alive. During a 3-day meeting in April, the Oncologic Drugs Advisory Committee (ODAC) voted on whether to maintain six accelerated approval indications for PD-1 and PD-L1 checkpoint inhibitors and discussed the possibility of a continued reevaluation of the accelerated approval process.

“I really welcome this reevaluation of the evidence for accelerated approval based on surrogate endpoints,” Joseph S. Ross, MD, MHS, professor of medicine and public health and member of the Center for Outcomes Research and Evaluation at Yale New Haven Hospital, told HemOnc Today. “A number of people, including myself, have been raising concerns about this over the years. I was pleasantly surprised to see attention paid to this issue, because my general sense is that this is a great way for the FDA to protect public health, promote innovation and do right by the broader U.S. community.”

Joseph S. Ross, MD, MHS
Joseph Ross

HemOnc Today spoke with experts about the utility of using surrogate endpoints to accelerate access to oncology drugs, the industrywide reckoning that has occurred as a result of confirmatory trials failing to show benefit, the results of the ODAC meeting, and what this broader discussion may mean for the future of the accelerated approval program.

Varied value of surrogates

Accelerated approval, and specifically the use of surrogate endpoints, proved to be crucial when first used in the setting of HIV treatment, according to Ross.

“The best example of a surrogate marker being used to support approval was actually for HIV and AIDS,” Ross said. “Surrogate markers of CD4 counts and HIV viral load work very well for predicting outcomes, and that was how we began to think about redesigning the trials to allow for accelerated evaluation programs.”

Although surrogate endpoints like PFS and response rate are valuable metrics in terms of expediting approval of potentially lifesaving drugs, they do not necessarily translate to longer OS or better quality of life, Ross said.

“In the oncology space, there’s been a lot of research showing that these surrogates do not always correlate reliably with observed improvements in survival,” he said. “When there’s a product for a disease as serious and difficult to treat as cancer, you want patients to have access to it while testing continues to confirm that benefit, but they have to hold companies responsible for doing the clinical trials that confirm the benefit.”

The value of each surrogate also varies based on tumor type, according to Vinay Prasad, MD, MPH, a hematologist-oncologist and associate professor in the department of epidemiology and biostatistics at University of California, San Francisco, and a HemOnc Today Next Gen Innovator.

Vinay Prasad, MD, MPH
Vinay Prasad

“Every surrogate endpoint in cancer is unique; the question always is, ‘Does this surrogate predict longer life with this cancer?’” Prasad told HemOnc Today. “There have been surrogates that are effective at predicting benefit. There are many, many surrogates that are mediocre or worse than mediocre; they have almost no correlation between living longer and living better.”

Prasad and Chul Kim, MD, MPH, assistant professor in the department of hematology and oncology at Georgetown University School of Medicine, co-authored a study published in JAMA Internal Medicine that evaluated 54 drugs approved between 2008 and 2012, including 36 approved based on a surrogate endpoint. The most common surrogate used was response rate (53%) followed by PFS or DFS (47%).

After median follow-up of 4.4 years, only five of those drugs were subsequently shown to extend OS. Overall, 86% of the drugs approved based on a surrogate had unknown impact on OS (n = 13) or had failed to extend OS in confirmatory trials (n = 18).

In some cases, confirmatory studies do not properly evaluate whether the previously used surrogate endpoints predict clinical benefit.

“It’s important to run the post-marketing confirmation trials to know exactly whether the initial system of basing benefit on surrogate endpoints was legit or not,” Kim told HemOnc Today. “There have also been some issues in post-marketing trials, because sometimes they use the same surrogate endpoint for the follow-up study. What’s the point?”

Prasad said the FDA has embraced surrogates across the board, calling the process “a product of their enthusiasm to approve drugs.”

“The entire ecosystem of cancer is in crisis,” he said. “The crisis is that everybody involved in the process is driven to get more drugs, to give them early and often. Everybody makes more money doing that. I don’t know if, on balance, accelerated approval is better for patients. I know it’s better for profits.”

Speed and thoroughness

Post-marketing trials may not only fail to confirm benefit but, in some cases, they are delayed or never conducted at all. This noncompliance with confirmatory studies is potentially misleading to clinicians and patients who equate approval with efficacy.

“In recent years, there has been some growing concern among the scientific and medical communities that some manufacturers are not following through — they’re not doing the studies,” Marjorie Zettler, PhD, MPH, director of clinical science at Regor Pharmaceuticals, told HemOnc Today. “So, the drugs with unconfirmed benefit are staying on the market.

“The second concern is that when studies are done and they fail, the manufacturers are not moving to voluntarily withdraw the drug from the market, or the FDA is not requiring it,” she added. “So, the drugs that we know don’t provide a clinical benefit to patients are remaining on the market.”

Zettler co-authored a 2018 study of 49 novel oncology drugs that were approved between 2011 and 2016. Results, published in JAMA Oncology, showed 23 (47%) of the agents had received accelerated approval, 17 (74%) of which had post-marketing requirements to complete, including 34 clinical trials.

Of the 34 clinical trials that were required, 15 (44%) were completed, 14 (41%) were ongoing at the time of the study, two (6%) were terminated — both evaluated idelalisib (Zydelig, Gilead) and were halted due to safety concerns — and three (9%) were pending. Of the 15 completed trials, three (20%) failed — these trials sought to confirm benefit for the checkpoint inhibitors atezolizumab (Tecentriq, Genentech/Roche), nivolumab (Opdivo, Bristol Myers Squibb) and pembrolizumab (Keytruda Merck). But, at the time of the study, those drugs remained on the market for their indications.

One drug, ponatinib (Iclusig, Millennium/Takeda) was temporarily removed from the market contingent on further studies and a risk evaluation and mitigation plan. However, the post-marketing clinical study eventually was completed.

The researchers concluded that the post-marketing requirement studies were critical in identifying the safety concerns of idelalisib and ponatinib and that all other pending and ongoing studies were adhering to their original schedule as posted in the FDA’s post-marketing requirements database.

“Most oncologists in practice are not looking at these drug approvals with a magnifying glass, and they’re not tracking every drug to ensure that these confirmatory trials get done and show a clinical benefit,” Zettler said. “They’re too busy providing care to their patients, and they’re trusting that FDA is taking care of this. That trust is understandable and reasonable. So, this evaluation the FDA is conducting is a positive for both the oncologists and their patients.”

Given the risk for erroneous approval of an ineffective cancer drug, the “acceleration” is not necessarily much faster, Prasad said.

In a study of the duration of registration trials of oncology drugs approved between 2006 and 2017, published in JAMA Internal Medicine, Prasad and colleagues reported that the median duration of studies that used PFS as a primary endpoint did not differ from those that used OS (31 months vs. 33 months), although median duration for trials that used response rate was significantly shorter (25 months; P = .001). Multivariate analysis showed study durations were shorter by a mean 11 months (95% CI, 5-17) with PFS as a primary endpoint and 19 months (95% CI, 13-25) with response rate as a primary endpoint compared with studies that assessed OS.

“In terms of speed, we found that over an 8-year period, the use of surrogates generally speeds drug approval by an average of 11 months,” Prasad said. “Some of us think it would be better to wait those 11 months if you’ve already waited 8 years.”

An ongoing reassessment

Still, the program has led to more than 150 accelerated approvals, with almost half of these converted to regular approval within a median 3 years, according to data from the FDA presented during last month’s ODAC meeting. Only 10 of these agents have been withdrawn from the market.

The ODAC meeting focused on checkpoint inhibitors — out of 76 approved indications for checkpoint inhibitors, 35 were accelerated.

“Initial clinical development and subsequent accelerated approval of the anti-PD-1 and anti-PD-L1 antibodies were aimed at indications for which there were high levels of unmet medical need,” Julia A. Beaver, MD, acting deputy director in the Office of Oncologic Diseases at FDA, and Richard Pazdur, MD, director of FDA’s Oncology Center of Excellence, wrote in a perspective published ahead of the meeting in The New England Journal of Medicine. “The period of rapid, dynamic drug development provided evolving information, including optimal biomarker-driven patient selection, doses and schedules, adverse-event management and endpoint selection for clinical trials. The fact that some of this information was unavailable at the time of many of the approvals must be balanced against the benefits offered to thousands of patients with cancer whose lives have been greatly enhanced by early access to these antibodies.”

During her introductory remarks at the meeting, Beaver noted that when a drug does not meet its clinical trial endpoint or fails to demonstrate a meaningful outcome, it does not necessarily follow that the drug will be ineffective.

“This failure to demonstrate meaningful efficacy, rather than indicating a true lack of efficacy, can potentially be explained by differences in trial design, differences in endpoints, statistical testing or biomarker selection,” Beaver said. “When clear reasons exist for a trial not to achieve its primary endpoint, or to demonstrate a small benefit that is not meaningful and an unmet clinical need still exists, the FDA will work with companies to identify subsequent clinical trials to verify that benefit while retaining the original accelerated approval.”

According to Beaver, the FDA’s reevaluation of its accelerated approval program is appropriate and will likely continue as more oncology drugs are developed.

“To increase transparency in the future, we may continue public discussions of these evaluations on a more periodic basis,” she said.

An FDA review of the 35 accelerated approvals of checkpoint inhibitors showed 10 were “dangling,” or still on the market despite a lack of confirmatory benefit. Four of these were subsequently withdrawn, and the remaining six were the subject of the ODAC meeting.

“While these antibodies have definite disease activity in specific patients, the results of confirmatory studies suggest that the risk-benefit calculation for these indications may have changed in the contemporary treatment landscape, and this may warrant further investigation,” Beaver said.

Difficult decisions

ODAC ultimately voted to maintain four of the six indications reviewed for atezolizumab, pembrolizumab and nivolumab (see Table).

ODAC Voting Results 

The panel voted to maintain atezolizumab’s indication for use with paclitaxel protein-bound (Abraxane, Bristol Myers Squibb) for a subset of patients with triple-negative breast cancer, despite lack of benefit in the confirmatory IMpassion131 trial. In a near-unanimous vote, ODAC also recommended maintaining an indication for atezolizumab for the treatment of locally advanced or metastatic urothelial carcinoma pending OS results of IMvigor130.

The panel discussed several indications for pembrolizumab, including for locally advanced or metastatic urothelial carcinoma; third-line treatment of recurrent, locally advanced or metastatic, PD-L1-positive gastric or gastroesophageal junction adenocarcinoma; and hepatocellular carcinoma previously treated with sorafenib (Nexavar, Bayer). The committee opted to maintain all but the gastric cancer indication.

Pembrolizumab received that accelerated approval in 2017 based on results of the KEYNOTE-059 trial. Although this trial showed an ORR of 13.3%, the confirmatory second-line KEYNOTE-061 trial and first-line KEYNOTE-062 trial failed to show OS and PFS benefits.

“While we are disappointed in the outcome of the gastric cancer vote, especially given the discussion of the unmet need that remains for immunotherapy-naive patients who have progressed to needing third-line treatment, we are committed to continuing discussions with the FDA to help ensure access to Keytruda for these patients,” Roy Baynes, MD, PhD, senior vice president, head of global clinical development and chief medical officer at Merck Research Laboratories, said in a company-issued press release.

In addition to citing the lack of confirmatory evidence, panel members who voted against maintaining the indication noted the changing therapeutic landscape, as nivolumab plus chemotherapy has since been approved in the first-line setting.

“The landscape of therapeutics and even companion diagnostics looks so different now than when accelerated approval was first granted,” Mark A. Lewis, MD, director of gastrointestinal medical oncology at Intermountain Healthcare, said after the vote. “This meeting is providential, coming after the seismic CheckMate-649 data and subsequent approval. I think we are going to see therapy earlier in treatment lines. The reason I voted ‘no’ is not to turn a deaf ear on the needs of patients — it’s that none of the pending studies, as best as I could tell, would definitively answer the questions of [benefit of] monotherapy or of the [appropriate] therapeutic sequencing for this agent.”

The decision to maintain pembrolizumab’s HCC indication was made despite that the confirmatory KEYNOTE-240 trial failed to reach its prespecified requirement for statistical significance of OS vs. placebo. The panel’s vote reflected concerns over the treatment need for patients ineligible for bevacizumab (Avastin, Genentech), approved as part of a combination with atezolizumab.

Yet, the panel voted not to maintain approval of nivolumab monotherapy for this population in part due to the low 14% response rate in the CheckMate-040 trial. Additionally, panel members cited the negative confirmatory CheckMate-459 trial of nivolumab monotherapy vs. sorafenib in the first-line setting.

“I voted ‘yes’ ... [for pembrolizumab] while today I voted ‘no’ [for nivolumab],” Anthony D. Sung, MD, assistant professor of medicine at Duke University School of Medicine and member of Duke Cancer Institute, said after the vote. “I would say the reason I voted ‘yes’ was that there was a potential biomarker signal that could help distinguish who would respond. I do think there are patients who benefit from nivolumab monotherapy. I just think the data show that the benefit isn’t there for the overall population, and I would encourage Bristol Myers Squibb and other members of the pharmaceutical industry to dig a little bit deeper to find biomarkers that could identify which patients may benefit.”

In a press release issued by Bristol Myers Squibb, Ian M. Waxman, MD, the company’s development lead of gastrointestinal cancers, acknowledged the changing HCC treatment landscape.

“Immunotherapy is an important next treatment option for patients who have progressed on sorafenib or were unable to tolerate it,” Waxman said in the release. “We are disappointed with [this] outcome for patients, and we will work closely with the FDA as it completes its review. Immunotherapies have changed the way we treat many forms of advanced cancer. We’re proud of the role Opdivo has played in helping to evolve the HCC treatment landscape, and we remain committed to delivering innovative therapies for [patients with HCC] in need.”

Responding to requests for comment prior to the ODAC meeting, Genentech and Bristol Myers Squibb referred HemOnc Today to statements on their websites.

Merck did not reply to HemOnc Today’s inquiry. However, in a press release issued after the meeting, the company applauded FDA for the accelerated approval program and for its diligence in reviewing confirmatory data.

“The FDA’s successful application of the accelerated approval program has been critically important to provide people with cancer earlier access to potential treatment options,” Baynes said in the release.

Maintaining the standard

Although clinicians with whom HemOnc Today spoke applauded the FDA’s reevaluation of the accelerated approval program, not all were convinced that the outcomes of the ODAC meeting exemplified the agency’s commitment to this reassessment.

“The proviso is that full approval necessitates confirmation with endpoints that show these patients benefit by living longer and the therapy does not have significant toxicity,” Andrae L. Vandross, MD, medical oncologist and phase 1 clinical investigator at NEXT Oncology in Austin, Texas, said in an interview. “Based on the overall result of this meeting, it’s not clear that the overall goal has been met. Since the confirmatory trials did not meet those important endpoints, I thought the proverbial writing was on the wall — which, I imagine, is why there were voluntary withdrawals ahead of this meeting.”

Vandross said he is not sure how to view the accelerated approval program after the outcomes of the meeting.

“It will be challenging to believe in a process that appears to deviate from its own set standard,” he said.

Still, as Beaver noted, the reassessment of the accelerated approval program is unlikely to end with the meeting.

Additionally, the Institute for Clinical and Economic Review (ICER) published a white paper outlining potential policy strategies for improving the program. Authors of the paper explored such options as strengthening the selection of surrogate endpoints; developing standardized templates to improve consistency and evidence standards; requiring increased use of randomized controlled trials; providing economic incentives for the completion of confirmatory trials; changing the conditions for ongoing marketing approval when confirmatory evidence is not furnished by a predetermined date; and offering “safety-only” approval, without the need for public or private insurance coverage, for treatments that have demonstrated safety but not the ability to improve the lives of patients.

“Since 1992, the FDA’s accelerated approval program has brought important new treatments to patients faster than would have been possible through the traditional approval process,” Steven D. Pearson, MD, MSc, president of ICER, said in a press release. “The successes are many, and the benefits for patients overall have been substantial.

“But criticisms have emerged, as well, including concerns that the FDA has lacked consistency in its application of evidentiary standards, and that the incentives and procedures in place today have proven inadequate to get high-quality confirmatory trials completed in a reasonable time frame,” Pearson added. “From various points on the spectrum, concerns are also raised that patients may not be getting fair access to these drugs, while some question whether the high prices for most of the drugs approved through this program are merited, given the level of uncertainty about their effectiveness at launch.”

Zettler agreed the outcome of the ODAC voting was not what she had expected.

“From my perspective, it’s difficult to justify keeping these drugs on the market while waiting for additional confirmatory clinical trials to be completed,” she said. “However, FDA is not bound to take ODAC’s recommendations, so no decision has been made yet.”

In response to an inquiry from HemOnc Today, the FDA issued a statement saying the independent expert advice offered by ODAC and other committees helps the FDA make complex regulatory decisions regarding drugs approved under the accelerated approval pathway.

“FDA considers this independent advice very seriously, while simultaneously weighing the severity of the disease, the [availability] of other treatment options and the benefit-risk algorithm of a particular drug,” the statement said. “The accelerated approval pathway provides an expedient avenue to enable drugs for serious, potentially fatal diseases to reach the market and ultimately the patient’s bedside. However, determining efficacy of these drugs, even in the post-market setting, can be extremely difficult.”

Vandross said he will feel compelled to clarify to his patients that these FDA approvals do not necessarily ensure that a drug will help them.

“It will indeed be oncologists’ responsibility to discuss the results of the trials, and also of this meeting, in order to provide context for the approvals,” he said. “As our patients continue to increase their medical savvy, we will need to be prepared to discuss and clarify that just because there was an approval does not mean that the optimal standard was met.


Beaver JA and Pazdur R. N Engl J Med. 2021;doi:10.1056/NEJMp2104846.

Chen EY, et al. JAMA Intern Med. 2019;doi:10.1001/jamainternmed.2018.8351.

FDA. Accelerated approval. Available at: Accessed on May 5, 2021.

Kaltenboeck A, et al. ICER. Strengthening the accelerated approval pathway: An analysis of potential policy reforms and their impact on uncertainty, access, innovation and costs. Available at: Accessed May 5, 2021.

Kim C and Prasad V. JAMA Intern Med. 2015;doi:10.1001/jamainternmed.2015.5868.

Zettler M and Nabhan C. JAMA Oncol. 2018;doi:10.1001/jamaoncol.2018.0610.

For more information:

Chul Kim, MD, MPH, can be reached at MedStar Georgetown University Hospital, 3800 Reservoir Road NW, Room 412, Washington, D.C. 20007; email:

Vinay Prasad, MD, MPH, can be reached at University of California, San Francisco, 550 16th St., 2nd Floor, San Francisco, CA 94158; Twitter: @vprasadmdmph.

Joseph S. Ross, MD, MHS, can be reached at Yale School of Medicine, 367 Cedar St., Suite 405 B, New Haven, CT 06510; email:

Andrae L. Vandross, MD, can be reached at NEXT Oncology 12221 Renfert Way, Austin, Texas 78758; email:

Marjorie Zettler, PhD, MPH, can be reached at Regor Pharmaceuticals, 50 Soldiers Field Place, Boston, MA 02135; email: