Regulatory Endpoints for Approval: ORR, OS, or PFS?

Alexander M. Castellino, PhD

July 06, 2016

In a move to get cancer drugs to patients more quickly, some recent accelerated drug approvals have been based on overall response rates (ORRs) rather than the more traditional oncology endpoints of overall survival (OS) or progression-free survival (PFS).

For this news feature, Medscape Medical News approached several oncology experts to ask what they think about this development.

A major move to faster drug approvals was made in 2013, when the US Food and Drug Administration (FDA) introduced several new approaches. To expedite the development of drugs that "address unmet medical needs in the treatment of a serious or life-threatening condition," the agency established fast-track designation, breakthrough therapy designation, accelerated approval, and priority review designation.

In these new regulatory pathways, single-arm trials, which have the potential for bias because they lack a comparator arm, have provided support for regulatory approval. In some cases, new drugs have been granted accelerated approval solely on the basis of ORR endpoint.

This is an appropriate endpoint, concluded a recent study published in the June issue of JAMA Oncology.

Specifically, the authors concluded that their "data suggest that high ORR (eg, statistically exceeding an ORR of 30%) is an appropriate end point for single-arm trials aiming to demonstrate breakthrough activity of a single-agent anticancer therapy."

With corresponding author Laurence H. Schwartz, MD, from the Department of Radiology at Columbia University, New York, these researchers also report "that single-agent regimens with ORRs statistically exceeding 45% had a 100% approval rate, while those with ORRs statistically exceeding 30% had an 89% approval rate.

"This analysis was restricted to trials in non-small cell lung cancer (NSCLC), colorectal cancer (CRC), renal cell carcinoma (RCC) and melanoma — all of which may be easily measured by the Response Evaluation Criteria In Solid Tumors (RECIST).

"[F]urther study will be needed before these data can be applied to such cancers as breast cancer, prostate cancer, and ovarian cancer, which can have a substantial burden of disease that is more difficult to measure on computed tomography," the researchers write in their discussion.

Approached for comment, Alan P. Venook, MD, from the Helen Diller Family Comprehensive Cancer Center at the University of California at San Francisco, explained that some variants of the cancers included in this analysis may not be easily measurable. For example, patients with BRAF V600E mutant metastatic CRC often have peritoneal disease or malignant ascites and therefore do not have easily "measurable" disease — this is a critical issue because different research strategies are being taken to target this specific subset of patients with CRC.

"One advantage of ORR (as opposed to time-to-event endpoints such as PFS or OS) is that it can be accurately measured in single-arm trials because a tumor response can be directly attributed to therapy, and spontaneous regression is extremely rare in the absence of therapy," write the authors of an invited commentary.

Gideon M. Bluementhal, MD, and Richard Pazdur, MD, from the FDA's Center for Drug Evaluation and Research, are the authors. They contend that the decades-long experience with RECIST allows for "comparisons with historic controls and established benchmarks for novel therapies to exceed in order to be better than available therapy."

The FDA commentators also indicated that the analysis presented in the JAMA Oncology report was consistent with their own meta-analysis of therapies in NSCLC, "in which a large magnitude of ORR effect was associated with a large magnitude of progression-free survival improvement."

Details of the Retrospective Analysis

The analysis was carried out on 578 trials from the Aggregate Analysis of database from the Clinical Trials Transformation Initiative at Duke University.

In total, 874 treatment arms were analyzed, of which 542 arms reported ORR. Of the ORR arms examined, 46% were in studies for NSCLC, 28% for CRC, and 13% each for melanoma and RCC. Sixty percent were phase 2 studies and 22% were phase 3 studies. Twenty-eight percent of the ORR arms were looking at single-agent therapies and 72% at combination regimens. Of single-agent arms, 15% led to regulatory approval.

Although ORR for some of the single agents exceeded maximum ORRs reported from previous studies, not all were approved on the basis of ORR. Sutinib (Sutent, Pfizer) in RCC was approved on the basis of an ORR of 53% and crizotinib (Xalkori, Pfizer) was approved in NSCLC on the basis of an ORR of 74%.

However, other agents were approved according to PFS or OS: afatinib (Gilotrif, Boehringer Ingelheim) in NSCLC, dabrafenib (Tafinlar, Novartis) in melanoma, axitinib (Inlyta, Pfizer) in RCC, and vemurafenib (Zelboraf, Genentech) in melanoma.

The researchers point out that while a high ORR may be adequate for regulatory approval for single agents, combination therapy is more likely to require a randomized trial to show that adding a second agent is likely to improve PFS or OS.

What Endpoint Is Appropriate?

The FDA commentators pointed out that several other factors are considered in regulatory approval: clinical pharmacology, safety profile, context of the malignant neoplasm, unmet medical need, and the availability of other safe and effective therapies.

A high ORR is, therefore, not always a slam-dunk for regulatory approval.

For example, for dacomitinib (developed by Pfizer) in NSCLC, an ORR of 54% in a phase 2 trial did not lead to regulatory approval for the agent in a disease for which other effective agents are available, including gefitinib (Iressa, AstraZeneca), erlotinib (Tarceva, Genentech), afatinib, and icotinib.

The FDA regulators also indicate that their experience with immune checkpoint inhibitors suggests that ORR may not fully capture the benefits of these agents, and one may need to consider other factors, such as tumor growth kinetics, depth of response, durability of response, and tumor volume.

These alternate metrics may provide insights into the clinical benefit of an agent, and Drs Blumenthal and Pazdur advise drug developers and researchers to use these metrics "to assist in compound prioritization, optimization of combinatorial approaches, and to better inform 'go/no-go' decision making."

"For regulators, more sophisticated and refined metrics will assist in identifying future breakthrough therapies and in developing better surrogates to predict long-term clinical outcome," the regulators conclude.

ASCO Recommendations

Recommendations for which endpoints to consider in clinical trials were laid out by the American Society of Clinical Oncology (ASCO) Clinically Meaningful Outcomes Working Group in 2014.

This document noted that for pancreatic, lung, colon, and breast cancer, PFS and OS were considered appropriate treatment goals.

For other tumor types, a clinically meaningful improvement of 25% and an absolute increase of 2.5 months in PFS and/or OS compared with standard-of-care treatment was considered appropriate.

How many drugs approved by the FDA achieve these clinically meaningful improvements suggested by ASCO? That was the question posed in a study published in June in JAMA Oncology, with corresponding author Sham Mailankody, MBBS, from the Memorial Sloan Kettering Cancer Center, New York, New York.

The team found that of 47 therapies approved, 10 (21%) received accelerated approval on the basis of single-arm studies, which precluded quantification of OS and/or PFS and comparison with standard treatment.

Of the 47 therapies, 25 (53%) met the requirements of PFS and only 9 (19%) met the standards for OS.

"Although we recognize the importance of incremental gains in oncology, we must also accept that the concept of building on incremental gains by combining marginally effective regimens has not brought the substantive progress for patients with cancer that we need to achieve," the team comments.

"[We] believe these data reinforce the need for continued engagement of all stakeholders in ensuring we do better for our patients," they conclude.

Current Thinking for Endpoints for Regulatory Approval

Medscape Medical News reached out to several researchers involved in clinical trials to determine what may be appropriate endpoints to determine drug efficacy.

Dr Venook commented on the endpoint in trials on CRC.

Patients with CRC have probably received many lines of therapy late in their disease, Dr Venook explained to Medscape Medical News. It will be difficult to get meaningful responses in these patients. In considering regulatory approval, it is important to take into account the agent under consideration and the course of disease. He indicated that the ASCO 2014 recommendations would be appropriate for endpoints in CRC — for advanced disease, an improvement in OS and PFS of 3 to 5 months.

"In the current environment, lack of a response may not be relevant," he said. "Durable stable disease may be beneficial to patients who are clearly progressing when they begin study treatment," Dr Venook added.

Brian I. Rini, MD, from the Cleveland Clinic in Ohio, commented on clinical trials in RCC.

"ORR may be an acceptable endpoint for regulatory approval when not too many agents exist," he said. "It is an immediate and an attainable endpoint," he added. Sunitinib was among the first agents to have an approval based on ORR. However, there are now 11 approved agents for RCC. Most of the other agents have been approved on the basis of PFS as a regulatory endpoint, but one of the newest agents approved for this indication showed an OS benefit — that was the immunotherapy, nivolumab (Opdivo, Bristol-Myers Squibb).

Clinical trial end-points were discussed by Michael A. Postow, medical oncologist at the Memorial Sloan Kettering Cancer Center and Georgina Long, PhD, MBBS, from the Melanoma Institute Australia at the University of Sydney.

"With increasing personalized approach in treating patients with cancer, large clinical trials are going to be harder to do," Dr Postow told Medscape Medical News. "Although survival is a gold standard, we need to have shorter endpoints and not one that requires a long follow-up time," he added.

In addition, Dr Postow pointed out for melanoma, future drugs, even if effective, may be unable to exceed survival benefits seen with currently approved agents.

"PFS rate is the best, most accurate endpoint [in melanoma trials] because it encompasses all clinical benefits," Dr Long told Medscape Medical News.

She recently coauthored a commentary in the Lancet Oncology, with her colleague Paolo A. Ascierto, MD, from Naples, Italy. They argue for PFS rate as an important endpoint to be reported in clinical trials.

Dr Long explained to Medscape Medical News that PFS encompasses all patients who benefit, including those who do not show a response based on RECIST, but who have a prolonged stabilization of their cancer.

"Unlike the restrictive measure of duration of response, which only analyzes the benefit of responders, progression-free survival records a negative event if a patient progresses, dies, or changes anticancer therapy," Dr Ascierto and Dr Long, write, arguing that these events are "the most relevant clinical endpoints in treating patients with advanced cancer."

"But the question is, what measure of progression-free survival should be used?" they add.

They suggest that median PFS may not consistently reflect the long-term benefits of a drug, and landmark PFS rates at 1 year, 2 years, and 3 years should be consistently reported in clinical trials.

"This analysis incorporates both tumour control and duration of control, is patient centred (unlike hazard ratios which are difficult for patients to understand), is easy to understand in terms of benefit, and can be determined in a timely manner without post-progression treatment confounding its interpretation," they argue.

According to Dr Long, ORR may be the worst endpoint for targeted therapies. "It is not the best for time to progression," she said. Even when patients respond, primary resistance quickly sets in, she explained.

Several authors receive consulting fees and honoraria from several pharmaceutical companies.

JAMA Oncol. 2016;2:772-779. Abstract Commentary

Follow Medscape Oncology on Twitter: @MedscapeOnc


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.