Long-term or open-ended treatment with PD-1/PD-L1 inhibitors is a burden for patients and the healthcare system and does not fit the biological paradigm of cancer immunotherapies, experts say; however, challenges to studying shorter durations include fears of undertreatment and patient objections to stopping a therapy that appears to be working.
More studies are needed on the optimal duration of therapy with immuno-oncology agents such as PD-1/PD-L1 inhibitors, industry representatives and academic researchers said at a recent US workshop on immunotherapy combination treatment.
There is growing evidence that patients may not need to be treated with immunotherapies indefinitely, or even for as long as the two years commonly studied in clinical trials. Rather, shorter courses of treatment may be just as effective and provide an option for retreatment should a patient progress after coming off a defined period of therapy, speakers said at the meeting sponsored by FDA and the American Association for Cancer Research (AACR).
"There are a couple of concepts that leaped into our practice that I think need to be proven correct before we adopt them. One of them is this concept of unlimited treatment," said David McDermott, an oncologist at Beth Israel Deaconess Medical Center and Dana Farber/Harvard Cancer Center who specializes in melanoma and kidney cancer.
"Many of our trials are unlimited PD-1," McDermott said. "We haven't proven you need that, and I think before we adopt that as a rule … we need to prove that that's real."
Shorter durations of immunotherapy could mean less financial burden for payers and patients and reduced risk of long-term toxicities for patients, but also less revenue for drug sponsors.
Unlimited immunotherapy treatment "is actually burdensome," said Israel Lowy, vice president of global clinical development and head of translational science and clinical oncology at Regeneron Pharmaceuticals Inc. "It's not just burdensome for the patient. It's burdensome for the healthcare system."
"If it's really not necessary, then maybe we shouldn't be doing it," Lowy said. "There's really no data that suggest that if you have a response, that continuing to treat … is going to make that duration of that response when you stop longer."
The optimal duration of immune-based therapy "is an unanswered question for our field," said Daniel Chen, vice president and global head of cancer immunotherapy development at Genentech Inc.
Chen suggested the answer to the question about optimal treatment duration may be found in patient subsets that are not well understood. "Some patients when you stop therapy have very durable responses, other patients have some durability but ultimately progress. Other patients might immediately progress upon stopping immune-based therapy," he said. "What I think has been difficult is there haven't been any clear identifiers for which patients belong to which subset."
Shorter durations of therapy could mean less financial burden on payers and patients and reduced risk of long-term toxicities for patients. The Catch-22 for the industry, however, is that shorter courses of treatment could mean reduced sales for immunotherapy drug sponsors.
While there was general agreement that the idea of shorter durations of therapy requires further study, researchers and industry executives cited barriers to developing the data needed to change from the current treatment paradigm of "more is better."
These hurdles include physician concerns that prematurely taking a patient off treatment will lead to cancer progression, and worries that clinical trial participants will seek supplemental immunotherapy elsewhere, thereby muddying the results of studies on shorter therapy duration. In addition, the duration of therapy, like the dosage amount and frequency, may depend upon the type of cancer being treated, which complicates extrapolation across indications.
Closing The Door On Open-Ended Treatment
The FDA/AACR meeting focused on approaches to combination therapy and best practices in patient and dose selection, biomarkers to aid in selection, and novel endpoints.
FDA and AACR have hosted several recent meetings on unique challenges in immunotherapy development, including an October 2016 session on new ways to measure efficacy. (Also see "Cancer Immunotherapies Have FDA, Industry Looking For New Endpoints" - Pink Sheet, 19 Oct, 2016.)
One theme that quickly emerged from discussions on optimizing the dose for studying immuno-oncology agents in combination therapy was uncertainty as to how long such drugs should be given, particularly in light of the somewhat arbitrary duration (often up to two years) that treatments such as PD-1/PD-L1 inhibitors are administered in clinical trials.
The Dosage and Administration section of labeling for all five currently approved PD-1/PD-L1 inhibitors recommends giving drug until disease progression or unacceptable toxicity. In addition, labeling for Merck & Co. Inc.'s PD-1 inhibitor Keytruda (pembrolizumab) advises dosing for up to 24 months in patients without disease progression.
Before immunotherapy agents were established as a viable approach to treating cancer, the treatment paradigm was to give toxic drugs until either disease progression or patient intolerance, said Regeneron's Lowy said, who previously worked at Medarex Inc. and was involved in that company's early development of the anti-CTLA-4 Yervoy (ipilimumab) and the PD-1 inhibitor Opdivo (nivolumab). Bristol-Myers Squibb Co. acquired Medarex in 2009.
Open-ended treatment "just doesn't fit the biologic paradigm. It just scientifically doesn't make sense. We don't do this, and we shouldn't do this just because we can." – Regeneron's Lowy
"What that reflects actually is that you're relying on the drugs to really do the entire job of controlling the pathogenic process, in this case the tumor," Lowy said. However, If you are "looking at an approach where you're enabling the immune system … and you really reset the balance, then maybe you can actually comfortably lay off of it."
Lowy and other speakers pointed to research suggesting that the major pharmacodynamic effect with PD-1/PD-L1 inhibitors occurs in the first few weeks or months of treatment.
"I think If you look across a variety of trials, different indications, different agents, it generally emerges that most people have their response by six months, and certainly by 12 months they've stabilized whatever response they've had," Lowy said. "[It's] very unusual to see somebody who [after] a year-and-a-half has a major change and further shrinkage in their tumor."
Open-ended treatment "just doesn't fit the biologic paradigm," Lowy said. "It just scientifically doesn't make sense. We don't do this, and we shouldn't do this just because we can."
Stopping PD-1/PD-L1 immunotherapy after a set period of time also opens the door to future use of the drugs in the event of progression, he said.
"If somebody relapses or progresses while they're on this extended treatment, I'm pretty sure that they're resistant to PD-1," Lowy said. "Whereas if I take the position that I'm going to put them in control with the defined treatment and follow them and then should they recur later, then I say well I have a crack now at redosing them."
"Just like if you have strep throat and you take penicillin for 10 days and a year later you get strep throat, it will respond to penicillin. But if you were taking penicillin for the entire year, I'm not sure you would want to stay on the penicillin," he said.
Novel Trial Designs Necessary But Challenging
Stopping immunotherapy early to see what happens "is an interesting concept" that begs further research using novel trial designs, such as randomized discontinuation, McDermott said.
However, "those kind of designs are very hard to execute in humans because they don't like to be taken off treatment when it's working, particularly if it's not adding side effects," he said. "One of the interesting things about PD-1 is that most of the tox happens in the first six months, so patients are not as eager to come off, in my experience, PD-1 as they might be Taxol for example, where they're accumulating side effects," McDermott said.
"We need to look at different durations, we need to look at novel designs because there are probably some patients who need chronic blockade and others that can come off, particularly those with deep responses," McDermott said. "We need to be doing short courses of treatment, stopping the drug, see what happens, reintroduce the drug if patients progress."
Randomized trials are needed to explore not just stopping early, but also possible predictors of success with an earlier stop, McDermott said, noting that optimal duration may depend, in part, on the type of cancer.
Randomized discontinuation trials "are very hard to execute in humans because they don't like to be taken off treatment when it's working, particularly if it's not adding side effects." – Beth Israel's McDermott
In addition, randomized studies are necessary to demonstrate that treatment beyond progression with immunotherapy is useful before the clinical community adopts this approach as standard of care, McDermott said.
This issue has been "explored in retrospective analyses of giving PD-1 after a patient progresses and showing that the patients who continue on drug do better than those who stop," he said. "I think those analyses are very biased. I think they are a set up for people in the community overtreating patients. And I think that's a particular problem because most progression in my experience outside of melanoma is real progression, and if you don't move that patient onto something else they're going to be in deep trouble."
Fear Of Undertreatment
Industry representatives and researchers acknowledged that fear of undertreatment can make drug manufacturers, clinicians and patients reluctant to pursue more abbreviated dosing schedules with immunotherapy.
Shortening the duration of treatment potentially can have life or death consequences, Chen said, recounting an experience from his company's development of the PD-L1 inhibitor Tecentriq (atezolizumab).
"Back in 2011, I was leading the atezolizumab development and I was really championing the idea of more limited duration of therapy, with the idea that you could retreat," Chen said. "And in that experience I know we lost some patients … because we stopped treatment."
"At least for the early portion of this field, we moved to treating more because of that concern over what happens if you undertreat." – Genentech's Chen
"Few things have more deeply affected me personally than knowing that while I was pushing what I thought was an important scientific point, that these things have very real ramifications to people's lives," Chen said. "That affected the way I look at duration of treating and the idea of overtreating."
"We tend to overtreat because it's really important, because a patient's life is at stake," Chen said. "We're willing to tolerate a little bit of overtreatment because we know it's that chance to save someone's life. … I think what you've seen is that at least for the early portion of this field, we moved to treating more because of that concern over what happens if you undertreat."
A Changing Tide In Study Treatment Duration?
Lowy described Regeneron's efforts to study more limited durations of therapy with its investgational PD-1 inhibitor REGN2810. An ongoing Phase I trial caps treatment at 48 weeks, with an opportunity for retreatment in the event of recurrence, Lowy said.
Despite the more limited duration in the Phase I trial, "some of our Phase III studies that we have started have been planned to dose out to two years, and this is the 'just in case' from the people who don't want me to screw up with this expensive development plan," Lowy said. One consideration in this design decision was the concern that if the protocol called for a shorter duration of immunotherapy, clinical trial participants may subsequently seek PD-1 treatment elsewhere, which could confound the study results, he said.
Merck and other companies have begun reducing the duration of immunotherapy from two years to one in adjuvant studies, where patients are generally healthier than in the later-line treatment settings, said Eric Rubin, vice president and therapeutic area head of oncology early clinical development at Merck.
Figuring out the answers to duration of treatment questions is critical to optimizing use of immunotherapies, Lowy said.
"Given that this is currently our best agent out there as the anchor of almost everything else we do … it's really important that we further understand how to use it optimally," Lowy said. "I think we need to figure out how to find studies that we can do that would help us get comfortable with this."
"Maybe it's just getting more and more data. Maybe there are other biologic indicators apart from PK that we could use to say you've hit your mark and you can stop," Lowy said. "This is an important issue which will actually make a huge impact not only patient care, burden, healthcare cost, but also even how when we start doing combinations how they actually work out in terms of being tolerable and effective."