[Editor’s note: Ben Fidler co-authored this report.] A decade from now, it’s possible that immunotherapy will have made a big dent in lung cancer, which is by far the deadliest type of the disease.
Drugs from several companies have worked their way forward, initially for patients who have failed chemotherapy, and now, in some cases, for patients newly diagnosed with advanced cancer.
They’re not miracle drugs. Only a minority of patients have positive benefits. But those benefits tend to be long-lasting—often better than the results from chemotherapy and targeted therapies such as Pfizer’s crizotinib (Xalkori) or Roche’s erlotinib (Tarceva). Will immunotherapies combined with each other, or with chemotherapy and other drugs, help more patients? That’s what some of the latest clinical studies aim to find out.
The latest results for lung cancer are out today at the American Association for Cancer Research meeting in Chicago. Last week we asked lung cancer specialists what they would be looking for, and today we’ll break the studies down here. (Note: All of the trials were for patients with advanced, previously untreated or first-line lung cancer.)
Pembrolizumab (Keytruda), made by Merck (NYSE: MRK), was approved nearly a year ago, when combined with chemotherapy, for this patient group.
But that FDA approval was based on limited evidence. The new, much larger study, dubbed Keynote-189 and including 616 patients, is meant to reinforce the previous study. Funded by Merck, it showed that patients taking pembrolizumab and chemotherapy (let’s call it pembro-chemo) fared better in several ways than patients just taking chemo. The study is not finished, but it has reached its first planned milestone. The results, published in the New England Journal of Medicine this morning, are very good. The risk of death or worsening disease for the pembro-chemo patients was half that of the chemo-only patients. “We’ve never seen this magnitude of benefit over chemotherapy with any other regimen,” said study leader Leena Gandhi, director of thoracic medical oncology at NYU Langone’s Perlmutter Cancer Center in New York City.
The data reinforce pembrolizumab’s advantage over rival nivolumab in first-line lung cancer, ISI Evercore analyst Umer Raffat wrote in a research note Monday. But he cautioned that the advantage does not apply to other cancers or even to second-line lung cancer, and it’s not clear why. “I still scratch my head,” wrote Raffat.
In a measure of how long the drugs halted the march of a patient’s cancer, median progression-free survival was 8.8 months for the pembro-chemo patients and 4.9 months for the chemo-only patients. Overall survival one year after starting treatment was 69 percent for the pembro-chemo patients and 49 percent for the chemo patients.
Safety is always a big deal in Phase 3. Problems often do not emerge until larger groups of patients are tested in randomized settings. In Keynote-189, in which two-thirds received pembrolizumab and chemo, the rate of serious side effects was nearly equal (67.2 percent versus 65.8 percent). However, the pembrolizumab arm had twice the rate of serious side effects (8.9 percent versus 4.5 percent) with an immune-related cause. Gandhi says that was expected with the addition of immunotherapy. The worry with immunotherapy is that it will work too well and kick the patient’s immune system into dangerous overdrive. But the only side effects standing out for Gandhi were higher liver problems, which are associated with the chemotherapies used in Keynote-189. Patients with pre-existing kidney problems would not be put on this particular regimen, she says.
Another important data point was that the pembro-chemo combination was an improvement no matter how much of the protein PD-L1 was on the patient’s tumor. Pembrolizumab and other drugs like it are designed to block PD-L1, which some tumors use to evade the immune system. Measuring a tumor’s PD-L1 abundance has been the only way to predict which patients might respond better to these new drugs. It is an imperfect measurement and open to interpretation, but “it is the best we have,” says Gandhi. (Pembrolizumab was approved in 2016 without chemo to treat patients with 50 percent or more PD-L1 on their tumors.)
Keynote-189 showed that the survival benefit of pembro-chemo accrued to all PD-L1 categories—even patients who might not have any (less than 1 percent) on their tumor. Their reduction of risk of death was about 40 percent, and those with abundant PD-L1 saw their risk reduced nearly 60 percent.
Merck threw a layer of intrigue over Keynote-189 and other lung cancer results from AACR when, one week ago, it announced that similar patients in a different trial, Keynote-042, survived “significantly longer” after taking pembrolizumab only—no chemo—than patients taking chemo only. Merck did not divulge details, and the study did not include patients with less than 1 percent PD-L1 on their tumors.
So talk of pembrolizumab alone knocking aside chemo not just for high-PD-L1 tumors but for most first-line lung cancer patients is a bit premature, says Gandhi. “We know from the Keynote-042 release there was a survival benefit, but the magnitude is unknown. Was it driven by high PD-L1 expressors? We can’t interpret it yet or compare.”
Even with more details, however, she emphasizes that treatment decisions won’t be based just on trial data but a patient’s condition and, in coming months and years, other biological indicators beyond PD-L1 status that are under exploration—as we’ll see in a moment. “The exciting thing is that we have more possibilities than ever before,” she says.
Merck’s success with pembrolizumab and pembro-chemo have amped up the pressure on rival Bristol-Myers Squibb (NYSE: BMY), whose immunotherapy nivolumab (Opdivo) has fallen behind its rival in lung cancer. Bristol is trying to show that the combination of nivolumab and its other approved immunotherapy, ipilimumab (Yervoy), has a role to play. (Let’s call this nivo-ipi from now on.) It’s focusing on people whose tumors have unusually high levels of genetic mutations. The measure, known as tumor mutational burden, or TMB, is emerging as a new way, beyond PD-L1, to predict who might respond to immunotherapy. The study’s lead investigator, Memorial Sloan Kettering Cancer Center medical oncologist David Hellman, says Checkmate-227 “validates the value of TMB.” (Bristol and its partner Ono Pharmaceutical paid for the study.)
“We’re carving out a distinct population of patients with lung cancer that can benefit from immunotherapy by itself and avoid the need for chemo in the first-line setting,” he says.
The study dosed patients with TMB of 10 mutations per megabase or greater. This is Bristol’s definition of high TMB; there is not yet a “community consensus about what TMB is precisely,” as The Jackson Laboratory president and CEO Edison Liu, a cancer genomics expert, notes.
Another caveat: a subset of the more than 1,700 patients enrolled in Checkmate-227 were measured as high TMB; 139 received nivo-ipi and 160 received chemotherapy alone. Those numbers make for a small sample size compared to many other Phase 3 immuno-oncology trials. But Hellmann says it nonetheless found a “representative population and a statistically relevant degree of benefit.”
The results were published this morning in the NEJM: 42.6 percent of nivo-ipi patients with high TMB registered progression-free survival at the one-year mark, versus 13 percent of chemo patients. Median progression-free survival was 7.2 months for nivo-ipi patients with high TMB, versus 5.5 months for chemo patients. Nivo-ipi led to a 42 percent reduction in risk of disease progression or death for high TMB patients. The combo’s benefit over chemotherapy was “broadly consistent within subgroups,” the NEJM paper says, including patients with very low levels of PD-L1. According to the paper, 38 nivo-ipi patients had less than 1 percent of PD-L1 on their tumors. and 101 had more than 1 percent. (Nivo-ipi didn’t fare nearly as well all-comers: 1-year progression-free survival was 30.9 percent for the combo compared to 17 percent for chemo patients, and median progression-free survival was 4.9 months. The results equaled a 17 percent risk reduction in disease progression or death for nivo-ipi versus chemo.)
There was improvement over time as well: 68 percent of the high-TMB patients who responded to nivo-ipi had their tumors held in check for at least a year, compared to just 25 percent of the chemo responders, suggesting that those who had a response to the combo would see a durable benefit.
Safety is an important consideration for this study. Bristol’s ipilumumab, which attacks a different tumor protein called CTLA-4, historically has harsher side effects than other immunotherapy products. The results are in Bristol’s favor. Serious side effects hit the nivo-ipi patients at a slightly lower rate than the chemo patients, 31 percent versus 36 percent. To that end: Dangerous lung inflammation (pneumonitis), which has been associated with ipilimumab at a higher dose for melanoma patients, occurred in just 3 percent of patients on the combination, says Hellmann.
“I could see that it would be a reasonable option for patients that do not want chemotherapy,” says Laura Chow, the associate director of the Phase 1 Clinical Trials Program at Seattle Cancer Care Alliance. “But you still need exceptionally fit patients.” She adds that the response rate to nivo-ipi, while impressive, is comparable to what pembrolizumab alone has shown. “Do you want that extra toxicity?” she says.
So who might be eligible for high-TMB immunotherapy? About 44 percent enrolled in Checkmate-227 had high TMB by Bristol’s definition. Hellmann expects that is “probably in the ballpark” of the worldwide NSCLC population.
Health economist Gary Lyman, who specializes in the cost of cancer care, says he’s skeptical and wants to see more data before putting a solid number on the high-TMB population. “Such population claims can rarely be made from one study,” says Lyman, co-director of the Hutchinson Institute for Cancer Outcomes Research in Seattle.
Whatever the population, it is no sure thing that nivo-ipi will become the preferred option. Here are some concerns: lung cancer experts want to see how much the combination extends patients’ lives. Bristol doesn’t have the final numbers yet, but early results published today show 67 percent of the TMB-high patients on nivo-ipi were alive after one year of treatment, compared to 59 percent of those on chemotherapy. Though early, Hellmann says that slight separation “is certainly encouraging.”
Doctors also expressed concern about the two to three weeks it might take to test for TMB, which requires DNA sequencing, rather than getting patients started faster on another regimen—say, pembrolizumab, or pembro-chemo. Hellmann acknowledges “the turnaround time matters” but argues that tests are become more frequent and routine, meaning delays could start to shrink.
What of the price of lumping two expensive immunotherapies together, rather than one and a cheaper chemotherapy? Confining nivo-ipi to high-TMB patients could mitigate the impact of its cost, says Hellmann. But when people respond to immunotherapy, those responses tend to last, and they keep getting treated. Chow has patients who have been on immunotherapy more than 6 years.
“Nobody wants to go off of it,” she says. “I’m not sure that’s sustainable.”
That concern extends to all checkpoint inhibitors—another name for nivolumab, pembrolizumab, and their rivals—all with a price tag of roughly $150,000 a year. Lyman doesn’t see competition between Merck, Bristol, and others bringing costs down anytime soon. Instead, he says, to reduce unnecessary costs, biomarkers like TMB “will be key” to identify the patients who should be taking a drug, and who shouldn’t.
To some extent, Roche has already answered the two main questions for this 1,200-patient study: whether its four-drug cocktail can lengthen the time a patient remains alive (overall survival); and how long her cancer is kept from spreading (progression-free survival). Its immunotherapy atezolizumab, when added to Roche’s bevacizumab (Avastin) and a one-two punch of chemotherapy, held tumors in check—8.3 months, compared to 6.8 months for Avastin plus chemo. Those results were detailed at a medical meeting in December.
Roche has also previously revealed that IMPower150’s four-drug combo led to a median overall survival benefit of 19.2 months, versus 14.4 months for Avastin plus chemo. (Updated OS details won’t emerge until the American Society of Clinical Oncology conference in June, according to a spokeswoman.)
Here’s what’s new at AACR this week: Roche is revealing how patients with varying levels of PD-L1 fared in the study. Patients with higher PD-L1 levels had better results, although Roche used two different tests to measure PD-L1 levels, with slightly different readings, which underscores how PD-L1 readings are not black-and-white.
That said, the benefits for people with less than 1 percent PD-L1 levels were minimal: either a 0.2-month or 1.2-month difference in PFS. For those with the highest PD-L1 levels, the immunotherapy combination provided a 2.9 month or 5.8 month PFS difference, depending on the PD-L1 test.
With one diagnostic, for instance, Roche’s regimen kept cancers in check 12.6 months for patients with at least 50 percent PD-L1 compared to 6.8 months for Avastin plus chemo (the results were 9.1 months to 6.9 months, comparatively, using the other test).
That seems encouraging for Roche, but Merck has beaten it to the punch. As already noted, pembrolizumab without chemo is already approved for high PD-L1 patients, and the top-line Keynote-042 results Merck announced last week have experts wondering if pembrolizumab alone will have significant benefit for all patients in this area, regardless of PD-L1 status.
While options with fewer agents and potentially less toxicity look promising, Roche’s four-drug regimen uses paclitaxel, which gives pause to specialists such as Young Kwang Chae, co-director of the clinical trials unit at the Lurie Cancer Center in Chicago, because of side effects like peripheral neuropathy. And bevacizumab, which causes people to cough up blood, can be tough to deal with as well. The Keynote-189 and Checkmate-227 regimens don’t use paclitaxel or bevacizumab, and Bristol’s combo doesn’t use chemo at all.
IMpower150 investigator Mark Socinski, the executive medical director of the Florida Hospital Cancer Institute, says the Roche regimen could be for those who can handle bevacizumab. “It’s not a drug for everybody,” Socinski says, but “I think there’s a role for it.” Other specialists have also noted the IMPower150 regimen has shown some benefit for patients with a mutation in the EGFR gene who haven’t fared well on targeted drugs like erlotinib.
Still, Chow of the Seattle Cancer Care Alliance thinks the combo’s high costs will be problematic: “I don’t think most insurance companies are going to be on board with this, even if it does get FDA approved,” Her group tends not to use either paclitaxel or bevacizumab for NSCLC because of their side effects.
All the regimens are going to be expensive, says Socinski. One next step to control costs, he says, should be to figure out how to reduce unnecessary doses. Some patients have stopped taking immunotherapies because of side effects, but their responses have continued anyway. “Does [immunotherapy] really need to be given as frequently as it is?” he says. “We haven’t put a lot of effort into studying that aspect.”
Image of lung cancer metastasis courtesy of the National Cancer Institute.