‘Striking’ lung cancer data emerge, but uncertainty remains about upfront immuno-oncology combinations

This year’s American Association for Cancer Research Annual Meeting featured several studies that could potentially transform the care of patients with non-small cell lung cancer.

They included the KEYNOTE-189 trial, which assessed the addition of pembrolizumab (Keytruda, Merck) to chemotherapy for certain patients with metastatic disease; the IMpower150 study, which evaluated the addition of atezolizumab (Tecentriq, Genentech) to bevacizumab (Avastin, Genentech) plus chemotherapy for patients with NSCLC; and the CheckMate227 trial, which compared the combination of two immunotherapies — nivolumab (Opdivo, Bristol-Myers Squibb) and ipilimumab (Yervoy, Bristol-Myers Squibb) — with standard chemotherapy for patients with newly diagnosed advanced NSCLC with high tumor mutational burden.

HemOnc Today spoke with Corey J. Langer, MD, director of thoracic oncology at Abramson Cancer Center at University of Pennsylvania Health System and professor of medicine at University of Pennsylvania, about the results of these investigations and their potential clinical implications.

KEYNOTE-189

The double-blind, randomized phase 3 KEYNOTE-189 trial showed the addition of the anti-PD-1 therapy pembrolizumab to chemotherapy significantly improved outcomes for patients with newly diagnosed metastatic nonsquamous NSCLC who did not harbor EGFR or ALK alterations.

The regimen halved the risk for death, demonstrating “an unprecedented effect of therapy in the first-line setting” for this patient population, according to researcher Leena Gandhi, MD, PhD, associate professor in the department of medicine and director of the thoracic medical oncology program at Perlmutter Cancer Center at NYU Langone Health.

Results from cohort G of the randomized phase 2 KEYNOTE-021 trial showed the addition of pembrolizumab to chemotherapy with pemetrexed plus carboplatin significantly improved PFS and overall response rate among patients with advanced nonsquamous NSCLC.

Gandhi and colleagues assessed whether an immunotherapy-chemotherapy combination would confer greater benefit as first-line therapy.

The KEYNOTE-189 analysis included 616 patients with previously untreated stage IV nonsquamous NSCLC. All patients had ECOG performance status of 0 or 1, and they had no EGFR or ALK alterations.

Researchers randomly assigned 410 patients to pembrolizumab 200 mg, pemetrexed 500 mg/m², and either carboplatin area under the curve 5 or cisplatin 75 mg/m². The other 206 patients received placebo plus chemotherapy.

Patients received their assigned treatment every 3 weeks for four cycles, followed by maintenance therapy with pemetrexed plus either pembrolizumab or placebo.

Eligible patients assigned the placebo regimen who developed verified progressive disease could cross over to pembrolizumab monotherapy.

OS and PFS in the intention-to-treat population served as primary endpoints. Median follow-up was 10.5 months (range, 0.2-20.4).

Patients assigned pembrolizumab plus chemotherapy achieved significantly longer median OS (not reached vs. 11.3 months; HR = 0.49; 95% CI, 0.38-0.64) and median PFS (8.8 months vs. 4.9 months; HR = 0.52; 95% CI, 0.43-0.64). The benefits persisted regardless of patients’ PD-L1 tumor proportion scores.

“These were the most striking data we saw [at AACR Annual Meeting],” Langer told HemOnc Today. “The P value for OS had seven zeroes in it. ... When have we seen an HR below 0.5 in oncology? And remember, that was an aggregate. Among patients with [PD-L1 tumor proportion scores of 50% or higher], the HR was 0.42. In my entire career, I don’t think I’ve seen a separation in OS curves like this.”

A similar percentage of patients assigned pembrolizumab and placebo experienced grade 3 or higher adverse events (67.2% vs. 65.8%). However, a higher percentage of patients assigned pembrolizumab discontinued any treatment due to adverse events (27.7% vs. 14.9%), discontinued all treatment due to adverse events (13.8% vs. 7.9%) or died due to adverse events (6.7% vs. 5.9%).

Most toxicities were expected, except for an elevated rate of acute kidney injury in the pembrolizumab group (overall incidence, 5.2% with pemetrexed vs. 0.5% with placebo; grade 3 to grade 5 incidence, 2% vs. 0%).

“There was some slight added toxicity, but it was clearly manageable,” Langer said. “I pay attention to treatment-related deaths. The overall rate of treatment-related grade 5 adverse events was around 2%, and the rate for immune-related grade 5 adverse events was 0.7%. That is well below the threshold that I would establish for advanced lung cancer.”

Sixty-seven patients assigned placebo crossed over to pembrolizumab. This equated to a 41.3% crossover rate in the intention-to-treat population, or a 50% rate when excluding patients who remained on treatment at the time of analysis.

“Despite the built-in crossover in the study design, we still see a major PFS and OS advantage,” Langer said.

It remains unclear — particularly for patients with 50% or higher PD-L1 expression — whether pembrolizumab alone or a triplet regimen is the best approach, Langer said.

Results of a planned ECOG/SWOG trial may answer that question.

The design calls for patients with advanced nonsquamous NSCLC to be randomly assigned to one of three regimens: first-line pembrolizumab, followed by second-line carboplatin and pemetrexed; first-line pembrolizumab, followed by second-line carboplatin, pemetrexed and pembrolizumab; or first-line induction therapy with carboplatin, pemetrexed and pembrolizumab, followed by maintenance therapy with pemetrexed and pembrolizumab. Second-line treatment will be up to the individual clinician.

“In patients with PD-L1 scores of 50% or higher, the decision of pembrolizumab alone or chemotherapy plus pembrolizumab is a practical one,” he said. “If a patient is younger and has few comorbidities and a high metastatic burden — so they’re highly symptomatic from the cancer but not terribly sick otherwise — I figure they have one shot, and I’m going to give them the triplet,” Langer said. “If they’re older, frailer and have multiple comorbidities, maybe with a lower metastatic burden, I’ll go with the single agent.

“I have no way to discern in the [group of patients with PD-L1 expression of 50% or higher] what the best approach is,” Langer said. “In the [group with 1% to 50% expression], I wouldn’t necessarily say the triplet is the standard, but it is a standard.”

IMpower150

The addition of the anti-PD-L1 antibody atezolizumab (Tecentriq, Genentech) to bevacizumab (Avastin, Genentech) plus chemotherapy extended PFS for patients with nonsquamous NSCLC, according to results from the IMpower150 study.

Atezolizumab is approved in the United States for second-line or later treatment of NSCLC, regardless of PD-L1 expression.

IMpower 150 assessed the addition of atezolizumab in the first-line setting for 692 patients with nonsquamous NSCLC. Patients received either 1,200 mg atezolizumab plus 15 mg/kg bevacizumab, carboplatin and paclitaxel, or a control regimen of bevacizumab plus carboplatin and paclitaxel.

Researchers reported longer median PFS in the experimental arm (8.3 months vs. 6.8 months) across all PD-L1 expression subgroups, including patients with PD-L1 expression of less than 1% or from 1% to 50% (HR = 0.62; 95% CI, 0.5-0.76).

For patients with PD-L1-low tumors, defined as those with expression between 1% and 50% (n = 140), median PFS was 9.7 months for those assigned atezolizumab and 6.9 months for those assigned the control regimen (HR = 0.57; 95% CI, 0.38-0.84). Researchers observed a similar difference among patients with PD-L1 expression of 50% or higher (9.1 months vs. 6.2 months; HR = 0.5; 95% CI, 0.33-0.77).

Patients with EGFR exon 19 deletion of L858R (n = 59) achieved improved PFS with the addition of atezolizumab (10.2 months vs. 6.1 months; HR = 0.41; 95% CI, 0.22-0.78).

Langer described that HR as “astounding.”

“The problem with this study has nothing to do with atezolizumab or bevacizumab. It has everything to do with the taxane,” Langer said.

Before IMpower150 was designed, Langer typically treated these patients with pemetrexed and carboplatin, with or without bevacizumab. His approach changed when immuno-oncology agents emerged.

“I just don’t see myself going back to taxanes in this population,” Langer said. “Pemetrexed is a much safer drug. There’s virtually no neuropathy, and there’s no hair loss. Patients look and feel like normal people. Put them on a taxane and they don’t feel normal.”

Still, the trial is important and the formal OS data are eagerly awaited, Langer said.

CheckMate227

Immunotherapy with the PD-1 inhibitor nivolumab plus the anti-CTLA-4 antibody ipilimumab significantly prolonged PFS compared with standard chemotherapy among patients with newly diagnosed advanced NSCLC with high tumor mutational burden, according to results of the randomized phase 3 CheckMate227 trial (see page 43).

“This establishes tumor mutational burden as a distinct and definable new subgroup of lung cancer and broadens the clinical actionability of data we can derive from routine next-generation sequencing,” researcher Matthew Hellmann, MD, assistant attending at Memorial Sloan Kettering Cancer Center, said during a press conference.

Platinum chemotherapy has long been the standard for patients with newly diagnosed NSCLC. PD-1 immunotherapies have emerged as a new treatment option for this patient population, but only a small percentage of patients with NSCLC respond.

The combination of the PD-L1 inhibitor nivolumab and the CTLA-4 inhibitor ipilimumab showed promising activity and tolerability as first-line therapy in a phase 1 trial of patients with advanced NSCLC.

The open-label CheckMate227 study compared three experimental regimens — nivolumab plus ipilimumab, nivolumab alone, or nivolumab plus platinum doublet chemotherapy — with platinum doublet chemotherapy alone as first-line treatment for patients with advanced NSCLC.

The analysis included 1,739 patients with chemotherapy-naive, histologically confirmed stage IV or recurrent NSCLC, ECOG performance status of 0 or 1, and no known sensitizing EGFR or ALK mutations.

Researchers assigned patients with PD-L1 expression of 1% or higher (n = 1,189) to one of three regimens: nivolumab 3 mg/kg every 2 weeks plus ipilimumab 1 mg/kg every 6 weeks (n = 396), nivolumab monotherapy dosed at 240 mg every 2 weeks (n = 396), or histology-based doublet chemotherapy (n = 397).

Investigators assigned patients with PD-L1 expression less than 1% (n = 550) to one of three regimens: nivolumab plus ipilimumab (n = 187), nivolumab 360 mg every 3 weeks plus histology-based chemotherapy (n = 177), or chemotherapy alone (n = 186).

Treatment continued up to 2 years, or until disease progression or unacceptable toxicity.

Researchers established two coprimary endpoints:

• OS for nivolumab-ipilimumab compared with platinum doublet chemotherapy among patients with PD-L1-selected tumors; and
• PFS determined by blinded central review for nivolumab-ipilimumab compared with platinum doublet chemotherapy among patients with high tumor mutational burden (at least 10 mutations per megabase).

Minimum follow-up was 11.2 months.

Among all randomly assigned patients, researchers reported an HR for PFS of 0.83 (95% CI, 0.72-0.96) with nivolumab-ipilimumab compared with chemotherapy.

Among patients with high tumor mutational burden, those assigned nivolumab plus ipilimumab achieved significantly longer PFS than those assigned chemotherapy (HR = 0.58; 97.5% CI, 0.41-0.81).

Results appeared consistent across subgroups, including those with nonsquamous histology (HR = 0.55; 95% CI, 0.38-0.8), those with squamous histology (HR = 0.63; 95% CI, 0.39-1.04), those with PD-L1 expression of 1% or higher (HR = 0.62; 95% CI, 0.44-0.88), and those with PD-L1 expression of less than 1% (HR = 0.48; 95% CI, 0.27-0.85).

The immunotherapy combination did not confer a significant PFS benefit among patients with tumor mutational burden less than 10 mutations per megabase (HR = 1.07; 95% CI, 0.84-1.35).

OS data for the PD-L1-selected coprimary population have not been released.

“This analysis is incredibly provocative, but I would not say it is game-changing yet,” Langer told HemOnc Today. “For one thing, the ipilimumab-nivolumab combination is not commercially available yet. Also, OS data are not mature. Last I checked, OS trumps PFS. My other big issue with this trial is that it is a subset of a subset.”

Although the trial enrolled more than 1,700 patients, fewer than 300 had high tumor mutational burden.

“In addition, assessing tumor mutational burden is tissue intensive and labor intensive,” Langer said. “It takes 4 to 6 weeks to get next-generation sequencing done. Tumor mutational burden hinges on next-generation sequencing. Also, only 58% of patients had sufficient tissue for tumor mutational burden analysis, so you also have an issue in terms of who’s actually eligible for this approach.”

Among all treated patients, those assigned chemotherapy experienced a higher rate of any-grade treatment-related adverse events (81% vs. 75%), as well as grade 3 to grade 4 treatment-related adverse events (36% vs. 31%). A higher percentage of patients assigned nivolumab-ipilimumab discontinued therapy due to adverse events (12% vs. 5%). One treatment-related death occurred in each group.

Langer told HemOnc Today he felt the data, as shown, “soft-pedalled the toxicity” of the ipilimumab-nivolumab combination.

“When they showed the percentages, they implied that [ipilimumab-nivolumab] was no more toxic; in fact, it might have been a little bit less toxic than chemotherapy,” he said. “But think about the toxicities that occur [with this combination]. These are not paper toxicities. They are real clinical toxicities, such as pneumonitis and nephritis, potentially bad skin rash, and colitis with diarrhea. Ipilimumab is a much tougher drug than the PD-1 or PD-L1 inhibitors.

“A lot of the chemotherapy side effects that are grade 3 or grade 4 are paper toxicity,” Langer added. “I never had a patient complain to me about their grade 3 thrombocytopenia. They don’t feel it. Unless they’re bleeding, it’s scored as grade 3 but it has no clinical ramifications. There needs to be a sharp distinction between clinical and paper toxicity.”

‘Provocative, not definitive’

On the basis of these data, some members of the lung cancer community believe combinations of immuno-oncology agents “will be the state of the art” going forward, Langer said.

“I hope they’re right, but I don’t think we’re there yet,” he said. “This is provocative but it’s not definitive. It’s not yet ready for prime time. It’s labor intensive, there are delays in getting tumor mutational burden back, and we have yet to see formal prospective survival data. Until that happens, it’s not quite ready.”

Langer predicted a certain subpopulation — possibly patients with low-PD-L1 disease and high tumor mutational burden — will derive the greatest benefit from a combination of immuno-oncology agents.

“You can say that is sparing chemotherapy, but what are we sparing?” Langer said. “We’re delaying chemotherapy. We’ll still have it in our back pocket, and that’s nice to know. But unless [a combination of immuno-oncology agents] upfront translates to a survival advantage, I’m not sure there’s any long-term benefit.”

“It’s also important to remember there are about a hundred other immuno-oncology agents out there, many of which are far less toxic than ipilimumab,” Langer added. “They are being looked at in phase 1 and phase 2 studies, generally in heavily pretreated patients, and some of those studies are generating intriguing results.”

Langer said he expects one or more of them in combination with a checkpoint inhibitor upfront will compare favorably with chemotherapy, an immuno-oncology agent alone, or perhaps ipilimumab-nivolumab among the subset of patients with high tumor mutational burden.

“I don’t know when that’s going to happen — it’s probably several years off — but if we have equivalence in the PFS and OS benefit but less toxicity with these other newer immuno-oncology combinations, ipilimumab will be displaced,” Langer said.

“I’m very much open to combinations of immuno-oncology agents but, at the end of the day, I take care of patients,” he added. “If they’re having horrendous toxicities, it’s bad for patients, and it’s bad for me and my nurses.”

References:

Langer CJ, et al. Lancet Oncol. 2016;doi:10.1016/S1470-2045(16)30498-3.

The following were presented at American Association for Cancer Research Annual Meeting; April 14-18, 2018; Chicago:

Gandhi L, et al. Abstract CT075.

Kowanetz M, et al. Abstract CT076.

Ramalingam S. Abstract CT078.

Disclosure: Langer reports an investigator role with KEYNOTE-021 cohort G. KEYNOTE-189 enrolled patients at his institution but he was not involved in the study. He also reports consultant and advisory board roles with Genentech/Roche and Eli Lilly.