During an OncLive Peer Exchange program, an international panel of lung cancer experts provided perspective on key findings regarding various immune checkpoint inhibitor (ICI) combinations and data concerning potential biomarkers of response to ICIs in non–small cell lung cancer.
Suresh S. Ramalingam, MD, FASCO
Immune checkpoint inhibitors (ICIs) have advanced rapidly in the management of non—small cell lung cancer (NSCLC), graduating from a salvage option to a frontline regimen. Despite the agents’ ability to produce durable responses and prolong overall survival (OS) in some patients with cancer, primary and acquired resistance remain significant obstacles.1
Investigators have invested significant effort over the past several years into identifying synergistic combinations, pairing ICIs that target different immune checkpoint pathways or combining 1 or more ICIs with chemotherapy with promising results.
In August 2018, the FDA approved the PD-1 inhibitor pembrolizumab (Keytruda) in combination with chemotherapy as a frontline regimen for metastatic nonsquamous NSCLC.2 In December 2019, the PD-L1 inhibitor atezolizumab (Tecentriq) gained an expanded frontline indication in combination with chemotherapy for the same patient population.3
The past year proved to be a pivotal time for new data on the prospects for ICI combinations in NSCLC. At the 2019 World Conference on Lung Cancer (WCLC) and the European Society for Medical Oncology (ESMO) Congress 2019, both of which took place in September, clinical trial investigators presented results from studies that assessed various ICI combinations as a treatment strategy for NSCLC. They also presented data concerning potential biomarkers of response to ICIs in NSCLC.
An international panel of lung cancer experts put the key findings into perspective during an OncLive Peer Exchange® program. Suresh S. Ramalingam, MD, who served as moderator, said the data suggest clinicians will eventually have 3 viable strategies for patients with advanced newly diagnosed NSCLC who lack targetable mutations: ICI monotherapy, immune checkpoint blockade plus chemotherapy, and dual immune checkpoint blockade pairing a PD-1/PD-L1 inhibitor with a CTLA-4 inhibitor. “I think in the upcoming months, we’ll be talking about the relative merits of 1 approach versus the other,” he said.
Dual Immune Checkpoint Blockade
Johan F. Vansteenkiste, MD, PhD, explained the rationale for combining an anti—CTLA-4 monoclonal antibody with an anti–PD-1/PD-L1 monoclonal antibody: “You tackle 2 points in the cancer immunity cycle, which may have a more pronounced effect than single-agent immunotherapy.”
At ESMO, investigators for the phase III CheckMate 227 trial reported that the study met its primary OS end point in a subgroup of patients with advanced or recurrent PD-L1—positive NSCLC and no EGFR or ALK alterations.4 Patients were randomly assigned to a combination of the PD-1 inhibitor nivolumab (Opdivo) and the CTLA-4 inhibitor ipilimumab (Yervoy), nivolumab monotherapy, or chemotherapy until disease progression or intolerance for up to 2 years.4
In the PD-L1—positive subgroup, median OS was 17.1 months in the nivolumab/ipilimumab arm, 15.7 months in the nivolumab monotherapy arm, and 14.9 months in the chemotherapy arm. The difference in OS between the nivolumab/ipilimumab arm and the chemotherapy arm was significant (HR, 0.79; 97.72% CI, 0.65-0.96; P = .007).4
Vansteenkiste said the persistence of the OS benefit in the nivolumab/ipilimumab arm was particularly interesting, with “plateaus at 1 year and at 2 years for nivolumab/ipilimumab- treated patients that were clearly above [plateaus for patients] given chemotherapy.” An OS benefit with nivolumab/ipilimumab versus chemotherapy was also observed in patients with PD-L1 expression levels <1% (17.2 vs 12.2 months; P value not reported).
Pasi A. Jänne, MD, PhD, agreed that the results were impressive. “I don’t think we see that with chemotherapy, even with maintenance chemotherapy,” he said. “It clearly opens up an option for patients to start with combination immunotherapy in the frontline setting as opposed to systemic combination chemotherapy.”
Byoung Chul Cho, MD, PhD, noted that combining PD-1 and CTLA-4 inhibition increases the risk and diversity of immune-related adverse events (IRAEs), making toxicity management especially important. Ramalingam said the dose of ipilimumab used in CheckMate 227 was lower than that used in the nivolumab/ipilimumab regimen for treating melanoma, which could improve tolerance. Cho and Jänne agreed that the dual ICI regimen is a reasonable alternative for patients with new diagnoses who wish to avoid cytotoxic chemotherapy.
CheckMate 817, which included a cohort of patients with worse performance status or prespecified comorbidities, found that these patients tolerated nivolumab/ipilimumab as well as healthier patients did but had a lower response rate and worse progression- free survival (PFS).5 Vansteenkiste said findings suggested that nivolumab/ipilimumab is a viable alternative to chemotherapy for sicker patients.
ICI Plus Chemotherapy
Both pembrolizumab and atezolizumab have first-line indications in combination with chemotherapy for metastatic nonsquamous NSCLC without EGFR or ALK alterations.2,3 Pembrolizumab is approved in combination with pemetrexed (Alimta) and platinum chemotherapy.2
In December 2018, atezolizumab gained a frontline indication for nonsquamous disease in combination with bevacizumab (Avastin), paclitaxel, and carboplatin. A year later, the FDA expanded the chemotherapy regimens available for use with atezolizumab in the frontline setting to include carboplatin plus nab-paclitaxel (Abraxane) based on an OS benefit in the phase III IMpower130 trial.6,7
Cho noted that a pembrolizumab/ chemotherapy regimen is also approved for frontline use in metastatic squamous NSCLC based on KEYNOTE-407 trial results, which showed significant improvement in OS and PFS versus placebo.8
At WCLC, Jotte and colleagues presented final OS results from the phase III IMpower131 trial, which compared atezolizumab plus carboplatin plus nab-paclitaxel versus chemotherapy alone in patients with advanced squamous NSCLC.9 Whereas KEYNOTE-407 and IMpower130 showed an OS benefit regardless of PD-L1 expression level, Vansteenkiste said that only patients with the highest PD-L1 expression levels derived an OS benefit from the atezolizumab/ chemotherapy combination.8,9
“At this point, we can’t just say ‘Use chemotherapy plus atezolizumab for all PD-L1—positive patients [with squamous NSCLC],’ ” Ramalingam said. He recommended chemotherapy plus pembrolizumab unless a patient’s PD-L1 expression level is very high. “That could be a consideration for using chemotherapy plus atezolizumab…or singleagent pembrolizumab,” he said.
Long-Term Data for ICI Monotherapy
Five-year outcomes from the phase III CheckMate 017 and 057 trials comparing nivolumab versus docetaxel in patients with previously treated squamous or nonsquamous NSCLC were presented at WCLC.10 Jänne said the 5-year OS rate was 13% with nivolumab versus 3% with docetaxel (HR, 0.68; 95% CI, 0.59-0.78),10 which translated to a 5-fold improvement in 5-year survival. The 5-year PFS rates were 8% versus 0%, respectively (HR, 0.79; 95% CI, 0.68-0.92).
Three-year survival data from the phase III KEYNOTE-024 trial were also presented at the conference.11 The study included patients with advanced NSCLC and a PD-L1 tumor proportion score (TPS) ≥50% who were randomly assigned to pembrolizumab or platinum-based chemotherapy.
The 3-year OS rate was 44% in the pembrolizumab arm versus 25% in the chemotherapy arm, and median OS was 26.3 versus 14.2 months, respectively (HR, 0.65; 95% CI, 0.50-0.86).11 “These are really meaningful numbers that we can use to tell our patients about the likelihood of being alive at different landmark points,” Jänne said.
Predictive Biomarkers of ICI Response
“Only a fraction of patients benefit from adding immunotherapy…and it’s very difficult to define that fraction,” Vansteenkiste said. Both PD-L1 expression and tumor mutational burden (TMB) are imperfect biomarkers of ICI efficacy in patients with NSCLC. He noted that in data presented at the ESMO 2019 Congress, “PD-L1 expression and TMB—in tissue or plasma—seem to fail to predict which patients will benefit from immunotherapy.”
Data also suggest no correlation between PD-L1 expression and TMB in a given tumor, according to Vansteenkiste.12 “There are patients who are TMB high and PD-L1 low, and the reverse is there, as well, so they’re not concordant at all,” he said. He added that they could be additive, however, so it may be necessary to “take both into account,” but more clarity is needed.
Jänne explained the rationale for using TMB as a predictive biomarker: Cancer cell—specific mutations express unique antigens that bind to human leukocyte antigen receptors on the cell surface for presentation as neo-epitopes to immune cells.1
Tumors with more mutations produce more neoantigens and thus have a greater chance of producing neo-epitopes recognized and targeted by immune cells as nonself. He said that with checkpoint inhibition, which unleashes cytotoxic immune cells, “you sort of have more chances for the immune system to go after these neoepitopes.” He added that lack of standard thresholds for high TMB, variations in genetic sequencing approaches, and the time needed for sequencing limits TMB’s usefulness as a biomarker.
Investigators for the KEYNOTE-021, 189, and 407 trials explored the relationship between tissue TMB and efficacy of pembrolizumab plus platinum-based chemotherapy in newly diagnosed squamous or nonsquamous NSCLC.13 The threshold between high versus low TMB was 175 mutations. They analyzed OS and PFS by TMB status in KEYNOTE-187 and 407 and objective response rate (ORR) in KEYNOTE-021.13 “These studies showed no difference in outcomes for the chemotherapy plus pembrolizumab as a function of TMB,” Ramalingam said.
In contrast, Ramalingam said, an exploratory analysis of the KEYNOTE-010 and 042 trials, which compared pembrolizumab monotherapy versus chemotherapy in previously treated or treatment-naïve patients with PD-L1—positive advanced NSCLC, did reveal a significant correlation between tissue TMB and response to pembrolizumab.14 Overall, a TMB ≥175 was associated with greater improvement in OS, PFS, and ORR than a TMB <175 in the pembrolizumab arm, but TMB did not correlate with outcomes in the chemotherapy arms of both trials.14
Ramalingam said the findings align with results published in 2018 from the CheckMate 227 study, which observed significantly longer PFS among patients with a high TMB who received nivolumab/ ipilimumab versus those treated with chemotherapy.15 Vansteenkiste noted that the trial’s recently presented OS data found no association between OS and TMB, however.4
Cho said data from the phase III MYSTIC trial, presented earlier in the year, showed “very good correlation” between bloodbased and tissue-based tests for TMB in patients with metastatic NSCLC.16 He said a high bloodbased TMB level was “predictive for PFS, as well as OS, to a durvalumab [Imfinzi] and tremelimumab combination versus chemotherapy.”16,17 The blood-based TMB cutoff used in MYSTIC was ≥20 mutations per megabase.
Cho and Jänne agreed that a lot of controversy surrounds which cutoffs to use for blood-based versus tissue-based TMB tests and how to make cross-platform comparisons. “In the future, we need to elaborate bloodor tumor-based TMB for development of immunotherapy combinations,” Cho said.
Vansteenkiste said the importance of PD-L1 expression levels also seems to diminish when an ICI is combined with chemotherapy. “Chemotherapy plus immunotherapy is superior to chemotherapy alone in all PD-L1 subsets, with some variation according to expression,” he explained. The FDA initially granted pembrolizumab monotherapy a frontline indication for patients with metastatic NSCLC (with no ALK or EGFR alteration) and a PD-L1 TPS ≥50%, which was expanded in April 2019 to include patients who have a PD-L1 TPS ≥1%.18
Vansteenkiste suggested that different methods of measuring PD-L1 can also make it difficult to interpret study results. He mentioned that the IMpower investigators used “combined tumor cell [TC] and immune cell [IC] score by the SP142 antibody, which is perhaps not the best antibody” to measure PD-L1 levels. Jänne agreed, noting that US oncologists rarely screen patients with NSCLC for TC/IC scores during a diagnostic workup and are more likely to obtain PD-L1 TPS. Cho said the situation was similar in South Korea, where he practices.
The phase III IMpower110 trial compared atezolizumab monotherapy versus a platinum-based chemotherapy doublet as firstline treatment for PD-L1—positive metastatic NSCLC regardless of histology.19
Results were analyzed according to PD-L1 levels, which were defined as TC3 or IC3 (TC ≥50% or IC ≥10% PD-L1+), TC2/3 or IC2/3 (TC ≥5% or IC ≥5% PD-L1+), and TC1/2/3 or IC1/2/3 (TC ≥1% or IC ≥1% PD-L1+). Vansteenkiste said interim data showed patients with TC3 or IC3 status had significantly prolonged OS when treated with atezolizumab versus chemotherapy (20.2 vs 13.1 months, respectively; HR, 0.59; P = .0106).19 For patients with TC2/3 or IC2/3, he said, the difference in OS between treatments was borderline, and OS data were not yet analyzed for patients with TC1/2/3 or IC1/2/3.
Vansteenkiste said combining the secondand third-tier PD-L1 groups makes it difficult to interpret the data, but they seem to suggest “atezolizumab as a single-agent is a valid option in patients who have TC3 or IC3 scoring.” Jänne said he would like to know whether the benefit is the same in TC3 and IC3 patients or whether they are separate populations. Vansteenkiste said there are plans to reanalyze the data using other assays, which he expects to provide more detailed information.