Waiting to exhale as three distinct checkpoint combinations read out in first-line NSCLC at AACR

by Megan Epperson | Apr 17, 2018

Checkpoint immunotherapy entered the NSCLC space in late 2015, and since that time, this tumor type has been a flurry of activity, filings, and failures that continue to push our understanding of the abilities and boundaries of immunotherapy. Currently, three checkpoint inhibitors now have approvals for NSCLC patients in the relapsed / refractory setting: Keytruda® (pembrolizumab, Merck & Co / MSD), Opdivo® (nivolumab, Bristol-Myers Squibb/Ono Pharmaceuticals), and Tecentriq® (atezolizumab, Genentech / Roche / Chugai), with Keytruda having a PD-1 positivity requirement associated with the approval. The next frontier is the first-line setting, and both Keytruda and Opdivo initially vied for approval as monotherapies in this setting.  Keytruda succeeded in its Phase III KEYNOTE-024 trial by demonstrating significant progression-free survival (PFS) and overall survival (OS) improvements as monotherapy in patients who are PD-L1 positive (Tumor Proportion Score (TPS) ≥ 50%) in all histologies1.  With data reading out at the same time, Opdivo was unable to meet their primary endpoint in a similar subset of patients in their CheckMate 026 trial, using a cutoff of PD-L1 positivity at ≥5% in their own assay2. These contradistinctive results have been dissected and discussed and ultimately lead to an approval in first-line NSCLC for Keytruda, and an important dialogue about the utility and limitations of PD-L1 as a biomarker.

Keytruda is now utilized in approximately 50% of PD-L1+ (≥ 50%) patients in the first-line setting as a monotherapy regardless of histology3. It was also recently announced that the Phase 3 KEYNOTE-042 trial assessing Keytruda monotherapy in patients with a PD-L1 TPS of ≥1 percent has met its primary endpoint of overall survival4. This is likely to lead to a label expansion for Keytruda monotherapy to include >1% PD-L1 positive patients, capturing more of this market.

Additionally, in May 2017 Keytruda gained FDA accelerated approval for use in first-line in combination with Alimta® (pemetrexed, Eli Lilly) and carboplatin based on the results from the randomized cohort G from the Phase II KEYNOTE-021 trial in non-squamous NSCLC patients, irrespective of PD-L15. This approval marked the first immunotherapy/chemotherapy combination to receive FDA approval in any tumor type. It was previously speculated that the one-two punch of using cytotoxic chemotherapy to release tumor antigens into the microenvironment combined with the re-invigorated T cell responses brought on by PD-1 checkpoint therapy could be beneficial to patients, but KEYNOTE-021 showed that it was indeed a possible marriage. While exciting, the small sample size of this study left many physicians and other global regulators seeking confirmation of activity from a randomized Phase III trial. The KEYNOTE-189 trial evaluating Keytruda in combination with platinum/Alimta would hopefully serve to alleviate any residual doubts about the immunotherapy/chemotherapy combination, and support full approval for patients with non-squamous NSCLC in the first-line setting.

Although Keytruda has, thus far, enjoyed approvals in the first-line setting uncontested by any other checkpoint inhibitor monotherapy or combination, this sovereignty could soon be contested. In the Clinical Trials Plenary session at the 2018 annual meeting of the American Association of Cancer Research (AACR), three pivotal trials investigating checkpoint therapy as part of distinct combination strategies in first-line metastatic NSCLC presented results: KEYNOTE-189, IMpower 150, and CheckMate 227. KEYNOTE-189, as introduced above, showed results using a checkpoint inhibitor/chemotherapy combination strategy. IMpower 150 detailed results of Tecentriq in combination with both chemotherapy and targeted agent, Avastin® (bevacizumab, Roche / Genentech / Chugai). CheckMate 227 provided initial results of Opdivo in combination with CTLA4-inhibitor, Yervoy® (ipilimumab, Bristol-Myers Squibb / Ono Pharmaceuticals). The results of these trials have been awaited with baited breath as it was hoped the data would shed light on what the best combination strategy will be for immunotherapy combinations in NSCLC. The AACR NSCLC showdown began with the results of KEYNOTE-189 presented in the first talk of the session.

The presentation of KEYNOTE-189 (NCT02578680), an international randomized (2:1) Phase III trial of 616 patients testing Alimta in combination with platinum chemotherapy (carboplatin or cisplatin) with or without Keytruda in first-line non-squamous, metastatic NSCLC was anticipated to show positive results, consistent with the efficacy data shown for KEYNOTE-021 cohort G that garnered this combination’s accelerated approval. Without disappointment, the results from KEYNOTE-189 showed that the addition of Keytruda in this combination indeed met both co-primary endpoints of PFS and OS, regardless of PD-L1 status6.  The addition of Keytruda to Alimta and platinum chemotherapy showed an overall response rate (ORR) of 47.6% versus 18.9% in the control arm. The triplet therapy also significantly improved PFS (8.8 mos versus 4.9 mos; HR=0.52, p<0.00001). A significant improvement in the co-primary endpoint of OS was also observed with the addition of Keytruda to chemotherapy showing that the median OS was not yet reached in the experimental arm versus 11.3 months in the control arm (HR 0.49, p<0.00001); at one-year, 69% of patients in the Keytruda arm were still alive compared to 49% in the control arm. This OS benefit was observed even with 41% of the patients in the Keytruda arm having crossed over to receive pembrolizumab or another checkpoint inhibitor post-study, which argues in favor of using Keytruda in first-line rather than delaying its use until later lines.  Importantly, the Keytruda combination arm showed a significant efficacy benefit across all subgroups regardless of PD-L1 status. Interestingly, for PFS, the lowest PD-L1 cutoff (TPS<1%) showed a less-than-robust HR of 0.75, but while it was of some concern, it was not thought to be a major limitation of the study, especially considering that the OS data showed a strong HR of 0.59 in TPS<1%.

The toxicity data showed the Keytruda/chemotherapy combination to be relatively well-tolerated. In the experimental arm, 99.8% of patients had treatment-related adverse events (TRAEs) regardless of grade versus 99.0% for the chemotherapy only arm. (Grade 3-5 adverse events 67.2% versus 65.8%). The only adverse events that appeared slightly more commonly in the Keytruda arm were diarrhea (~30% versus 20% all grades) and rash (~20% versus ~10% all grades), and immune-related AEs occurred in 23% of patients receiving the Keytruda combination versus 12% in the control arm. Specifically called out, were the renal adverse events associated with the Keytruda/chemotherapy combination, with acute kidney injury occurring in 5% of patients versus 0.5% in the chemotherapy only arm; it was noted that these events were mostly low-grade.  The rates of adverse event-related treatment discontinuation were low overall, although slightly higher in the Keytruda arm (13.8% versus 7.9%).

All told, this data from KEYNONTE-189 showed an efficacy benefit with the addition Keytruda to chemotherapy with robust hazard ratios, including a demonstrated benefit in overall survival, an important endpoint for NSCLC physicians and patients. These results will almost certainly alleviate any remaining concern surrounding the accelerated FDA approval based on the Phase 2 trial, promote the readiness to prescribe this triplet to first-line patients, and support global regulatory filings outside of the U.S.

KEYNOTE-189 was not, however, the only trial in first-line NSCLC to strike a chord at AACR. The Keytruda presentation was immediately followed by data from the Phase III IMpower 150 trial, detailing the results of the quadruplet therapy of Tecentriq in combination with Avastin, carboplatin, and paclitaxel in non-squamous first-line NSCLC. The IMpower 150 trial (NCT02366143) is a three-arm study that compared Tecentriq plus carboplatin and paclitaxel versus Tecentriq plus Avastin and carboplatin and paclitaxel versus Avastin plus carboplatin and paclitaxel as first-line therapy for advanced or metastatic non-squamous NSCLC. The trial initiated in March 2015 and the first data from Impower150 read out in the Fall of 2017. Interestingly, the IMpower 150 trial was set up to accept ALK and EGFR mutant patients, as well as to evaluate multiple strategies for patient segmentation, including a newer biomarker population call the T-effector (Teff) gene signature (mRNA expression of three genes: PD-L1, CXCL9, and IFNg). Additionally, as secondary endpoints, this trial examined the role of PD-L1 levels by IHC staining on efficacy. Primary endpoints included PFS and OS in the intent to treat (ITT) population of non-mutated patients (WT) (No ALK or EGFR mutations), and PFS in the Teff-high non-mutated population (WT). Previously reported data showed that in patients who received Tecentriq plus Avastin and chemotherapy (n=356) the median PFS was 8.3 mos compared to 6.8 mos in patients who received Avastin plus chemotherapy (n=336) (HR 0.62; p<0.0001)7. For Teff high patients (n=284) the mPFS was 11.3 mos versus 6.8 mos in the Tecentriq plus Avastin and chemotherapy arm versus the Avastin plus chemotherapy arm (HR 0.51).  The Teff low patients (n=374) showed a mPFS of 7.3 mos for the Tecentriq plus Avastin and chemotherapy versus 7.0 mos for Avastin plus chemotherapy (HR 0.76)8. Data presented Monday at AACR showed a PFS benefit across PD-L1 subtypes as assessed by the Ventana SP263 or SP142 IHC assays, although the strongest data was seen for PD-L1 high patients (IC3 or TC3, HR 0.39) versus the PD-L1 negative population (IC0 or TC0, HR 0.77). Interestingly, data presented assessing PFS in ALK/EGFR mutants (n=108) showed significant benefit for this patient population. This patient subgroup showed a mPFS of 9.7 mos in the Tecentriq plus Avastin and chemotherapy arm versus 6.1 mos in the Avastin plus chemotherapy arm (HR 0.59); this benefit was also observed despite lower PD-L1 expression in this population. One last subgroup worth mentioning for IMpower 150 would be patients with confirmed liver metastases (n=110), who showed a significant benefit with the quadruplet therapy over Avastin plus chemotherapy alone, with a mPFS of 8.2 mos versus 5.4 mos (HR 0.40). No additional adverse event data was presented, but previously reported data show that treatment-related Grade 3-5 AEs occurred in 59% of patients receiving Tecentriq plus Avastin and chemotherapy recipients, compared to 50% of the patients receiving Avastin plus chemotherapy9.  In summary, while the IMpower 150 data showed increased PFS with the quadruplet therapy (both overall as well as across multiple patient subsets), excitement for this trial is somewhat curbed by the lack of mature OS data in the presentation; although it was stated that the OS endpoint was met positively in March 2018, no data on this endpoint was reported at AACR (presumably this will be on the docket for ASCO in June 2018). The lack of reported OS results, and the lack of PFS benefit that was previously reported between Arm A (Tecentriq + platinum doublet) and Arm C (Avastin + platinum doublet)10 create residual uncertainty about how this combination will fit in first-line NSCLC and how well it will be able to compete. While it appears there may be a place for this combination, especially in certain patient subtypes, it is not yet clear how well, if at all, Tecentriq plus Avastin and chemotherapy will be able to compete with Keytruda plus Alimta / platinum in first-line non-squamous histology.

Not to be left out, CheckMate 227 results were presented following the data from IMpower 150. CheckMate 227 is an international, four-arm Phase III trial (NCT02477826), with Part 1 evaluating Opdivo in combination with Yervoy versus Opdivo monotherapy versus platinum-doublet chemotherapy in first-line NSCLC. This trial recruited both squamous and non-squamous histologies. Originally, the trial had two main tracks: one for PD-L1 positive patients, and the other for PD-L1 negative patients; however, the study was amended prior to initial analysis to include PFS in tumor mutational burden (TMB) subsets, as assessed by the FoundationOne CDx assay [a 10 mutation/mega base cutoff was used to define “TMB-high” as a co-primary endpoint patient cohort]; 44% of randomized evaluable patients were found to be TMB-high by this definition, although only 58% of all randomized patients were evaluable by this assay (reasons for being non-evaluable included insufficient tissue quantity or quality11). Data presented at AACR for CheckMate 227 showed an efficacy benefit for patients with high TMB treated with Opdivo in combination with Yervoy (n=139) versus those treated with chemotherapy appropriate for their histology (n=160)11. The ORR for TMB-high patients treated with Opdivo plus Yervoy was 45.3% versus 26.9% for TMB-high chemotherapy-treated patients. There was also an improved mPFS in the Opdivo plus Yervoy treated TMB-high patients with compared to the same patients treated with chemotherapy alone (7.2 mos vs 5.4 mos; HR 0.58, p=0.0002); at one-year, 43% of TMB-high patients in the Opdivo + Yervoy arm were alive and progression-free compared to 13% in the control arm. It was noted that the increased PFS benefit was seen across histologies, as well as across PD-L1 expression subsets, driving home the point that TMB was functioning as an independent biomarker.  There was no PFS benefit in patients who were TMB-low (<10 mut/Mb; HR 1.07).  While overall survival was immature, BMS presented initial data showing a trend toward improved overall survival for the Opdivo plus Yervoy-treated patients who were TMB high, with a median OS of 23.0 mos versus 16.4 mos for the chemotherapy-treated TMB-high patients (HR 0.79).

The safety report for Opdivo in combination with Yervoy showed a similar rate of Grade 3/4 AEs in the combination arm versus the chemotherapy arm in all patients (31% versus 36%), and treatment-related adverse events led to treatment discontinuations in 17% of patients in the combination arm versus 9% in the chemotherapy arm. The most common individual toxicities observed more frequently in the Opdivo + Yervoy arm versus chemotherapy arm were all grades rash (17% versus 5%), diarrhea (16% versus 10%), and pruritis (14% versus 1%). 

The results of this trial were highly anticipated, and “Tumor Mutational Burden” was the buzz-word of the week at AACR across many different sessions and talks. The data shown are promising, but immature. Still more questions seemed to arise from this talk than were answered, and how the Opdivo plus Yervoy combination will compete with Keytruda + Alimta/platinum is still not answered.  One could envision that a patient who is PD-L1 negative but TMB high could have an option for first-line treatment that spares chemotherapy until the second-line, but is there inherent value in this option if this non-chemotherapy option still retains similar levels of adverse events and leads to more treatment discontinuations?  A key question remaining is whether these results from CheckMate 227 can support regulatory approval.  Filing and approval may need to wait until OS data is mature and demonstrates a positive benefit.  At worst the regulatory authorities will consider the use of TMB for analysis of trial results as retrospective in nature due to the change in protocol mid-study, which may necessitate a new confirmatory study.

Cross comparisons of these three important trials are quite difficult given the designs of each, and several key differences should be noted.  First, patient crossover was allowed in the KEYNOTE-189 trial but was not allowed in IMpower 150 or CheckMate 227.  Second, EGFR and ALK mutant patients were enrolled in the IMpower 150 trial, and results show a clear benefit for the quadruplet combination in these patients who were excluded from the other two trials. Third, the CheckMate 227 trial results segmented patients according to tumor mutational burden, and there were no correlations with PD-L1 expression, making comparisons with the other trial data difficult.  Even with these vast differences, it is easy to jump to the simplest analysis, which is that the largest magnitude of OS benefit (measured as Hazard ratio between two arms) reported in the broadest patient population (non-squamous histology, biomarker unselected) belongs to Keytruda + platinum/Alimta (HR 0.49).  In terms of immediate application to clinical decisions, it is clear that KEYNOTE-189 came out singing. Any concerns surrounding the accelerated approval based on a small cohort of patients should now be quelled and this data can immediately translate into clinical practice. Discussant leader Dr. Roy Herbst says that the answer to the question of ‘Will this regimen become the new standard of care?’ should be “An absolute yes.” This news will, undoubtedly, keep Keytruda on top in first-line NSCLC, albeit limited to non-squamous histology for now (efficacy in squamous histology awaits the results of other ongoing trials).

We would be remiss, however, to discount the important data gleaned from both IMpower 150 and CheckMate 227. For starters, no checkpoint monotherapy has yet demonstrated a meaningful efficacy benefit in patients with ALK or EGFR mutations. It is also known that PD-L1 and even mutational burden do not appear to be enriched in these mutant patient populations12,13. The fact that a strong PFS benefit was observed  in this group of patients in the IMpower 150 study receiving Tecentriq in combination with Avastin and chemotherapy merits further research, and suggests that there could be a role for checkpoint inhibition in earlier lines of therapies for these patients.  However, the magnitude of PFS benefit with this regimen still lags far behind the levels achieved with first-line tyrosine kinase inhibitors, so in the absence of a demonstrated OS benefit we may not see checkpoint inhibition with the Tecentriq quadruplet regimen move earlier than second-line. They also presented data showing that this quadruplet combination is particularly robust in patients with liver metastasis, a patient group with a historically poor prognosis. The addition of Tecentriq to Avastin and chemotherapy significantly enhanced responses and thus could represent a good therapy options for these patients in the near future.

Finally, the CheckMate 227 data has shown that tumor mutational burden is a distinct and separate biomarker from PD-L1 that can be used predictively in treatment decisions about immunotherapy. This is important information that will continue to inform the biomarker debate, and enhance the opportunity to incorporate new biomarkers and gene signatures into treatment decisions to provide the most benefit possible to patients. While the data for PFS was strong for CheckMate 227’s Opdivo + Yervoy in TMB-high patients, overall survival data is still immature. The discussant leader, Dr. Naiyer Rizvi, expressed both intrigue as well as concern for TMB as a potential biomarker. While there is a trend toward OS benefit significance, Dr. Rizvi questioned whether tumors with high mutational burden may initially show response, but subsequently escape or become resistant to therapy.  Additionally, questions surrounding the feasibility of labs to accurately test for TMB are likely to go unanswered for the moment. On one hand, FoundationOne CDx diagnostic was successfully used in CheckMate 227, has FDA approval, and a preliminary National Coverage Determination from the Centers for Medicare and Medicaid Services. BMS also announced a partnership with Illumina to create and commercialize a companion diagnostic for TMB. This is a good set up for TMB to become a viable and reliable lab test. On the other hand, many hospitals still employ their own testing methods, some of which are more or less accurate than FoundationOne. This leaves more questions surrounding the issue of potential cross trial or company standardization for TMB in the future, which is something that PD-L1 struggles with as a potential biomarker currently without a great answer. As a final thought, the discussant leader Dr. David Rimm noted that TMB testing costs 5-10 times as much as standard IHC testing, and requires almost 10 times as much tumor sample. While questions remain surrounding the role of TMB as a predictive biomarker, and what the final results of the CheckMate 227 trial will ultimately show, it does appear that TMB represents a valuable biomarker that, while perhaps not quite ready for prime time, will be an important feature to examine in future trials and perhaps eventually will help discern treatment options for certain patient subsets.   

While many held their breath at AACR this year, looking for details from these three pivotal trials that would clarify their respective roles and places in first-line NSCLC, it seems we must continue to hold our breath for more clarification from IMpower 150 and CheckMate 227 over the next year, while Keytruda, for now at least, can breathe a sigh of relief.





  1. Reck M, Rodríguez-Abreu D, Robinson AG, et al.; “Pembrolizumab versus Chemotherapy for PD-L1-Positive Non-Small-Cell Lung Cancer;” N Engl J Med, 375(19):1823-1833, 2016.
  2. Carbone DP, Reck M, Paz-Ares L, et al.; “First-Line Nivolumab in Stage IV or Recurrent Non–Small-Cell Lung Cancer;” N Engl J Med, 376(25):2415-2426, 2017.
  3. Kantar Health, CancerMPact® Treatment Architecture, accessed April, 2018. Available at www.cancermpact.com.
  4. Merck Press Release: “KEYTRUDA® (pembrolizumab) Monotherapy Met Primary Endpoint in Phase 3 KEYNOTE-042 Study, Significantly Improving OS as First-Line Therapy in Locally Advanced or Metastatic NSCLC Patients Expressing PD-L1 in at Least 1 Percent of Tumor Cells”; APRIL 09, 2018.
  5. Borghaei H, Langer CJ, Gadgeel S, et al.; “Updated results from KEYNOTE-021 cohort G: A randomized, phase 2 study of pemetrexed and carboplatin (PC) with or without pembrolizumab (pembro) as first-line therapy for advanced nonsquamous NSCLC;” Ann Oncol, Volume 28, Issue suppl_5, Abstract LBA49, 2017.
  6. Gandhi L, Rodriguez-Abreu D, Gadgeel S, et al.; “Pembrolizumab plus chemotherapy in metastatic non small-cell lung cancer;” N Engl J Med, [published online April 16, 2018], doi: 10.1056/NEJMoa1801005. AACR Abstract CT075, 2018.
  7. Reck M, Socinski MA, Cappuzzo F, et al.; “Primary PFS and safety analyses of a randomized phase III study of carboplatin + paclitaxel +/− bevacizumab, with or without atezolizumab in 1L non-squamous metastatic NSCLC (IMPOWER150);” Annals of Oncology, 28(11), 2017
  8. Kowanetz M, Zou W, McCleland M, et al.; “Pre-Existing Immunity Measured by Teff Gene Expression in Tumor Tissue is Associated with Atezolizumab Efficacy in NSCLC;” Proc WCLC, Abs MA05.09, 2017.
  9. Kowanetz M, et al.; “IMpower150: PFS results for atezolizumab plus bevacizumab and chemotherapy across PD-L1 expression subgroups defined by the SP142 and SP263 IHC assays confirm all-comer benefit in 1L metastatic NSCLC;” AACR 2018, Abstract CT076. 
  10. Reck M, Socinski MA, Cappuzzo F, et al.; “Primary PFS and safety analyses of a randomized Phase III study of carboplatin + paclitaxel +/- bevacizumab, with or without atezolizumab in 1L non-squamous metastatic NSCLC (IMpower 150);” Annals of Oncology, 28(suppl 11, abstract LBA1_PR), 2017.
  11. Hellman MD, Ciuleanu T-E, Pluzanski A, et al.; “Nivolumab plus Ipilimumab in Lung Cancer with a High Tumor Mutational Burden;” N Engl J Med, published online April 16, 2018; doi: 10.1056/NEJMoa1801946.
  12. Gainor JF, Shaw AT, Sequist LV, et al.; “EGFR Mutations and ALK Rearrangements Are Associated with Low Response Rates to PD-1 Pathway Blockade in Non-Small Cell Lung Cancer (NSCLC): A Retrospective Analysis;” Clin Cancer Res, 22(18):4585-93, 2016.
  13. Nahar R, Zhai W, Zhang T, et al.; “Elucidating the genomic architecture of Asian EGFR-mutant lung adenocarcinoma through multi-region exome sequencing;” Nature Comm, 9(1):216, 2018.

Leave a comment

More than one Google Analytics scripts are registered. Please verify your pages and templates.