Thank you for your interest in CCO content. As a guest, please complete the following information fields. These data help ensure our continued delivery of impactful education.
Become a member (or login)? Member benefits include accreditation certificates, downloadable slides, and decision support tools.
In this module, Patrick Forde, MBBCh, reviews the latest data on incorporating immune checkpoint inhibitors into the neoadjuvant and adjuvant settings for the treatment of early-stage non-small-cell lung cancer (NSCLC) from the point of view of a medical oncologist.
The key points discussed in this module are illustrated with thumbnails from an accompanying downloadable PowerPoint slideset that can be found here or downloaded by clicking any of the slide thumbnails in this module alongside the expert commentary.
Clinical Care Options plans to measure the educational impact of this activity. Two questions will be asked twice: once at the beginning of the activity and then once again after the discussion that informs the best choice. Your responses will be aggregated for analysis, and your specific responses will not be shared.Before continuing with this educational activity, please take a moment to answer the following questions.
The treatment algorithm for advanced NSCLC has evolved dramatically during the past several years. The FDA has approved multiple new targeted agents for patients with actionable mutations or alterations, and testing of at least 7 genes is now recommended for patients with newly diagnosed NSCLC.1
For patients whose tumors do not harbor an actionable alteration, immune checkpoint inhibitor–based therapy has become standard of care.1 To determine the range of immunotherapy options suitable for these patients, it is essential to perform immunohistochemistry to assess PD‑L1 expression in the tumor. Patients with the highest PD‑L1 expression (PD-L1 ≥50%) may qualify for treatment with anti–PD-1 or anti–PD-L1 antibody monotherapy or in combination with chemotherapy. Those with intermediate PD‑L1 expression (PD-L1 1%-49%) or with PD-L1–negative (PD-L1 <1%) tumors would be optimally treated with combination therapy.The advent of immune checkpoint inhibitors such as pembrolizumab, nivolumab, atezolizumab, and cemiplimab—with even more now emerging—has improved survival in our patients with advanced NSCLC, with the median OS now ranging from approximately 14 months to more than 2 years for patients with no actionable alterations.2 By contrast, we have had relatively few advances in the management of resectable early‑stage disease until the recent emergence of studies evaluating targeted therapy and immunotherapy in the neoadjuvant and adjuvant settings, which will be the focus of our discussion.
Chemotherapy is currently the predominant perioperative systemic therapy used in the treatment of resectable early-stage NSCLC and, according to the National Comprehensive Cancer Network guidelines, may be given either preoperatively in the neoadjuvant setting or postoperatively in the adjuvant setting.1 In the diagram here, the red outline denotes patients for whom perioperative systemic therapy currently has a management role.
Furthermore, before we discuss emerging clinical trial data, it is important to mention the different editions of the American Joint Committee on Cancer (AJCC) staging manual, which was updated in 2017 to the 8th edition.3,4 A key difference between the 7th and 8th editions is that patients with NSCLC with what the 7th edition categorized as stage IB tumors are now categorized as having stage II disease if their tumors are larger than 4 cm.5 Which edition was used for staging can affect trial eligibility and applicability of results for some patients.
A major paper influencing management of early-stage NSCLC was published in 2008. Using data from 5 large, randomized clinical trials, the LACE meta-analysis assessed the benefit of adding adjuvant cisplatin-based chemotherapy to surgery in 4584 patients with stage I-III disease who were enrolled from 1995-2001.6 The meta-analysis found that the addition of adjuvant cisplatin-based chemotherapy to surgery led to a 5.4% improvement in OS at 5 years. Despite the OS benefit with cisplatin chemotherapy, there is still an obvious need for improved therapeutic options for our patients with early-stage disease.
Of note, adjuvant cisplatin was associated with a 6.9% decrease in cancer‑specific death and a 1.4% increase in noncancer‑related death. This finding reflects the study population, who tended to be older and with a greater burden of comorbidities.
A key advance in the treatment of resectable early-stage NSCLC was the ADAURA trial.7 This phase III trial randomized 682 patients with completely resected stage IB-IIIA (AJCC 7th edition) nonsquamous NSCLC harboring EGFR mutations to either osimertinib, an oral third-generation EGFR tyrosine kinase inhibitor, or placebo for 3 years following surgery with or without adjuvant chemotherapy. The primary endpoint was investigator-assessed DFS in patients with stage II-IIIA tumors.
ADAURA met is primary endpoint and demonstrated significantly improved DFS with adjuvant osimertinib for patients with stage II-IIIA NSCLC and EGFR sensitizing mutations.7 Median DFS was not reached in the osimertinib arm vs 19.6 months in the placebo arm (HR: 0.17; 95% CI: 0.11-0.26; P <.001). These positive results led to the FDA expanding the indications for osimertinib to include adjuvant treatment following tumor resection in adults with early-stage NSCLC and EGFR exon 19 deletions or exon 21 L858R mutations.8 We continue to await OS data from this trial.
These results may apply to the approximately 15% of White patients and almost 50% of Asian patients with nonsquamous NSCLC harboring these EGFR mutations.9 However, for patients without these mutations, for whom the standard option remains perioperative chemotherapy, there is a need for new options.
Shown here are some of the considerations when deciding whether to use neoadjuvant or adjuvant therapy for a patient with early-stage NSCLC.
First, neoadjuvant therapy delivered prior to resection provides the earliest opportunity to eradicate micrometastatic disease.10 Second, more patients both receive and complete neoadjuvant therapy compared with those undergoing adjuvant therapy. In a phase III trial comparing outcomes in patients with early-stage NSCLC undergoing neoadjuvant vs adjuvant therapy, 97.0% of patients who planned to receive neoadjuvant therapy actually commenced treatment vs 66.2% of those who planned to receive adjuvant therapy.11 Furthermore, 90.4% vs 60.9% of patients completed the planned course of neoadjuvant or adjuvant therapy, respectively. Finally, assessment of the pathologic response or degree of pathologic progression with neoadjuvant therapy gives us an early indicator as to the efficacy of systemic therapy.12
On the other hand, adjuvant therapy delivered after resection allows for the fastest time to curative surgery, along with avoiding any presurgical complications from systemic therapy that may preclude surgery altogether.13 Furthermore, adjuvant therapy can be delivered for a longer period of time—which may have particular relevance for immunotherapy, as we will discuss later.14
In 2021, there have been some major developments in the use of immunotherapy in the neoadjuvant and adjuvant settings for the treatment of resectable early-stage NSCLC. At the 2021 American Association for Cancer Research Annual Meeting, my colleagues and I15 presented results from the phase III CheckMate 816 trial that compared the addition of the anti–PD-1 antibody nivolumab to neoadjuvant chemotherapy vs chemotherapy alone in 358 patients with resectable early-stage NSCLC.
Patients in this study were required to have newly diagnosed, resectable NSCLC that was stage IB (≥4 cm), II, or IIIA (AJCC 7th edition), an Eastern Cooperative Oncology Group performance status of 0/1, and no known sensitizing EGFR mutations or ALK alterations. They were stratified by stage, PD‑L1 status, and sex, and randomized to either neoadjuvant nivolumab plus chemotherapy or chemotherapy alone. Those in the nivolumab arm received 3 cycles of standard platinum-doublet chemotherapy, and those in the control arm also received 3 cycles of the same platinum-doublet chemotherapy, with squamous disease having 2 additional options (cisplatin with vinorelbine or docetaxel). Patients were recommended to undergo surgery within 6 weeks of the last dose of chemotherapy, and postoperative chemotherapy with or without radiotherapy was permitted per investigator’s decision. No further immunotherapy was delivered in the postoperative setting.
The coprimary endpoints were pathologic complete response (pCR) and event‑free survival (EFS), both determined by blinded independent review. At the 2021 meeting, I presented the results for pCR along with some correlative endpoints.
Pathologic assessment was performed by blinded independent pathologic review based on immune‑related pathologic criteria described by Cottrell and colleagues16 in the Annals of Oncology in 2018. These criteria defined pCR as 0% residual viable TC in both the primary tumor and sampled lymph nodes. Major pathologic response was defined as ≤10% residual viable TC in both the primary tumor and the sampled lymph nodes.
In CheckMate 816, we assessed several biomarkers, including PD‑L1 and TMB.15 TMB was defined as high if there were ≥12.3 mutations/Mb per the Illumina TSO500 assay—a cutoff correlating with 10 mutations/Mb per the FoundationOne assay.17 We also analyzed circulating tumor DNA (ctDNA) in plasma samples collected on Day 1 before each of the 3 treatment cycles using a tumor‑guided personalized ctDNA panel (ArcherDX Personalized Cancer Monitoring).15 Clearance of ctDNA was defined as a change from detectable levels at cycle 1 to undetectable at cycle 3.
At baseline, patient characteristics were relatively well balanced between arms.15 The median age was 64-65 years, and 63% to 64% of patients had stage IIIA disease. There was an almost even balance between squamous and nonsquamous histology in both arms, and approximately 50% of patients had tumor PD‑L1 expression ≥1%.
Turning now to patient disposition, all 3 cycles of chemotherapy were completed by 94% of patients in the nivolumab arm vs 85% in the control arm.15 More patients in the nivolumab arm underwent definitive surgery (83% vs 75% in the control arm). Of note, more patients were able to have lung‑sparing surgery in the nivolumab arm (77% vs 61%), and fewer patients had pneumonectomies (17% vs 25%).
Definitive surgery was canceled for 16% of patients in the nivolumab arm vs 21% in the control arm. Approximately one half of these cancellations were due to disease progression or adverse events (AEs), whereas other cancellations were due to patient decision based on response to therapy, patients choosing chemoradiotherapy over surgery, and, in some cases, patients being determined to not have resectable disease after enrollment.
There was a significant improvement in the coprimary endpoint of pCR with the addition of nivolumab to neoadjuvant chemotherapy.15 In the intention-to-treat (ITT) population, the pCR rate improved from 2.2% with chemotherapy alone to 24.0% with nivolumab plus chemotherapy, translating to a statistically significant odds ratio of 13.94 (99% CI: 3.49-55.75; P <.0001). In a more restricted population of patients who underwent resection, the pCR rate also improved from 3.2% to 30.5%. Comparable results were obtained when we examined only the primary tumor in the ITT population, which showed a pCR rate of 2.8% for chemotherapy vs 25.7% for nivolumab plus chemotherapy.
The improvement in pCR rate with the addition of nivolumab to neoadjuvant chemotherapy appeared to be consistent across all analyzed subgroups, including PD-L1 expression level, stage, or histology.
The overall response rate increased from 37% among those receiving chemotherapy alone to 54% with nivolumab plus chemotherapy.15 Slightly more patients had radiographic downstaging with the addition of nivolumab to chemotherapy (31% vs 24% with chemotherapy alone).
The most common treatment-related AEs occurred at a similar rate between the treatment arms.15 There did not appear to be a significant increase in AEs with the addition of nivolumab to chemotherapy. Indeed, most of these common treatment-related AEs were predominantly AEs related to chemotherapy.There were no unexpected immune‑related AEs with the addition of nivolumab to chemotherapy, and those that did occur were low grade. Of note, immune‑mediated pneumonitis occurred in only 2 patients in the nivolumab arm, and both had low‑grade events.
In summary, the CheckMate 816 trial showed a significant improvement in pCR—one of its coprimary endpoints—with the addition of nivolumab to neoadjuvant chemotherapy vs chemotherapy alone in patients with resectable stage IB-IIIA NSCLC.15 This benefit was seen across all disease stages, histologies, and levels of TMB and PD‑L1 expression. Major pathologic response and objective response rate also were improved with the addition of nivolumab, but we await data on the coprimary endpoint of EFS, which was immature at the time of this analysis.
Furthermore, the addition of nivolumab to chemotherapy maintained a tolerable safety profile and did not impede the feasibility of surgery. Finally, in an exploratory subset analysis, ctDNA clearance was more frequent among patients treated with nivolumab plus chemotherapy, and those with added nivolumab who cleared ctDNA were more likely to achieve a pCR.
In addition to CheckMate 816, other ongoing phase III trials are studying the addition of anti–PD‑1 or anti–PD‑L1 antibodies to neoadjuvant chemotherapy in resectable NSCLC. The 4 other trials with ongoing accrual are KEYNOTE‑671, IMpower030, AEGEAN, and CheckMate 77T.18-21 Each trial has a primary endpoint of EFS. CheckMate 816 and AEGEAN also have a coprimary endpoint of pCR, whereas KEYNOTE‑671 is evaluating both EFS and OS. A key difference between CheckMate 816 and the other trials is that the latter had patients receive adjuvant immunotherapy as well as neoadjuvant immunotherapy.
Turning now to the adjuvant setting, we will discuss the phase III IMpower010 trial, a major adjuvant trial presented at the 2021 American Society of Clinical Oncology Annual Meeting by Dr. Heather Wakelee and colleagues.22 This trial was the first of several large adjuvant trials to be enrolled worldwide and the first to report data.
This trial enrolled patients with completely resected stage IB (≥4 cm) to IIIA NSCLC (AJCC 7th edition). Patients were required to have had an anatomic resection—either a lobectomy or pneumonectomy—as well as tumor tissue available for PD‑L1 assessment. This was a relatively selected population because all patients had to have received at least 1 cycle of adjuvant cisplatin‑based chemotherapy. After chemotherapy, patients were randomized to the anti–PD‑L1 antibody atezolizumab for 1 year or BSC.
The primary endpoint was rather complicated and involved hierarchical evaluation of investigator‑assessed DFS in 3 populations. The first population evaluated was those with PD‑L1 ≥1% in stage II-IIIA disease. If that assessment was statistically significant (defined as a 2-sided α = 0.05), then DFS was assessed in all randomized patients with stage II-IIIA disease irrespective of PD‑L1 expression. If that assessment was significant (2-sided α = 0.05), then DFS was assessed in a broader population of patients with stage IB-IIIA disease irrespective of PD‑L1 expression (2-sided α = 0.05). Finally, if all of the above analyses were positive, then OS would be assessed in the ITT population.
As can be seen in this table of baseline patient characteristics, relatively few patients enrolled on IMpower010 had stage IB tumors (12.2% of all randomized patients), and approximately two thirds had nonsquamous histology.22
A key difference between IMpower010 and CheckMate 816 was that patients with EGFR mutations or ALK alterations were permitted to enroll in IMpower010 but were excluded from CheckMate 816.15,22 In IMpower010, 11.6% of all randomized patients were EGFR positive, and 3.3% were ALK positive.
Adjuvant atezolizumab significantly prolonged DFS vs BSC in patients with stage II-IIIA NSCLC and PD‑L1 TC ≥1%.22 Median DFS was not estimable vs 35.3 months, respectively (HR: 0.66; 95% CI: 0.50-0.88; P = .004).
This benefit appeared consistent across most analyzed subgroups. The only subgroups that did not appear to derive benefit from adjuvant atezolizumab were current smokers and patients with ALK rearrangements.
Turning to the next hierarchical assessment given that the first was positive, atezolizumab again significantly prolonged DFS vs BSC—but this time in all randomized patients with stage II-IIIA NSCLC irrespective of PD‑L1 status.22 The median DFS was 42.3 vs 35.3 months, respectively (HR: 0.79; 95% CI: 0.64-0.96; P = .02). Although positive, this HR suggested less benefit with adjuvant atezolizumab compared with the first population assessed with PD‑L1 TC ≥1%.
The subgroup analyses of all randomized patients with stage II-IIIA NSCLC showed an interesting breakdown by PD‑L1 status. There was a pronounced benefit in patients with PD-L1 TC ≥50% (HR: 0.43; 95% CI: 0.27-0.68) but no benefit in those with PD‑L1 TC <1% (HR: 0.97; 95% CI: 0.72-1.31). Although there was benefit in the subgroup with PD-L1 TC ≥1% (HR: 0.66; 95% CI: 0.49-0.87), we did not see data for those with PD-L1 TC 1% to 49% vs ≥50%, which I consider to be key data needed to interpret this study in more detail.
Consistent with the first assessment, we did not see benefit in patients with either EGFR‑mutated or ALK‑rearranged tumors.
At the time of data cutoff for the preplanned interim DFS analysis, the OS data were immature.22 There was a slight separation of the curves in favor of atezolizumab in patients with stage II-IIIA NSCLC and PD‑L1 TC ≥50%.
In summary, the IMpower010 trial showed a significant improvement in DFS with atezolizumab when given after adjuvant cisplatin-based chemotherapy for patients with resected, PD-L1–positive stage II-IIIA NSCLC, with the magnitude of benefit being greatest for tumors with PD-L1 ≥50% (HR: 0.44).22 Further data are awaited on outcomes for those with PD-L1 1% to 49%, as well as for other hierarchical endpoints.
There are some interesting considerations around the endpoints used for neoadjuvant trials, such as CheckMate 816, and adjuvant trials, such as IMpower010 and ADAURA.
With ADAURA, many oncologists have accepted the positive DFS data as sufficient to incorporate adjuvant osimertinib into practice, because that is how the study was designed and powered. However, others are waiting for the OS data, despite osimertinib being approved by the FDA in this setting. The controversy, particularly with targeted therapies, is whether a DFS benefit will indeed translate into an OS benefit. It is unusual for a targeted therapy to actually cure metastatic lung cancer, with remissions rarely being sustained when patients are off treatment.
However, DFS has been accepted by the FDA as a surrogate endpoint for survival in the adjuvant setting,23 and waiting for mature OS data for early-stage disease can take many years. For metastatic EGFR-mutated NSCLC, the median OS is approximately 3-4 years with osimertinib treatment.24 Patients with early‑stage disease should be expected to have an even longer median OS.
With immunotherapies, there is the idea that you can achieve a tail on the survival curve where approximately 15% of patients with metastatic NSCLC may be effectively cured by immunotherapy and can be off treatment.25 Because of this tail, the thought is that immunotherapy may be more likely to deliver a cure in early‑stage disease. Thus, I imagine that if the FDA approves adjuvant atezolizumab for the treatment of PD-L1–positive resectable early-stage NSCLC—for which it is currently under consideration—there may be less resistance to its uptake in the clinic than there currently is for adjuvant osimertinib.
In trials of neoadjuvant therapy, the common endpoints are pCR and EFS, with pCR being the first to report. pCR has not been accepted by the FDA as a surrogate endpoint for survival, so we likely will be waiting for EFS to report before neoadjuvant nivolumab will be considered for approval. However, perhaps CheckMate 816 will be the first prospective study to provide robust data on pCR as a surrogate endpoint for EFS.
Along with IMpower010, 3 other key ongoing phase III trials studying anti–PD‑1 or anti–PD‑L1 antibodies as adjuvant therapy in NSCLC have completed accrual: ANVIL, PEARLS, and BR31.26-29 Accrual for the ALCHEMIST Chemo-IO (ACCIO) trial is still ongoing.30 All of these trials have DFS as a primary endpoint, and ANVIL and ALCHEMIST Chemo-IO also have OS as a coprimary endpoint.
Each trial is examining a different anti–PD-1 or anti–PD-L1 antibody, either alone or with additional adjuvant therapy. As we just discussed, IMpower010 is comparing adjuvant atezolizumab vs BSC following adjuvant chemotherapy. ANVIL is comparing adjuvant nivolumab vs observation, PEARLS is comparing adjuvant pembrolizumab vs placebo, and BR31 is comparing adjuvant durvalumab vs placebo. All 3 of these trials permit, but do not require, additional adjuvant therapy prior to receipt of immunotherapy. Finally, ALCHEMIST Chemo-IO has 3 arms: concurrent adjuvant pembrolizumab plus chemotherapy followed by pembrolizumab monotherapy, sequential adjuvant pembrolizumab monotherapy after adjuvant chemotherapy, and observation after adjuvant chemotherapy.
ANVIL, PEARLS, and BR31 are due to report the primary analysis within the next several years. IMpower010 is the only trial for which we have any data at this time.
To summarize, in 2021 we have started to see data on incorporating immunotherapy into the treatment of resectable early-stage NSCLC. The CheckMate 816 trial showed that adding PD‑1 blockade to neoadjuvant chemotherapy increased response and pCR rates without significantly increasing toxicity.15 In the adjuvant setting, IMpower010 showed that PD‑L1 blockade after surgery and chemotherapy improved DFS in patients with PD‑L1–positive NSCLC.22 These trials are notable as the first phase III trials to show a statistically significant benefit with neoadjuvant or adjuvant immunotherapy, respectively, for the treatment of resectable NSCLC.
Based on these positive phase III trials, I consider it likely that immunotherapy will become a mainstay of perioperative management of stage II-IIIA NSCLC in the near future. Exactly how immunotherapy will be incorporated into neoadjuvant and/or adjuvant therapy will be determined by the final data from these trials, particularly subgroup analyses of benefit by PD‑L1 status and disease stage. More patients with stage IIIA disease were enrolled in the neoadjuvant CheckMate 816 trial, whereas more patients with stage IB-II disease were enrolled in the adjuvant IMpower010 trial. This could certainly affect our determination of whether neoadjuvant or adjuvant immunotherapy is appropriate for particular patients. We also must await data on how early use of immunotherapy affects treatment selection for our patients experiencing relapse and advanced NSCLC.