From Hope to Reality: Durable Overall Survival With Immune Checkpoint Inhibitors for Advanced Lung Cancer

After several decades of frustratingly small advances, a revolution in the therapeutic landscape of advanced non–small-cell lung cancer (NSCLC) has occurred in the span of only four short years. During this time, several immune checkpoint inhibitors (ICIs)—pembrolizumab, nivolumab, atezolizumab, and durvalumab—have secured approvals from the United States Food and Drug Administration (FDA) for the management of locally advanced and metastatic NSCLC.^1-5 ICI alone or in combination with chemotherapy is now the standard first-line therapy for advanced NSCLC without other actionable genomic alterations.

First published in 2015, KEYNOTE-001 was a landmark study exploring the use of the programmed cell death protein 1 (PD-1) antagonist pembrolizumab in treatment-naïve and previously treated advanced NSCLC. In this trial, single-agent pembrolizumab resulted in radiographic responses and established the programmed death ligand 1 (PD-L1) tumor proportion score (TPS) using the anti–PD-L1 pharmDX 22C3 antibody as a useful biomarker for patient selection. Patients with tumor PD-L1 TPS ≥ 50% experienced the highest objective response rate (ORR) and progression-free survival (PFS) at 45.2% and 6.3 months, respectively.¹ In 2015, pembrolizumab was granted accelerated approval by the FDA for use in patients with previously treated advanced NSCLC and tumor PD-L1 TPS ≥ 50%.

On the heels of KEYNOTE-001, multiple ensuing studies explored pembrolizumab use in the second-line setting. In KEYNOTE-010, patients with advanced NSCLC and tumor PD-L1 TPS ≥ 1% with disease progression after front-line platinum doublet chemotherapy achieved superior overall survival (OS) when treated with pembrolizumab as compared with those receiving docetaxel, and with fewer adverse events (AEs).⁶ The superior efficacy and toxicity profile of ICI monotherapy in the second-line setting when compared with docetaxel was reaffirmed in multiple contemporaneous studies of both nivolumab (anti–PD-1, CHECKMATE-017, CHECKMATE-057) and atezolizumab (anti–PD-L1, OAK).^2-4 Notably, tumor PD-L1 positivity was not a requirement in these latter studies.

Next, KEYNOTE-024 launched pembrolizumab into the front-line setting, demonstrating improved ORR, PFS and OS with ICIs compared with platinum doublet chemotherapy in tumors with PD-L1 TPS ≥ 50%.⁷ Interestingly, CHECKMATE-026, a comparable study of first-line nivolumab versus platinum doublet chemotherapy, did not show improved outcomes; much has been explored in the ensuing analysis regarding variations in patient and biomarker selection that may have accounted for this difference.⁸ Recently published, KEYNOTE-042 further expanded the use of single-agent pembrolizumab for previously untreated patients with tumor PD-L1 TPS of ≥ 1%, demonstrating improved OS with pembrolizumab in the overall study population when compared with platinum doublet chemotherapy. Notably, this benefit was mostly driven by the population of patients already identified in KEYNOTE-024 (ie, those with tumor PD-L1 TPS ≥ 50%).⁹

But the story cannot end here, as a substantial proportion of patients remain for whom biomarker-driven single-agent immune or targeted therapy will not afford needed palliation of their disease. Numerous variations of platinum doublets combined with ICIs with or without biologic agents (such as bevacizumab) have been explored in KEYNOTE-021 cohort G (nonsquamous), KEYNOTE-189 (nonsquamous), KEYNOTE-407 (squamous), and IMpower150 (nonsquamous) and have demonstrated the superiority of up-front ICIs combined with chemotherapy.^10-13 As in the prior ICI-only trials, the hallmark features of these chemoimmunotherapy strategies are augmented ORR and improved median OS—often with similar nonimmune rates of AEs as would be seen with chemotherapy alone.

As a consequence of the progress made in these iterative studies, the FDA’s current approval for pembrolizumab as of this publication includes: (1) in the front-line setting for tumors lacking actionable alterations in epidermal growth factor receptor (EGFR) or anaplastic lymphoma kinase (ALK) and with tumor PD-L1 TPS ≥ 1%, (2) in the front-line setting combined with chemotherapy for tumors lacking actionable alterations in EGFR/ALK and regardless of tumor PD-L1 status, and (3) after platinum doublet or targeted therapies for ICI-naïve patients whose tumors have PD-L1 TPS ≥ 1%. Notably, the activity of ICIs in tumors with abnormalities in EGFR or ALK has been disappointing, for reasons that are as yet not fully understood. The benefits of ICI monotherapy in these patients therefore remain limited.^1-4,6-9 However, for the substantial subset of patients with advanced NSCLC without contraindications to immune-based therapies and lacking actionable genomic alterations, the evidence-based standard of care for initial therapy now necessarily involves the use of ICIs.

A key question that remains is how this approach will affect long-term outcomes in this hitherto recalcitrant disease. Certainly, the experience with use of tyrosine kinase inhibitors (TKIs) in oncogene-driven tumors has raised the bar regarding expectations of robust clinical efficacy, overall tolerability, and therapeutic durability, with potential to optimize long-term survival. For EGFR-mutated, ALK-rearranged, and ROS proto-oncogene 1 receptor tyrosine kinase (ROS1)-rearranged advanced NSCLC treated with relevant TKIs, the 5-year OS rates can exceed 15%, 35% to 50%, and 40%, respectively.^14-18 These numbers are only likely to further increase as more effective next-generation drugs enter the therapeutic sphere.

In the article that accompanies this editorial, Garon and colleagues¹⁹ provide much-awaited updated outcome data for patients with advanced NSCLC treated with single-agent pembrolizumab in KEYNOTE-001. Of the 449 previously treated and 101 treatment-naïve patients included in this analysis, median OS was 10.5 and 22.3 months, respectively. Perhaps even more compelling, the 2-, 3-, 4-, and 5-year OS were 30.1%, 20.9%, 18.2%, and 15.5% for previously treated patients and 49.0%, 37.0%, 31.0%, and 23.2% for treatment-naïve patients.¹⁹ The evolution of pembrolizumab use during the life span of the trial led to treatment assignments in long-term responders that better match the current use of this agent (ie, at a fixed dose of 200 mg every 3 weeks for up to 24 months in patients without disease progression). It is reassuring to note that late-onset immune AEs were rare in this long-term follow-up experience.¹⁹ PD-L1 TPS continued to be the most robust biomarker associated with long-term outcomes, with the most vigorous outcomes reported for treatment-naïve patients with tumor PD-L1 TPS ≥ 50%: a heretofore unheard of median OS of 35.4 months and 5-year OS of 29.6%.¹⁹ In addition, those patients receiving 24 months or more of pembrolizumab were enriched to have an ORR greater than 85% and 5-year OS greater than 75%, with most all responses noted within the first 6 months of therapy. Previously published long-term outcomes from the initial phase I study of nivolumab similarly showed an approximately 16% 5-year OS in a heavily pretreated population.²⁰ None of the other advanced-phase studies of these agents have yet reported their long-term (ie, 5-year OS and beyond) outcomes—and these, too, are eagerly awaited.

The successes of up-front tumor molecular profiling and therapeutic stratification in patients with advanced-stage disease have set NSCLC apart in the era of precision medicine. Therapeutic selection based on tumor-based biomarkers has been transformative for actionable genomic alterations in EGFR, ALK, and ROS1 (among others), the identification of the biomarker can clearly delineate patients who should (or should not) receive a given therapy. However, the same cannot be so simply said for immune-based biomarkers in clinical practice, as even well-established markers such as PD-L1 TPS cannot always identify patients who will or will not respond to a given ICI. KEYNOTE-001 highlights these persistent areas of uncertainty, as even patients at the most enriched upper quartile of PD-L1 TPS (75% to 100%) had an ORR of only 54%; conversely, a significant proportion of patients with protracted disease control were in the category of PD-L1 TPS less than 50%.¹⁹ Identifying a single robust predictive biomarker that can fully capture the breadth and scope of clinical responses to ICIs is perhaps an unrealistic goal. Rather, the field is focused on identifying additional metrics that include several distinct biomarkers that will likely be both necessary and feasible, including assessment of: tumor mutation burden, microsatellite instability, mutations in DNA damage repair pathways, tumor-infiltrating lymphocytes, tobacco exposure, presence/absence of tumor suppressor genes that affect innate immune function, and others.²¹

Many other questions about optimal use of immunotherapy strategies remain unanswered. The optimal duration of therapy with ICIs either alone or in combination with maintenance chemotherapy remains a clinical conundrum. To date, no rigorous prospective investigation has yet established whether treatment lasting shorter or longer than 24 months in stable/responding patients might achieve comparable or better outcomes—an issue that is germane to all involved stakeholders, given the escalating limitations on health care delivery and finances. Understanding when we must necessarily add chemotherapy to the immunotherapy backbone also remains of interest, as most recent studies of first-line combination versus single-agent regimens have compared pembrolizumab alone to the now-defunct standard of chemotherapy alone, rather than pembrolizumab (or another PD-1/PD-L1 antibody) versus chemoimmunotherapy. Exploring optimal sequencing of the immunotherapy backbone with cytotoxic chemotherapy, other immunologic/biologic agents, and/or palliative radiotherapy also remains critical, as emphasizing combination strategies alone is unlikely to permit maximal use of these agents among all who may benefit.

In conclusion, ICIs have rightfully earned their place as a mainstay of care for advanced NSCLC. KEYNOTE-001 and similar trials have unequivocally established that durable long-term responses with previously unseen extension of survival are feasible with ICIs. Five-year survival rates can now be expected for a growing number of patients with advanced NSCLC treated with biomarker-driven use of ICIs in those without actionable alterations or TKIs for those with actionable alterations, thus achieving the quest to offer each patient the most optimally durable, efficacious, and tolerable strategy for their care. As we look to the future, the full potential of ICIs and TKIs might only be fully realized when these agents move into the early-stage disease setting. Use of ICIs or TKIs in the neoadjuvant or adjuvant setting may well be transformative when partnered with ongoing worldwide efforts dedicated toward smoking cessation and early detection modalities. Given the progress made in the past decade, it is not unrealistic to think that within future decades, the dismal survival rates for all stages of lung cancer²² can be meaningfully—and durably—improved beyond what has already been achieved in studies like KEYNOTE-001.

AUTHOR CONTRIBUTIONS

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

From Hope to Reality: Durable Overall Survival With Immune Checkpoint Inhibitors for Advanced Lung Cancer

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/site/ifc.

Deepa Rangachari

Consulting or Advisory Role: Advance Medical, DynaMed

Research Funding: Bristol-Myers Squibb (Inst), Novocure (Inst), AbbVie/Stemcentrx (Inst)

Daniel B. Costa

Honoraria: Pfizer, Takeda/Millennium Pharmaceuticals

Consulting or Advisory Role: AstraZeneca, Takeda/Millennium Pharmaceuticals

Research Funding: Takeda/Millennium Pharmaceuticals (Inst), AstraZeneca (Inst), Pfizer (Inst), Merck (Inst), Merrimack (Inst), Bristol-Myers Squibb (Inst), Clovis Oncology (Inst), Genentech (Inst), Tesaro (Inst)

No other potential conflicts of interest were reported.