Data from completed studies evaluating immunotherapy in lung cancer alone or in combination with chemotherapy show that a breakdown of populations by PD-L1 status and additional considerations such as toxicity profiles should guide treatment decisions.
The therapeutic landscape for first-line treatment of metastatic non–small cell lung cancer (NSCLC) contains many options. Progress made in the area of targeted therapies has pushed treatment with small-molecule inhibitors to the front-line setting for patients with aberrations in EGFR, ALK, ROS1, BRAF, MET, and RET. Despite these advances, the best treatment options for patients with nondriver met-astatic NSCLC remain unclear. Recent studies have demonstrated a role for immuno-oncology (IO) agents in the first line, but the added benefit of combining them with chemotherapy (chemo IO) is not clear in all populations. Data from completed studies evaluating IO alone or in combination with chemotherapy show that a breakdown of populations by PD-L1 status and additional considerations such as toxicity profiles should guide treatment decisions.
Several studies have evaluated the role of immunotherapy as monotherapy in patients with PD-L1 expression of at least 50%. In KEYNOTE-024 (NCT02142738),1 pembrolizumab (Keytruda) was com-pared with platinum-based chemotherapy. Investigators reported a median over-all survival (OS) of 30.0 months (95% CI, 18.3–not reached [NR]) in the pembrolizumab arm and 14.2 months (95% CI, 9.8-19.0) in the chemotherapy arm with a hazard ratio of 0.63 (95% CI, 0.47-0.86) in patients with a PD-L1 tumor proportion score (TPS) of at least 50% by immunohistochemistry (IHC) 22C3 pharmDX assay.2 The KEYNOTE-042 (NCT02220894)3 open-label, phase 3 study was conducted in 213 medical centers in 32 countries. Eligible patients were adults (≥18 years). The study confirmed the benefit of pembrolizumab compared with chemotherapy in the population of patients with PD-L1 TPS of at least 50%, also evaluated by IHC 22C3. Data showed a median OS of 20.0 months (95% CI, 15.4-24.9) for pembrolizumab versus 12.2 months (95% CI, 10.4-14.2) for chemotherapy with a hazard ratio of 0.69 (95% CI, 0.56-0.85). Of note, the pembrolizumab regimen included up to 2 years of IO and reported a 31% survival rate for those who received IO at 3 years. These data indicate a durable effect at 1 year following completion of treatment.
A population of patients with PD-L1 expression of at least 50% also demonstrated benefit with IO in the IMpower110 study (NCT02409342), in which investigators evaluated the efficacy of atezolizumab (Tecentriq) versus chemotherapy.4 The OS for patients with high PD-L1 expression was 20.2 months with atezolizumab (95% CI, 16.5-not evaluable [NE]) com-pared with 13.1 months in the chemotherapy arm (95% CI, 7.4-16.5) with a hazard ratio of 0.59 (95% CI, 0.40-0.89). High PD-L1 expression in IMpower110 was defined as PD-L1 tumor cell expression of at least 50% or tumor-infiltrating immune cell expression of at least 10% by SP142 IHC.
The KEYNOTE-189 (NCT02578680)5 study compared pembrolizumab plus chemo-therapy with chemotherapy alone in a nonsquamous NSCLC population. Final analysis data presented at the 2020 American Society of Clinical Oncology (ASCO) Virtual Scientific Program showed a median OS of 27.7 months for the population of patients with PD-L1 TPS of at least 50%, evaluated by IHC 22C3 (95% CI, 20.4-NR) compared with 10.1 months (95% CI, 7.5-22.0) in the chemotherapy group with a hazard ratio of 0.59 (95% CI, 0.40-0.86).6
In CheckMate 227 (NCT02477826),7 nivolumab (Opdivo) plus ipilimumab (Yervoy) was compared with chemo-therapy. In the subgroup analysis of patients with PD-L1 TPS of at least 50%, those who received dual immunotherapy had an OS of 21.2 months (95% CI, 15.5-38.2) compared with 14.0 months (95% CI, 10.0-18.6) in the chemotherapy- alone population with a hazard ratio of 0.70 (95% CI, 0.55-0.90). In CheckMate 9LA (NCT03215706),8 which compared nivolumab plus ipilimumab plus chemo-therapy with chemotherapy alone, investigators reported a median OS of 18 months compared with 12.6 months, respectively, with a hazard ratio of 0.66 for patients with PD-L1 TPS of at least 50%, assessed by 28-8 IHC.
The relative parity in the findings from these studies is summarized in FIGURES 1A AND 1B. Given that the survival advantages provided by IO compared with chemo IO and chemotherapy alone are minimal, the decision for first-line therapy may be guided by disease burden and safety considerations from these agents.
KEYNOTE-024 showed 31.2% of grade 3-5 adverse events (AEs) in the pembrolizumab group compared with 53.3% in the chemotherapy group. Similarly, in the IMpower110 study, 30.1% of patients treated with atezolizumab experienced grade 3 or 4 AEs, compared with 52.5% for chemotherapy.
Further, the objective response rates report-ed for the KEYNOTE-189 study in the population of patients with PD-L1 TPS of at least 50% were 62.1% (95% CI, 53.3%-70.4%) compared with 25.7% (95% CI, 16.0%-37.6%) in the pembrolizumab plus chemotherapy and chemo-therapy-alone cohorts, respectively.6 Given the improved objective response rate, it would be reasonable to treat a patient with a higher disease burden and PD-L1 TPS of at least 50% with the KEYNOTE-189 regimen, but given the more mature data and durable responses and improved safety profile seen with pembrolizumab monotherapy in the KEYNOTE-024 study, IO alone is a reasonable first-line choice in patients with PD-L1 of at least 50% with a lower disease burden.
In patients with nonsquamous NSCLC with PD-L1 TPS less than 1%, data from KEYNOTE-189 showed an OS of 17.2 months (95% CI, 13.8-22.8) for combined pembrolizumab plus chemotherapy compared with 10.2 months (95% CI, 7.0-13.5) with a hazard ratio of 0.51 (95% CI, 0.36-0.71).6 In KEYNOTE-407 (NCT02775435), patients with squamous NSCLC were treated with either pembrolizumab plus chemotherapy or chemotherapy alone. In the population of patients with PD-L1 TPS less than 1%, the OS with combined therapy was 15.9 months (95% CI, 13.1-NE) versus 10.2 months (95% CI, 8.6-13.8) for chemotherapy alone, with a hazard ratio of 0.61 (95% CI, 0.38-0.98).9
The primary statistical analysis for CheckMate 227 for patients with PD-L1 TPS of at least 1% supported the FDA approval for nivolumab and ipilimumab in this patient population. However, the more interesting result was seen in the population with PD-L1 TPS less than 1%. Data showed an OS of 17.2 months in the immunotherapy group versus 12.2 months for those who received chemotherapy, with a hazard ratio of 0.64 (95% CI, 0.51-0.81).10 Similarly, in data from the population of patients with PD-L1 TPS less than 1% in the CheckMate 9LA study, the OS for the dual immunotherapy plus chemotherapy group was 16.8 months compared with 9.8 months with chemotherapy alone (HR, 0.62; 95% CI, 0.45-0.85).8
The findings from these studies are summarized in FIGURES 1C AND 1D. CheckMate 227 reported a similar safety profile for both regimens, with 33% of patients experiencing grade 3/4 AEs with immunotherapy and 36% experiencing AEs with chemotherapy. However, immune-related AEs (irAEs) related to ipilimumab-based combinations, in particular colitis, require careful monitoring. The rates of colitis were nearly twice as high with immunotherapy compared with chemo-therapy in CheckMate 227, with 17.0% of patients experiencing any-grade diarrhea compared with 9.6%, respectively. Further, grade 3 or higher AEs were reported in 1.7% and 0.7%, respectively.10
Although it is difficult to compare across trials, the rate of grade 3 or greater AEs were less frequent than reported in the combined immunotherapy and chemotherapy arm of the CheckMate 9LA trial (47%). Both event rates are lower than those reported in KEYNOTE-189 results—71.9% and 66.8% for grade 3 or higher AEs for patients with combination therapy and chemotherapy alone, respectively. Some of these differences are likely secondary to a longer follow-up period, but it is reasonable to suspect a lower rate of AEs if chemotherapy is removed from the treatment regimen. Data from both KEYNOTE-189 and KEYNOTE-407 present reasonable first-line therapy options for patients with PD-L1 expression less than 1%.
Nivolumab plus ipilimumab is a particularly attractive option for patients with PD-L1 expression less than 1%. However, selecting nivolumab and ipilimumab in combination with chemotherapy per the data from CheckMate 9LA is a reasonable approach, and more mature data are pending.
In the KEYNOTE-042 trial, patients with PD-L1 TPS of at least 1% had a reported OS of 16.7 months (95% CI, 13.9-19.7) in the pembrolizumab group compared with 12.1 months (95% CI, 11.3-13.3) with chemotherapy (HR, 0.81; 95% CI, 0.71-0.93). In a subgroup analysis of patients with PD-L1 TPS of 1% to 49%, no significant median OS advantage was seen between the 2 treatments (13.4 vs 12.1; HR, 0.92; 95% CI, 0.77-1.11).3
Atezolizumab monotherapy in the population of patients with PD-L1 of at least 1% in IMpower110 led to an OS of 17.5 months (95% CI, 12.8-23.1) compared with 14.1 months (95% CI, 11.0- 16.6) in patients receiving chemotherapy alone, with a hazard ratio of 0.83 (95% CI, 0.651.07). It is important to note that in this study population, patients had at least 1% of tumor cells (TC) or tumor-infiltrating immune cells (IC) expressing PD-L1 (TC1/2/3 or IC 1/2/3). However, no significant median OS advantage was seen in the subgroup analyses for TC1/2 or IC1/2 (< 50% TC PD-L1 versus < 10% IC PD-L1; HR, 1.04; 95% CI, 0.76-1.44), TPS 1% to 49% by IHC 22C3 (HR, 1.00; 95% CI, 0.63-1.58) or TC 1% to 49% by SP263 (HR, 0.94; 95% CI, 0.58-1.53).
In KEYNOTE-189, in patients with nonsquamous NSCLC and a PD-L1 TPS 1% to 49%, pembrolizumab plus chemotherapy led to an OS of 21.8 months (95% CI, 17.7-25.9) compared with 12.1 months (95% CI, 8.7-19.4) for chemotherapy, with a hazard ratio of 0.62 (95% CI, 0.42-0.92). Combined pembrolizumab plus chemotherapy versus chemotherapy alone in the squamous NSCLC population of KEYNOTE-407 with PD-L1 TPS of at least 1% led to an OS of 18.1 months (95% CI, 14.9 -22.2) versus 12.8 months (95% CI, 9.5-14.7) for chemotherapy alone, with a hazard ratio of 0.67 (95% CI, 0.51-0.87). In the patients with PD-L1 1% to 49% subgroup, combined treatment led to significant improvement in OS, with a hazard ratio of 0.59 (95% CI, 0.42-0.84).
CheckMate 227 compared the dual IO approach of nivolumab plus ipilimumab versus chemotherapy in patients with PD-L1 TPS of at least 1%. The combination led to an OS of 17.1 months versus 14.9 months with a hazard ratio of 0.79 (95% CI, 0.67-0.93). The subgroup analysis of patients with PD-L1 TPS 1% to 49% failed to show an OS advantage, with a hazard ratio of 0.94 (95% CI, 0.75-1.18). Adding nivolumab and ipilimumab to chemotherapy in CheckMate 9LA for patients with PD-L1 TPS of at least 1% led to an OS of 15.8 months compared with 10.9 months with chemotherapy alone, with a hazard ratio of 0.64 (95% CI, 0.50-0.82). In the subgroup with PD-L1 1% to 49%, the combined approach led to an OS of 15.4 months compared with 10.4 months, (HR, 0.61; 95% CI, 0.44-0.84).
These data are summarized in FIGURES 1E AND 1G. It is important to note that subgroup analysis of patients with PD-L1 1% to 49% failed to demonstrate an OS benefit to IO alone over chemotherapy alone. However, in the chemo IO trials, there was a significant benefit in these groups. As a result, chemoimmunotherapy is the standard of care for this subgroup of patients.
As mentioned above, targeted therapy has an increasing role in the first-line treatment of metastatic NSCLC. These approaches must be integrated into the management of patients with NSCLC. It is critical to establish the presence or absence of targetable driver mutations as soon as possible.
Blood-based and tissue-based next-generation sequencing (NGS) panels have emerged as the most comprehensive and efficient way to obtain this information. However, if a patient is in extremis and cannot wait the 1 to 3 weeks needed to complete this analysis, it is reasonable to initiate chemotherapy alone. Due to an increased risk of IO-mediated AEs, including pneumonitis and hepatitis, in patients treated with IO followed by tyrosine kinase inhibitors (TKIs), IO should be held until mutational status is known.
This is supported by 6 of 7 patients experiencing a TKI-induced AE following immunotherapy in a phase 2 study (NCT02879994) evaluating response of EGFR- mutant PD-L1–positive patients assigned to pembrolizumab prior to treatment with a TKI.11 This analysis included 1 case of fatal pneumonitis. Our work has also shown 3 of 9 patients with grade 3 transaminitis and 1 patient with grade 4 pneumonitis when crizotinib (Xalkori) was combined with pembrolizumab in a phase 1b study (NCT02511184) of previously untreated advanced NSCLC patients with ALK aberrations.12 The study was terminated based on low enrollment, but it is noted that the agents demonstrated high efficacy in first-line ALK-positive NSCLC. Following the initial cycle of chemotherapy, a targeted agent or IO can be implemented as directed by the NGS results.
The treatment landscape for metastatic NSCLC without a targetable alteration is complex. IO has a role in the first-line treatment of all patients with no targetable driver mutations. Although we have presented favored options for each specific PD-L1 population above, each patient should be evaluated in a personalized fashion.
Patients who are heavily symptomatic should be treated with chemotherapy-based combinations, with the addition of IO following a determination of targetable gene alteration. An asymptomatic patient with PD-L1 expression of at least 50% should receive IO alone. Given the durable OS response from KEYNOTE-024, one may be able to discontinue IO after 2 years per the data from frontline studies. It is reasonable to suspect that the patient’s immune system can maintain the antitumor effect at this point, a hypothesis that is potentially supported by the 31% survival data at 3 years.
In our practice, we typically restage the patient at 2 years and consider biopsy of positron emission tomography–avid lesions. We offer the option to stop IO therapy to patients who have no evidence of disease or a negative biopsy. We offer stereotactic body radiation therapy to patients who are oligometastatic and continue IO in patients with persistent active multifocal cancer. In a patient with PD-L1 expression less than 1%, ipilimumab-based combinations may be a reasonable option for squamous and nonsquamous metastatic NSCLC. The agent has shown comparable OS and an improved AE profile compared with chemo IO. The addition of chemotherapy to this regimen continues to be evaluated and may be a good option based on the findings from CheckMate 9LA.
It is notable that ipilimumab was given 1 mg/kg every 6 weeks in these trials, and nivolumab varied from 240 mg every 2 weeks to 360 mg every 3 weeks in CheckMate 227 and CheckMate 9LA, respectively. This dosing differs from combined regimens of these agents when used for other malignancies, such as melanoma, which may contribute to lower rates of irAEs secondary to lower doses of ipilimumab. The ongoing CheckMate 9LA study will help us understand if chemo IO with ipilimumab outperforms the current PD-L1–based chemo IO approaches, but for now regimens guided by KEYNOTE-189 and KEYNOTE-407 have longer-term follow-up. In January 2020, ASCO updated their clinical practice guidelines to reflect data from the KEYNOTE-042, KEYNOTE-189, and KEYNOTE-407 trials, among others (FIGURE 2).29
Ultimately, having more options in the treat-ment of nondriver metastatic NSCLC benefits both the patient and practitioner by providing the best therapeutic regimen to try to maximize the benefit to the patient. Patient selection for IO utilizing PD-L1 IHC is a pertinent but limit-ing factor in optimal treatment selection.