Nivolumab Plus Chemotherapy in Epidermal Growth Factor Receptor–Mutated Metastatic Non–Small-Cell Lung Cancer After Disease Progression on Epidermal Growth Factor Receptor Tyrosine Kinase Inhibitors: Final Results of CheckMate 722

Abstract

PURPOSE

The phase III CheckMate 722 trial (ClinicalTrials.gov identifier: NCT02864251) evaluated nivolumab plus chemotherapy versus chemotherapy in patients with epidermal growth factor receptor (EGFR)–mutated metastatic non–small-cell lung cancer (NSCLC) after disease progression on EGFR tyrosine kinase inhibitors (TKIs).

METHODS

Patients with disease progression after first- or second-generation EGFR TKI therapy (without EGFR T790M mutation) or osimertinib (with/without T790M mutation) were randomly assigned 1:1 to nivolumab (360 mg once every 3 weeks) plus platinum-doublet chemotherapy (once every 3 weeks) or platinum-doublet chemotherapy alone (once every 3 weeks) for four cycles. Primary end point was progression-free survival (PFS). Secondary end points included 9- and 12-month PFS rates, overall survival (OS), objective response rate (ORR), and duration of response (DOR).

RESULTS

Overall, 294 patients were randomly assigned. At final analysis (median follow-up, 38.1 months), PFS was not significantly improved with nivolumab plus chemotherapy versus chemotherapy (median, 5.6 v 5.4 months; hazard ratio [HR], 0.75 [95% CI, 0.56 to 1.00]; P = .0528), with 9- and 12-month PFS rates of 25.9% versus 19.8%, and 21.2% versus 15.9%, respectively. Post hoc PFS subgroup analyses showed a trend favoring nivolumab plus chemotherapy in patients with tumors harboring sensitizing EGFR mutations (HR, 0.72 [95% CI, 0.54 to 0.97]), one line of previous EGFR TKI (0.72 [95% CI, 0.54 to 0.97]), or both (0.64 [95% CI, 0.47 to 0.88]). Median OS was 19.4 months with nivolumab plus chemotherapy versus 15.9 months with chemotherapy, while ORR was 31.3% versus 26.7%, and median DOR was 6.7 versus 5.6 months, respectively. Grade 3/4 treatment-related adverse events occurred in 44.7% and 29.4% of patients treated with nivolumab plus chemotherapy and chemotherapy alone, respectively.

CONCLUSION

Nivolumab plus chemotherapy did not significantly improve PFS versus chemotherapy in patients with EGFR-mutated metastatic NSCLC previously treated with EGFR TKIs. No new safety signals were identified.

INTRODUCTION

Lung cancer is the leading cause of global cancer deaths (18%), with an estimated 1.8 million deaths each year.¹ In approximately 10%-15% of Caucasian patients and 30%-50% of Asian patients, non–small-cell lung cancer (NSCLC) is driven by mutations in the epidermal growth factor receptor (EGFR) gene^2-5; the reason for higher prevalence of EGFR mutations in Asian patients remains unknown.^6-8

CONTEXT

• Key Objective

• Although epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) are approved for first-line treatment of patients with EGFR-mutated metastatic non–small-cell lung cancer (mNSCLC), most patients develop resistance, limiting treatment options. CheckMate 722 evaluated nivolumab plus chemotherapy versus chemotherapy in patients with EGFR-mutated mNSCLC after disease progression on EGFR TKIs.

• Knowledge Generated

• Nivolumab plus chemotherapy did not significantly improve progression-free survival in patients with previously treated EGFR-mutated mNSCLC. Post hoc analyses showed a trend favoring nivolumab plus chemotherapy versus chemotherapy among patients with tumors harboring sensitizing EGFR mutations, only one line of previous EGFR TKI therapy, or both.

• Relevance (T.E. Stinchcombe)

• For patients with EGFR mutant NSCLC with disease progression after receiving an EGFR TKI, platinum-based therapy with immunotherapy did improve outcomes compared to chemotherapy alone.*

  *Relevance section written by JCO Associate Editor Thomas E. Stinchcombe, MD.

EGFR tyrosine kinase inhibitors (TKIs), including the third-generation osimertinib, are approved for the first-line treatment of patients with metastatic NSCLC harboring sensitizing EGFR mutations.^9-11 Studies of patients with EGFR-mutated NSCLC who received previous TKIs have reported 5-year survival rates of 15%-24%.^12,13 Most patients develop resistance to EGFR TKIs,² and in up to 60% of cases, resistance to first-/second-generation EGFR TKIs is due to acquired EGFR exon 20 T790M mutations,² and for these patients, osimertinib is the preferred treatment option.^9-11 For patients without tumors harboring T790M mutations whose disease progressed after first-/second-generation TKI or osimertinib, platinum-based doublet chemotherapy is the standard treatment.⁹ In this setting, tumor response rate and median progression-free survival (PFS) with chemotherapy are 31%-34% and 5.4-4.4 months, respectively,^14,15 thereby highlighting a clear unmet need for patients with EGFR TKI resistance.

Single-agent immunotherapies (eg, nivolumab, pembrolizumab, or atezolizumab),^16-18 dual immunotherapy regimens,¹⁹ and immunotherapy plus platinum-based chemotherapy combinations have shown clinical benefits in patients with metastatic NSCLC,^20,21 including patients with EGFR/ALK mutation after previous chemotherapy.²² Nivolumab is a PD-1 inhibitor approved globally, including in the United States, Europe, and Japan, for treatment of metastatic NSCLC progressing on or after platinum-based chemotherapy, including patients with tumors harboring EGFR mutations who have received previous EGFR TKIs.^23-25 Nivolumab in combination with chemotherapy has also shown encouraging clinical activity in a phase I study of patients with metastatic NSCLC that included a small subpopulation of patients with EGFR-mutated tumors whose disease progressed after EGFR TKI therapy.²⁶ However, the efficacy of nivolumab plus chemotherapy in patients with EGFR-mutated metastatic NSCLC after progression with EGFR TKIs is not well studied. Here, we report the final efficacy and safety results from CheckMate 722, a randomized, open-label, phase III study of nivolumab plus chemotherapy compared with chemotherapy in patients with EGFR-mutated metastatic NSCLC after disease progression on EGFR TKIs.

METHODS

Patients

Eligible patients were age 18 years and older with histologically confirmed stage IV or recurrent EGFR mutation-positive (ie, exon 19 deletion, L858R, T790M, L861Q, exon 20 insertion, G719X, S768I, other) NSCLC, measurable disease per RECIST v1.1,²⁷ and an Eastern Cooperative Oncology Group performance status of 0-1. Additionally, patients had disease progression on previous first-line first-/second-generation EGFR TKI therapy and had no evidence of EGFR T790M mutation, or disease progression on first-/second-line osimertinib regardless of EGFR T790M status after a protocol amendment during the enrollment period. Tumor tissue sample availability (fresh or archival within 6 months before first dose of study drug) was required at enrollment; an optional tumor biopsy was recommended but not required at disease progression. Patients with EGFR T790M mutation-positive NSCLC after disease progression on first-line, first-/second-generation TKI were eligible only if they also had disease progression on second-line osimertinib. Presence of treated or asymptomatic brain metastases or a history of smoking was allowed. Patients with known ALK translocations or any previous systemic therapy other than the above-listed EGFR TKIs were excluded.

Study Design and Treatment

CheckMate 722 was a randomized, open-label, phase III trial (ClinicalTrials.gov identifier: NCT02864251; Data Supplement, Fig S1 [online only]). Patients were randomly assigned 1:1 to nivolumab (360 mg) plus platinum-doublet chemotherapy comprising pemetrexed (500 mg/m²) plus cisplatin (75 mg/m²) or carboplatin (area under the curve 5 or 6) once every 3 weeks for four cycles, or to platinum-doublet chemotherapy once every 3 weeks for four cycles. Upon completion of four cycles, patients in the nivolumab plus chemotherapy arm without disease progression continued nivolumab and pemetrexed once every 3 weeks, until disease progression, unacceptable toxicity, withdrawal of consent, or for ≤2 years. Patients in the chemotherapy arm with stable disease or response received maintenance therapy (pemetrexed alone once every 3 weeks) until disease progression or unacceptable toxicity. Stratification factors included tumor PD-L1 expression (≥1% or <1%/not evaluable/indeterminate), brain metastases (presence or absence), smoking history (current/former or never), and previous osimertinib use (yes or no). Enrollment for a third randomized group for nivolumab (3 mg/kg once every 2 weeks) plus ipilimumab (1 mg/kg once every 6 weeks) was closed early following external trial data reported in 2018²⁸; no safety concerns were identified in patients who received nivolumab plus ipilimumab. Additional information on protocol amendments and assessments is provided in the Data Supplement.

The CheckMate 722 study was conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonization Good Clinical Practice guidelines. Institutional review boards or independent ethics committees approved the study Protocol (online only) and patient consent form at each site before study initiation. All patients provided written informed consent.

End Points and Assessments

The primary end point was PFS determined by blinded independent central review (BICR) or death due to any cause (whichever occurred first). Tumor progression or response was assessed by computed tomography or magnetic resonance imaging. Secondary end points included 9- and 12-month PFS rates, overall survival (OS), objective response rate (ORR), and duration of response (DOR) according to RECIST v1.1 by BICR. Exploratory end points included PFS and OS subgroup analyses by baseline tumor PD-L1 status, PFS2 (PFS after next line of therapy), safety and tolerability in all treated patients, and patient-reported outcomes assessed using the EQ-5D-3L questionnaire. Post hoc analyses included PFS and OS by sensitizing/nonsensitizing mutation, by lines of previous EGFR TKI, as well as by baseline patient and disease characteristics.

EGFR mutations were determined by local laboratory testing. Patients could have multiple EGFR tumor mutations. For subtype analyses, the classic sensitizing EGFR mutations were exon 19 deletions or L858R mutations.²⁹ All other EGFR mutations were classified as nonsensitizing (including L861Q, exon 20 insertion, G719X, and S768I). Although T790M mutation is known to be sensitive to osimertinib, such patients were eligible for enrollment only after a protocol amendment in November 2018. Patients with T790M-mutant tumors without another sensitizing mutation were categorized in the nonsensitizing mutation subgroup in the final EGFR mutation subanalyses. T790M-negative status was confirmed by a central or local laboratory using the cobas EGFR Mutation Test v2 (US-IVD; Roche Molecular Systems, Pleasanton, CA) on tumor tissue samples; T790M testing was not required in patients with previous osimertinib treatment.

Safety was assessed in all treated patients within 100 days after the last dose received. Adverse events (AEs) were graded by the National Cancer Institute Common Terminology Criteria for Adverse Events v4.0 (Data Supplement).

Statistical Analyses

The original planned sample size of 500 patients across the nivolumab plus chemotherapy and chemotherapy arms provided 90% power to detect an average hazard ratio (HR) of 0.735. Because of slow accrual and the COVID-19 pandemic, the sample size was reduced to approximately 270 patients in the protocol amendment effective on July 30, 2020, to provide 83% power to detect an average HR of 0.692 at alpha = .05 for PFS. If PFS was statistically significant for nivolumab plus chemotherapy versus chemotherapy, OS would then be tested at alpha = .05 in hierarchical order. Final analysis of PFS was to be conducted when 233 PFS events were reported and/or at a minimum follow-up of 6 months. Final analysis occurred at a minimum follow-up of 18.2 months, at which point only 212 actual PFS events had occurred in 294 patients, further decreasing the power from 83% to 76%.

PFS was analyzed with a stratified two-sided log-rank test. The HR and corresponding two-sided 95% CI were estimated using a Cox proportional hazards model, with treatment group as a single covariate, stratified by the study stratification factors. PFS and OS were estimated using Kaplan-Meier methodology, and a two-sided 95% CI for median PFS was calculated with the log-log transformation method. OS comparison was descriptive, as superiority was not demonstrated in PFS. ORR analysis was performed at the time of PFS analysis. An estimate of the difference in ORRs and corresponding 95% CI was calculated using stratified Cochran-Mantel-Haenszel methodology; ORRs and corresponding 95% exact two-sided CIs were calculated using the Clopper-Pearson method. DOR curves were estimated using the Kaplan-Meier product-limit method for subjects with complete response (CR) or partial response by BICR. Post hoc and exploratory analyses for additional subgroups were performed using an unstratified Cox proportional hazards regression model; analyses were descriptive and not adjusted for multiplicity. Descriptive statistics of safety were presented by treatment group.

RESULTS

Patients and Treatment

In the CheckMate 722 study, 367 patients were randomly assigned between March 2017 and June 2020 at 83 sites in nine countries (China, France, Hong Kong, Japan, South Korea, Singapore, Spain, Taiwan, and the United States; Fig 1). Of these patients, 294 were randomly assigned to the nivolumab plus chemotherapy arm (n = 144) and chemotherapy arm (n = 150), and the 73 patients assigned to nivolumab plus ipilimumab arm, which was closed early, were not evaluated in this final analysis. Overall, 141 (97.9%) patients in the nivolumab plus chemotherapy arm and 143 (95.3%) in the chemotherapy arm were treated.

FIG 1.

CONSORT diagram of patient disposition. ^aIncludes 73 patients randomly assigned to the nivolumab plus ipilimumab arm that was closed during enrollment. Database lock on April 25, 2022, with a minimum follow-up of 18.2 months and a median follow-up of 38.1 months. AE, adverse event.

Patient baseline characteristics were generally balanced between treatment arms (Table 1). Patients were predominantly Asian, with stage IV disease and adenocarcinoma. Approximately half of the patients in both treatment arms had tumor PD-L1 ≥1% and, of these, about 20% in each arm had tumor PD-L1 ≥50%. Most patients had tumors harboring EGFR-sensitizing mutations, with the most common being exon 19 deletions (55.6% in the nivolumab plus chemotherapy arm and 54.7% in the chemotherapy arm); nonsensitizing EGFR mutations occurred in <7% of patients in both arms. T790M mutation was reported in 6.3% of patients receiving nivolumab plus chemotherapy and 6.0% of patients receiving chemotherapy. Previous osimertinib use was reported in 22.2% and 24.0%, respectively.

TABLE 1.

Baseline Characteristics of All Randomly Assigned Patients

Patient/Disease Characteristic	Nivolumab Plus Chemotherapy (n = 144)	Chemotherapy (n = 150)
Age, years, median (range)	64 (29-80)	61 (35-86)
Male, No. (%)	61 (42.4)	56 (37.3)
Race, No. (%)
Asian	136 (94.4)	139 (92.7)
White	7 (4.9)	11 (7.3)
Other	1 (0.7)	0
ECOG performance status, No. (%)
0	56 (38.9)	41 (27.3)
1	88 (61.1)	109 (72.7)
Smoking status, No. (%)
Never	88 (61.1)	94 (62.7)
Current/former	56 (38.9)	56 (37.3)
Stage (at study entry), No. (%)
Stage IV	133 (92.4)	129 (86.0)
Recurrent	11 (7.6)	21 (14.0)
Histology, No. (%)
Adenocarcinoma	141 (97.9)	148 (98.7)
Squamous cell carcinoma	1 (0.7)	0
Other	2 (1.4)	2 (1.3)
Brain metastases, No. (%)
Presence	51 (35.4)	51 (34.0)
Absence	93 (64.6)	99 (66.0)
Liver metastases, No. (%)
Presence	17 (11.8)	30 (20.0)
Absence	127 (88.2)	120 (80.0)
Tumor PD-L1 expression,a,b No. (%)
<1%	54 (37.5)	60 (40.0)
≥1%	72 (50.0)	76 (50.7)
1%-49%	42 (29.2)	37 (24.7)
≥50%	30 (20.8)	39 (26.0)
Not evaluable	17 (11.8)	14 (9.3)
Baseline EGFR mutation,c,d No. (%)
Classic sensitizing mutationse
Exon 19 deletion	80 (55.6)	82 (54.7)
L858R	54 (37.5)	58 (38.7)
Nonsensitizing mutations
T790M	9 (6.3)	9 (6.0)
L861Q	6 (4.2)	2 (1.3)
Exon 20 insertion	5 (3.5)	2 (1.3)
G719X	2 (1.4)	6 (4.0)
S768I	2 (1.4)	4 (2.7)
Other	7 (4.9)	7 (4.7)
Line of previous EGFR TKI, No. (%)
1	121 (84.0)	127 (84.7)
2	23 (16.0)	21 (14.0)
3	0	2 (1.3)
Previous osimertinib use, No. (%)
Yes	32 (22.2)	36 (24.0)
No	112 (77.8)	114 (76.0)
Line of therapy (osimertinib), No. (%)
First line	10 (6.9)	14 (9.3)
Second line or third line	22 (15.3)	22 (14.7)

Abbreviations: ECOG, Eastern Cooperative Oncology Group; EGFR, epidermal growth factor receptor; TKI, tyrosine kinase inhibitor.

^a Determined by the PD-L1 IHC 28-8 pharmDx assay (Dako).

^b Tumor PD-L1 expression was not reported for one patient in the nivolumab plus chemotherapy arm.

^c Per local laboratory, patients could have multiple EGFR mutations.

^d Type of EGFR mutation was not reported for one patient in the nivolumab plus chemotherapy arm.

^e Defined as exon 19 deletions or L858R mutations.

At database lock (April 25, 2022), the minimum and median follow-ups were 18.2 and 38.1 months, respectively, and 212 actual PFS events were reported in the 294 randomly assigned patients. The median (range) duration of therapy was 4.9 (0.0-41.8) months with nivolumab plus chemotherapy and 3.8 (0.0-30.8) months with chemotherapy (Data Supplement, Table S1). The median (range) number of doses of chemotherapy received was similar for patients treated with nivolumab plus chemotherapy versus chemotherapy (pemetrexed, 7.0 [1.0-60.0] v 6.0 [1.0-44.0]; cisplatin, 4.0 [2.0-4.0] v 4.0 [1.0-4.0]; carboplatin, 4.0 [1.0-4.0] v 4.0 [1.0-4.0]). By database lock, most patients in both arms had discontinued treatment, mainly because of disease progression (Fig 1).

Efficacy

The median PFS was 5.6 months (95% CI, 4.5 to 6.8) with nivolumab plus chemotherapy versus 5.4 months (95% CI, 4.4 to 5.6) with chemotherapy. The primary end point of PFS with nivolumab plus chemotherapy versus chemotherapy was not met (HR, 0.75 [95% CI, 0.56 to 1.00]; P = .0528; Fig 2A). The 9-, 12-, and 18-month PFS rates with nivolumab plus chemotherapy versus chemotherapy were 25.9% versus 19.8%, 21.2% versus 15.9%, and 14.8% versus 5.9%, respectively. At database lock, 24.3% of patients in the nivolumab plus chemotherapy arm and 31.3% of patients in the chemotherapy arm were censored for PFS; most were censored upon receiving subsequent anticancer therapy before BICR-confirmed disease progression (16.0% with nivolumab plus chemotherapy and 19.3% with chemotherapy).

FIG 2.

PFS in all randomly assigned patients. (A) PFS by BICR and (B) PFS subgroup analysis with stratification factors in bold. PFS by BICR in patients with (C) sensitizing EGFR mutations, (D) one line of previous EGFR TKI therapy, and (E) sensitizing EGFR mutations and one line of previous EGFR TKI therapy. ^aStratified HR, 0.75. ^bRace was reported as other for one patient in nivolumab plus chemotherapy arm. ^cTumor PD-L1 expression was not evaluable in 31 patients (17 in nivolumab plus chemotherapy arm, and 14 in chemotherapy arm), and not reported for one patient in nivolumab plus chemotherapy arm. ^dType of EGFR mutation was not reported in one patient in nivolumab plus chemotherapy arm. ^eIncludes exon 19 deletion and L858R mutation. ^fWithout any sensitizing mutations (T790M, L861Q, exon 20 insertion, G719X, S768I, and other). ^gTwo patients in the chemotherapy arm had received three lines of previous EGFR TKI. BICR, blinded independent central review; ECOG PS, Eastern Cooperative Oncology Group performance status; EGFR, epidermal growth factor receptor; HR, hazard ratio; PFS, progression-free survival; TKI, tyrosine kinase inhibitor.

Exploratory and post hoc analyses by baseline characteristics showed no notable differences in median PFS across treatment arms in most subgroups (Fig 2B). However, in a post hoc analysis in patients with tumors harboring sensitizing EGFR mutations or who received one line of previous EGFR TKI, there was a trend of benefit in favor of nivolumab plus chemotherapy versus chemotherapy (Figs 2C and 2D). In these two subgroups, there was a trend of PFS benefit favoring nivolumab plus chemotherapy. For patients with tumors harboring EGFR-sensitizing mutations, median PFS was 5.6 versus 5.4 months, respectively (HR, 0.72 [95% CI, 0.54 to 0.97]; Fig 2C), and for patients who had received one line of previous EGFR TKI, median PFS was 5.6 versus 5.4 months, respectively (HR, 0.72 [95% CI, 0.54 to 0.97]; Fig 2D). As for patients with both tumor EGFR-sensitizing mutations and one line of previous EGFR TKI, median PFS was 6.3 versus 5.4 months, respectively (HR, 0.64 [95% CI, 0.47 to 0.88]; Fig 2E). Survival analysis by tumor PD-L1 expression showed no differences in median PFS between the two arms in patients with tumor PD-L1 <1%, ≥1%, and 1%-49% (Figs 3A-3C), while patients in the nivolumab plus chemotherapy arm with tumor PD-L1 ≥50% had a numerically improved PFS (HR, 0.65 [95% CI, 0.36 to 1.15]; Fig 3D).

FIG 3.

PFS by BICR in patients with (A) tumor PD-L1 <1%, (B) tumor PD-L1 ≥1%, (C) tumor PD-L1 1%-49%, and (D) tumor PD-L1 ≥50%. BICR, blinded independent central review; HR, hazard ratio; PFS, progression-free survival.

Median OS was 19.4 months (95% CI, 16.1 to 21.0) with nivolumab plus chemotherapy versus 15.9 months (95% CI, 14.0 to 18.8) with chemotherapy (HR, 0.82 [95% CI, 0.61 to 1.10]), with 18-month OS rates of 54% and 46%, respectively (Fig 4). OS was similar between treatments across exploratory subgroup analyses by baseline characteristics (Data Supplement, Fig S2). In patients with tumors harboring sensitizing EGFR mutations and patients with one line of previous EGFR TKI in a post hoc analysis, OS was also similar between treatment arms (Data Supplement, Fig S3). The ORR was 31.3% (95% CI, 23.8 to 39.5) with nivolumab plus chemotherapy including three patients with CR, versus 26.7% (95% CI, 19.8 to 34.5) with chemotherapy and no CR; the odds ratio was 1.32 (95% CI, 0.78 to 2.23) between arms (Fig 5). The median DOR was 6.7 months (95% CI, 4.2 to 12.4) and 5.6 months (95% CI, 4.1 to 9.9), respectively. ORR by tumor PD-L1 expression is summarized in the Data Supplement (Table S2).

FIG 4.

OS in all randomly assigned patients. HR, hazard ratio; OS, overall survival.

FIG 5.

ORR by BICR and DOR in all randomly assigned patients. BICR, blinded independent central review; DOR, duration of response; ORR, objective response rate.

Subsequent EGFR/ALK TKI therapy was administered in 35.4% and 31.3% of patients in the nivolumab plus chemotherapy and chemotherapy arms, respectively (Data Supplement, Table S3). Median PFS2 was 12.0 months for the nivolumab plus chemotherapy arm and 11.2 months for the chemotherapy arm (HR, 0.75 [95% CI, 0.57 to 0.99]).

Safety

Any-grade and grade 3/4 treatment-related AEs (TRAEs) were reported in 85.1% and 44.7% of patients in the nivolumab plus chemotherapy arm and 86.7% and 29.4% in the chemotherapy arm, respectively (Table 2). The most common any-grade TRAEs in both arms were anemia (39.7% with nivolumab plus chemotherapy and 35.0% with chemotherapy) and nausea (31.2% and 35.0%); the most frequent grade 3/4 events were anemia (15.6% and 9.1%) and decreased neutrophil count (11.3% and 11.2%). Any-grade TRAEs leading to treatment discontinuation occurred in 14.9% and 7.7% of patients, respectively (Table 2). Serious TRAEs of any grade were reported in 19.1% of patients treated with nivolumab plus chemotherapy and in 9.1% of patients treated with chemotherapy. In both arms, there were two treatment-related deaths (nivolumab plus chemotherapy: one due to pneumonitis and one due to interstitial lung disease; chemotherapy: one due to pneumonia and one due to interstitial lung disease).

TABLE 2.

Safety Summary and TRAEs in All Treated Patients

TRAE	Nivolumab Plus Chemotherapy (n = 141), No. (%)		Chemotherapy (n = 143), No. (%)
	Any Grade	Grade 3/4	Any Grade	Grade 3/4
Any TRAEsa	120 (85.1)	63 (44.7)	124 (86.7)	42 (29.4)
TRAEs leading to discontinuationa	21 (14.9)	9 (6.4)	11 (7.7)	4 (2.8)
Serious TRAEsa	27 (19.1)	23 (16.3)	13 (9.1)	9 (6.3)
Treatment-related deathsb	2 (1.4)c		2 (1.4)d
Any TRAE occurring in ≥10% of patients in either treatment arm
Anemia	56 (39.7)	22 (15.6)	50 (35.0)	13 (9.1)
Nausea	44 (31.2)	3 (2.1)	50 (35.0)	4 (2.8)
Decreased neutrophil count	40 (28.4)	16 (11.3)	42 (29.4)	16 (11.2)
Decreased WBC count	36 (25.5)	12 (8.5)	34 (23.8)	7 (4.9)
Increased ALT	29 (20.6)	6 (4.3)	22 (15.4)	3 (2.1)
Increased AST	28 (19.9)	0	23 (16.1)	0
Vomiting	27 (19.1)	4 (2.8)	15 (10.5)	1 (0.7)
Decreased appetite	27 (19.1)	3 (2.1)	38 (26.6)	3 (2.1)
Fatigue	21 (14.9)	3 (2.1)	13 (9.1)	1 (0.7)
Constipation	19 (13.5)	0	34 (23.8)	0
Decreased platelet count	18 (12.8)	7 (5.0)	28 (19.6)	4 (2.8)
Neutropenia	16 (11.3)	7 (5.0)	5 (3.5)	2 (1.4)
Increased blood creatinine	15 (10.6)	0	10 (7.0)	0
Pyrexia	15 (10.6)	0	3 (2.1)	0
Malaise	13 (9.2)	0	15 (10.5)	0

Abbreviation: TRAE, treatment-related adverse event.

^a Includes patients with TRAEs reported between first dose and 30 days after last dose of study therapy.

^b Includes treatment-related deaths reported between first day and 100 days after last dose of study therapy.

^c One due to interstitial lung disease and one due to pneumonitis.

^d One due to interstitial lung disease and one due to pneumonia.

In patients treated with nivolumab plus chemotherapy, the incidence of any-grade immune-mediated AEs (IMAEs) was ≤3% (Data Supplement, Table S4). The most common any-grade IMAEs were hypothyroidism/thyroiditis (2.8%), pneumonitis (2.1%), and hyperthyroidism (1.4%). Three patients reported grade 3/4 IMAEs, including pneumonitis (n = 2) and hypothyroidism/thyroiditis (n = 1).

Patient-Reported Outcomes

Completion rates for the EQ-5D-3L questionnaires for patient-reported quality of life were ≥85% at baseline through week 72 for both arms (Data Supplement, Table S5). Baseline mean EQ-5D-3L visual analog scale (VAS) scores were similar between treatment arms: 72.8% in the nivolumab plus chemotherapy arm and 73.9% in the chemotherapy arm. During on-treatment assessments, mean change from baseline EQ-5D-3L VAS scores were similar for both arms and fluctuated close to baseline, with no changes from baseline exceeding the minimum important difference (Data Supplement, Fig S4).

DISCUSSION

CheckMate 722 is a randomized phase III study comparing immunotherapy (nivolumab) plus chemotherapy with chemotherapy in patients with EGFR-mutated metastatic NSCLC after disease progression on EGFR TKIs. The study did not demonstrate a statistically significant result in the primary end point of PFS in the intention-to-treat patient population (median, 5.6 v 5.4 months, respectively; HR, 0.75; P = .0528). Treatment with nivolumab plus chemotherapy resulted in numerically prolonged median OS versus chemotherapy (19.4 v 15.9 months), while ORR and DOR were similar in both treatment arms. The safety profile of nivolumab plus chemotherapy was consistent with previous studies.^26,30,31

Similar to clinical outcomes from CheckMate 722, other phase III studies evaluating immunotherapies in patients with EGFR-mutated metastatic NSCLC also did not find a significant survival benefit with immunotherapy plus chemotherapy. Recently, in KEYNOTE-789, pembrolizumab plus chemotherapy prolonged PFS and OS but did not demonstrate a statistically significant survival benefit in patients with EGFR-mutated metastatic NSCLC after progression on previous TKI,³² while a subgroup analysis of 44 patients with EGFR or ALK genomic alterations from IMpower130 showed marginal improvement with atezolizumab plus chemotherapy.³³ A combination of antiangiogenic agents, such as bevacizumab with immunotherapy and chemotherapy, has also shown promising results in patients with EGFR-mutated NSCLC, such as the IMpower150 trial, in which a combination of atezolizumab with bevacizumab and chemotherapy showed improved OS compared with bevacizumab plus chemotherapy.³⁴ Also, addition of a bevacizumab biosimilar, IBI305, to the anti–PD-1 antibody sintilimab plus chemotherapy resulted in significant PFS improvement in the ORIENT-31 trial.³⁵ Furthermore, in ORIENT-31, the sintilimab plus chemotherapy arm showed a significant PFS benefit (HR of 0.72),³⁶ similar to that observed in the CheckMate 722 study, although PFS benefit in CheckMate 722 was not significant. However, cross-trial comparisons cannot be made due to differences in study designs and patient populations in these studies. The ongoing IMpower 151 study may help elucidate the benefit of antiangiogenic drugs in combination with immunotherapy and chemotherapy.

Notably, in the CheckMate 722 study, a post hoc analysis showed a trend of PFS benefit in favor of nivolumab plus chemotherapy versus chemotherapy in patients with tumors harboring sensitizing EGFR mutations (HR, 0.72 [95% CI, 0.54 to 0.97]) and those who received one line of previous EGFR TKIs (HR, 0.72 [95% CI, 0.54 to 0.97]). Although these results are not protected against the possibility of being random, a similar exploratory analysis in patients with tumors harboring sensitizing EGFR mutations and previous EGFR TKI experience treated with atezolizumab in combination with chemotherapy and bevacizumab also showed an OS benefit compared with bevacizumab plus chemotherapy; however, atezolizumab plus chemotherapy without addition of bevacizumab did not show any benefit.³⁷ Taken together, the optimal combinations of immunotherapy and chemotherapy with or without antiangiogenic agents are yet to be determined for patients with EGFR-mutated metastatic NSCLC who progressed on TKI therapy.

Tumor PD-L1 expression has been studied as a predictive biomarker of benefit with anti–PD-1/PD-L1 checkpoint inhibitors as single agents, dual immunotherapy combinations, and in combination with chemotherapy in patients with metastatic NSCLC without oncogenic driver mutations in several phase III studies.^17-20,38-40 Although EGFR TKIs may affect PD-L1 expression,⁴¹ the available information on the ability of PD-L1 to predict outcomes after immunotherapy in patients with EGFR-mutated NSCLC who progressed on EGFR TKIs is limited.⁴² In the CheckMate 722 study, a trend toward improved PFS with nivolumab plus chemotherapy was seen in patients with tumor PD-L1 ≥50%, but the small subgroup size limited definitive data interpretation. Recently, a retrospective analysis reported that OS improvements in this patient subgroup were observed with pembrolizumab plus chemotherapy and pembrolizumab plus anlotinib (an antiangiogenic agent) compared with pembrolizumab monotherapy.⁴³ Additional prospective research is needed to further assess the role of PD-L1 as a predictive biomarker in patients with EGFR-mutated metastatic NSCLC. Further research exploring resistance mechanisms and biomarkers for predicting response to immunotherapy plus chemotherapy in this patient population is needed.

In the CheckMate 722 study, no new safety concerns were identified, although a higher rate of grade 3/4 TRAEs was observed in the nivolumab plus chemotherapy arm versus the chemotherapy arm. The incidence of IMAEs in the nivolumab plus chemotherapy arm was low. Among IMAEs, pneumonitis of any grade was reported in three (2.1%) patients, similar to the rates of pneumonitis reported in previous studies after nivolumab monotherapy in patients with previously treated metastatic NSCLC.^44,45 Although a higher rate of TRAEs was observed for patients in the nivolumab plus chemotherapy arm than in the chemotherapy arm, there was no negative impact on patient-reported quality of life.

This study has limitations that may have affected the outcomes. Because of enrollment challenges compounded by the COVID-19 pandemic and lower PFS event numbers than expected at final analysis, the CheckMate 722 study was underpowered to ascertain a significant treatment effect with nivolumab plus chemotherapy versus chemotherapy in the intention-to-treat population. Furthermore, at the beginning of the CheckMate 722 trial, patients were excluded if they had tumors harboring T790M mutations or were previously treated with osimertinib after disease progression following first- or second-generation EGFR TKIs. However, the treatment landscape for patients with EGFR-mutated metastatic NSCLC changed significantly after initiation of this trial with the approval of osimertinib in first line.^10,46 Although a protocol amendment allowed for enrollment of patients treated with osimertinib regardless of T790M mutations, the proportion of patients who received first-line osimertinib or with tumors harboring T790M mutations was lower than the current real-world scenario among patients with EGFR-mutated NSCLC.⁴⁷ Thus, translating the outcomes of this study to a real-world population in the clinic is limited.

In summary, nivolumab plus chemotherapy did not significantly improve PFS in patients with EGFR-mutated metastatic NSCLC previously treated with EGFR TKI therapy compared with chemotherapy. No new safety signals were observed with nivolumab plus chemotherapy.

DATA SHARING STATEMENT

Data are available upon reasonable request. BMS policy on data sharing may be found at https://www.bms.com/researchers-and-partners/independent-research/data-sharing-request-process.html.

AUTHOR CONTRIBUTIONS

Conception and design: Tony Mok, Kazuhiko Nakagawa, Keunchil Park, Yuichiro Ohe, Nicolas Girard, Yi-Long Wu, Justin Gainor, Ann Forslund, James Chih-Hsin Yang

Administrative support: Kazuhiko Nakagawa, Lin Wu

Provision of study materials or patients: Kazuhiko Nakagawa, Keunchil Park, Yuichiro Ohe, Nicolas Girard, Yi-Long Wu, Sang-We Kim, Cheng-Ta Yang, Lin Wu, Meng-Chih Lin, Jens Samol, Mengzhao Wang, James Chih-Hsin Yang

Collection and assembly of data: Tony Mok, Kazuhiko Nakagawa, Keunchil Park, Yuichiro Ohe, Nicolas Girard, Hye Ryun Kim, Yi-Long Wu, Justin Gainor, Chao-Hua Chiu, Cheng-Ta Yang, Chien Liang Wu, Lin Wu, Meng-Chih Lin, Jens Samol, Kazuya Ichikado, Mengzhao Wang, Xiaoqing Zhang, Judi Sylvester, Ann Forslund, James Chih-Hsin Yang

Data analysis and interpretation: Tony Mok, Kazuhiko Nakagawa, Keunchil Park, Yuichiro Ohe, Nicolas Girard, Hye Ryun Kim, Justin Gainor, Se-Hoon Lee, Sang-We Kim, Meng-Chih Lin, Jens Samol, Xiaoqing Zhang, Judi Sylvester, Sunney Li, Ann Forslund, James Chih-Hsin Yang

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

Nivolumab Plus Chemotherapy in Epidermal Growth Factor Receptor–Mutated Metastatic Non–Small-Cell Lung Cancer After Disease Progression on Epidermal Growth Factor Receptor Tyrosine Kinase Inhibitors: Final Results of CheckMate 722

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. 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/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).