Pembrolizumab Plus Chemo Does Not Significantly Improve Survival in TKI-resistant, EGFR+ NSCLC

The addition of pembrolizumab (Keytruda) to pemetrexed and platinum-based chemotherapy resulted in a numerical, but not statistically significant, improvement in progression-free survival (PFS) or overall survival (OS) vs chemotherapy plus placebo in patients with TKI-resistant, EGFR-mutated, metastatic nonsquamous non–small cell lung cancer (NSCLC), according to data from the phase 3 KEYNOTE-789 trial (NCT03515837) presented at the 2023 ASCO Annual Meeting.

Findings from the second interim analysis of the trial showed that at a median follow-up of 28.6 months (range, 16.0-40.4), pembrolizumab plus chemotherapy (n = 245) resulted in a median PFS of 5.6 months (95% CI, 5.5-5.8) by blinded independent central review (BICR) compared with 5.5 months (95% CI, 5.4-5.6) for placebo plus chemotherapy (n = 247; HR, 0.80; 95% CI, 0.65-0.97; P = .0122), missing the prespecified efficacy boundary (P = .0117) for PFS. The 24-month PFS rate was 4.7% for the pembrolizumab regimen and 3.5% for the placebo regimen.

At a median follow-up of 42.0 months (range, 29.5-53.9) for the final analysis, patients in the pembrolizumab arm experienced a median OS of 15.9 months (95% CI, 13.7-18.8) vs 14.7 months (95% CI, 12.7-17.1) for patients in the placebo group (HR, 0.84; 95% CI, 0.69-1.02; P = .0362); this did not cross the prespecified boundary for efficacy (P = .0117).

The 36-month OS rate was 14.6% and 11.4% for the pembrolizumab and placebo arms, respectively.

“Results were consistent with prior findings for [patients with] TKI-resistant, EGFR-mutant, metastatic NSCLC, [who] derive less benefit from anti–PD-1 or –PD-L1 treatment [vs] EGFR wild-type patients,” lead study author James Chih-Hsin Yang, MD, PhD, of the National Taiwan University Hospital and National Taiwan University Cancer Center in Taipei, Taiwan, said in a presentation of the data. “Yet, there [was] a biomarker that was identified, which is PD-L1 [tumor proportion score (TPS)] more than 1%, which deserves further investigation in the future.”

KEYNOTE-789 enrolled patients with histologically or cytologically confirmed stage IV nonsquamous NSCLC who harbored an EGFR exon 19 deletion or L858R mutation and had an ECOG performance status of 0 or 1. Patients needed to have progressive disease by RECIST v1.1 criteria after a first- or second-generation EGFR TKI for those without a EGFR T790M mutation; after a first- or second-generation EGFR TKI and osimertinib (Tagrisso) for those with an EGFR T790M mutation; or after osimertinib as first-line therapy, irrespective of T790M status.

Patients were randomly assigned 1:1 to receive 200 mg of pembrolizumab or matching placebo in combination with 500 mg/m2 of pemetrexed plus area under the curve 5 of carboplatin or 75 mg/m2 of cisplatin once every 3 weeks for 4 cycles. Patients then received 200 mg of pembrolizumab or placebo in combination with 500 mg/m2 of pemetrexed once every 3 weeks for 31 cycles. Notably, those in the placebo arm who experienced progressive disease were allowed to cross over to receive 200 mg of pembrolizumab once every 3 weeks for up to 35 cycles.

Patients were stratified by PD-L1 TPS (<50% vs ≥50%), treatment history (osimertinib vs no osimertinib), and geographic region (East Asia vs not East Asia).

BICR-assessed PFS by RECIST v1.1 criteria and OS served as the trial’s dual primary end points. Secondary end points consisted of BICR-assessed overall response rate (ORR) and duration of response (DOR) per RECIST v1.1 criteria, as well as safety and patient-reported outcomes.

The study had a power of approximately 91% to detect a hazard ratio of 0.7 for PFS and approximately 86% to detect a hazard ratio of 0.72 for OS with 492 patients at 24 or more months of enrollment. The first and second interim analyses were conducted approximately 6 months and 16 months after the last patient was randomly assigned, respectively. The final analysis occurred about 29 months after the final patient was randomized. The data cutoff dates for the second interim analysis and final analysis were December 3, 2021, and January 17, 2023, respectively. P values were 1-sided.

At the final data cutoff, 3 patients in the pembrolizumab arm completed treatment, and 2 patients remained on treatment. A total of 240 patients discontinued treatment. Reasons for discontinuation included radiographic or clinical progressive disease (n = 209), adverse effects (AEs; n = 21), patient withdrawal (n = 9), or physician decision (n = 1). Eighteen patients (7.3%) went on to receive anti–PD-1/L1 therapy, and 96 patients (39.2%) received a subsequent TKI.

In the placebo arm, 2 patients completed treatment, and 1 remained on treatment at data cutoff. A total of 243 patients discontinued treatment, with the most common reason for discontinuation being radiographic or clinical progressive disease (n = 219), followed by AEs (n = 18), patient withdrawal (n = 5), and excluded medication (n = 1). Eighty-eight patients (35.6%) received subsequent anti–PD-1/L1 therapy, and 101 (40.9%) were given a subsequent TKI.

Yang said baseline characteristics were generally similar between the 2 arms, although he noted that the pembrolizumab arm had a higher rate of patients with an ECOG performance status of 1 (71.0%) compared with the placebo arm (62.8%).

The median age of patients in the experimental and control arms was 63 years (range, 34-87). Most patients were female (62.0% for pembrolizumab and 61.1% for placebo), enrolled in Asia (61.2% and 60.7%), had adenocarcinoma (97.6% and 98.4%), and received carboplatin with pemetrexed (84.9% and 82.2%). Rates of patients who were current or former smokers (34.3% and 33.6%), had a PD-L1 TPS of at least 50% (21.2% and 20.6%), and had brain metastases at baseline (20.8% and 19.0%) were comparable.

Prior therapy included neoadjuvant therapy (0.4% and 1.6% for pembrolizumab and placebo, respectively), adjuvant therapy (5.3% and 4.9%), and radiation (33.1% and 37.2%). Moreover, patients also previously received treatment with a TKI except for osimertinib (52.2% and 51.0%), first-line osimertinib (11.4% and 13.4%), second-line osimertinib (35.9% and 35.6%), or other TKI use (0.4% and 0%).

EGFR mutational status consisted of L858R (42.0% for the pembrolizumab arm and 41.3% for the placebo arm), exon 19 deletion (56.7% and 57.5%), and both exon 19 deletion and L858R (0.8% and 0.8%). In the pembrolizumab arm, 38.8% of patients harbored EGFR T790M mutations and 52.7% did not; those rates were 35.2% and 56.7%, respectively, in the placebo arm. EGFR T790M testing was not done in 8.6% of patients in the experimental arm and 8.1% of those in the control arm.

Additional data showed that in patients with a PD-L1 TPS of at least 1%, pembrolizumab plus chemotherapy (n = 101) resulted in a median OS of 18.6 months (95% CI, 12.5-22.9) compared with 14.1 months (95% CI, 11.4-18.1) with placebo plus chemotherapy (n = 123; HR, 0.77; 95% CI, 0.58-1.02). The 36-month OS rates for the pembrolizumab and placebo arms were 18.3% and 13.0%, respectively.

In those with a PD-L1 TPS of less than 1% who received pembrolizumab (n = 127), the median OS was 15.7 months (95% CI, 12.4-18.8) vs 14.7 months (95% CI, 12.2-17.3) for those who received placebo (n = 113; HR, 0.91; 95% CI, 0.70-1.19). The 36-month OS rates were 12.0% for pembrolizumab plus chemotherapy compared with 8.7% for placebo plus chemotherapy.

The ORR for all patients in the pembrolizumab arm was 29.0% (95% CI, 23.4%-35.1%) compared with 27.1% (95% CI, 21.7%-33.1%) for the placebo arm. Two percent of patients given pembrolizumab plus chemotherapy achieved a complete response (CR), 26.9% had a partial response (PR), 49.4% experienced stable disease, and 15.1% had progressive disease. Notably, 3.3% and 3.3% of patients were not evaluable for response or did not have a response assessment, respectively.

The rates of CR, PR, stable disease, and progressive disease in the placebo arm were 1.2%, 25.9%, 47.4%, and 21.1%, respectively. Two percent of patients were not evaluable for response, and 2.4% did not have a response assessment.

At the final analysis, the median DOR was 6.3 months (95% CI, 2.3 to 40.8+) for pembrolizumab plus chemotherapy and 5.6 months (95% CI, 1.8+ to 40.6+) for placebo plus chemotherapy. The 9-month DOR rates were 34.0% and 22.9%, respectively.

Regarding safety, rates of any-grade AEs were 97.6% for the pembrolizumab group and 98.0% for the placebo group; 55.9% and 58.1% of patients, respectively, experienced grade 3 to 5 AEs. AEs led to death in 5 patients (2.0%) and 12 patients (4.9%) in the pembrolizumab and placebo arms, respectively.

“There were no special AEs identified. When we look at the AE profile and the instances, there was [little] increase in [AEs when] adding pembrolizumab to chemotherapy,” Yang said.

Treatment-related AEs (TRAEs) of any grade occurred in 89.8% of patients in the pembrolizumab arm and 86.2% of those in the placebo arm. Rates of grade 3 to 5 TRAEs were 43.7% and 38.6% for the pembrolizumab and placebo regimens, respectively. Any-grade immune-mediated AEs were reported in 20.0% of patients in the pembrolizumab arm vs 8.1% of those in the placebo arm; these effects were grade 3 to 5 in 4.5% and 2.0% of patients, respectively.

AEs resulted in discontinuation of any treatment component in 16.3% of patients in the experimental arm. Rates of discontinuation of pembrolizumab, any chemotherapy, and all treatment components were 9.8%, 12.7%, and 2.9%, respectively. In the placebo arm, 11.8% of patients discontinued any treatment component due to AEs. In this group, rates of discontinuation of placebo, any chemotherapy, and all treatment components were 4.5%, 11.8%, and 2.0%, respectively.

The most common any-grade TRAEs reported in at least 10% of either treatment group included anemia (38% for pembrolizumab vs 40% for placebo), decreased neutrophil count (38% and 45%), nausea (35% and 41%), decreased white blood cell count (30% and 35%), fatigue (23% and 18%), increased alanine aminotransferase (22% and 21%), increased aspartate aminotransferase (22% and 20%), decreased platelet count (22% and 20%), decreased appetite (22% and 20%), constipation (19% and 17%), vomiting (15% and 14%), and asthenia (10% and 8%).

Rates of hypothyroidism and hyperthyroidism in the pembrolizumab arm were 5.7% and 5.3%, respectively; those rates were 2.4% and 1.2% in the placebo arm, respectively. Other any-grade immune-mediated AEs reported in the pembrolizumab and placebo arms, respectively, included infusion reactions (2.0% and 0.8%), pneumonitis (2.0% and 1.6%), adrenal insufficiency (0.8% and 0%), hepatitis (0.8% and 0.4%), myositis (0.8% and 0.4%), nephritis (0.8% and 0.4%), severe skin reactions (0.8% and 0%), thyroiditis (0.8% and 1.2%), colitis (0.4%, both), hypophysitis (0.4% and 0%), myelitis (0.4% and 0%), myocarditis (0.4% and 0%), and type 1 diabetes mellitus (0.4% and 0%).

“Additional biomarker research is needed to determine which patients will benefit from immune checkpoint inhibitor therapy in TKI-resistant, EGFR-mutant metastatic NSCLC, as there remains a great unmet need for this patient population,” Yang concluded.

Reference

Yang, JC-H, Lee DH, Lee JS, et al. Pemetrexed and platinum with or without pembrolizumab for tyrosine kinase inhibitor (TKI)-resistant, EGFR-mutant, metastatic nonsquamous NSCLC: phase 3 KEYNOTE-789 study. J Clin Oncol. 2023;41(suppl 17):LBA9000. doi:10.1200/JCO.2023.41.17_suppl.LBA9000