EGFR mutation subtypes and response to immune checkpoint blockade treatment in non-small-cell lung cancer

Abstract

Background: Although EGFR mutant tumors exhibit low response rates to immune checkpoint blockade overall, some EGFR mutant tumors do respond to these therapies; however, there is a lack of understanding of the characteristics of EGFR mutant lung tumors responsive to immune checkpoint blockade.

Patients and methods: We retrospectively analyzed de-identified clinical and molecular data on 171 cases of EGFR mutant lung tumors treated with immune checkpoint inhibitors from the Yale Cancer Center, Memorial Sloan Kettering Cancer Center, University of California Los Angeles, and Dana Farber Cancer Institute. A separate cohort of 383 EGFR mutant lung cancer cases with sequencing data available from the Yale Cancer Center, Memorial Sloan Kettering Cancer Center, and The Cancer Genome Atlas was compiled to assess the relationship between tumor mutation burden and specific EGFR alterations.

Results: Compared with 212 EGFR wild-type lung cancers, outcomes with programmed cell death 1 or programmed death-ligand 1 (PD-(L)1) blockade were worse in patients with lung tumors harboring alterations in exon 19 of EGFR (EGFRΔ19) but similar for EGFRL858R lung tumors. EGFRT790M status and PD-L1 expression did not impact response or survival outcomes to immune checkpoint blockade. PD-L1 expression was similar across EGFR alleles. Lung tumors with EGFRΔ19 alterations harbored a lower tumor mutation burden compared with EGFRL858R lung tumors despite similar smoking history.

Conclusions: EGFR mutant tumors have generally low response to immune checkpoint inhibitors, but outcomes vary by allele. Understanding the heterogeneity of EGFR mutant tumors may be informative for establishing the benefits and uses of PD-(L)1 therapies for patients with this disease.

Keywords: epidermal growth factor receptor; immune checkpoint blockade; non-small-cell lung cancer.

Figure 1.

Response, progression-free survival, and overall survival of EGFR^L858Rand EGFR^Δ19 mutant tumors to immune checkpoint blockade. (A) Response rate in tumors with EGFR^Δ19 (n = 76) or EGFR^L858R(n = 44) mutations, and wild-type for EGFR (WT) (n = 212). Overall response rate is indicated on each bar in white. Statistics were calculated using Fisher’s exact test. (B) Progression-free survival in tumors with EGFR^Δ19 (n = 77) (HR 0.449, 95% CI 0.338–0.595, log-rank P < 0.001) or EGFR^L858R(n = 44) (HR 0.578, 95% CI 0.412–0.811, log-rank P = 0.001) alterations compared with lung tumors that are EGFR wild-type (n = 212). (C) Overall survival in tumors with EGFR^Δ19 (n = 77) (HR 0.69, 95% CI 0.493–0.965, log-rank P = 0.03) or EGFR^L858R(n = 45) (HR 0.917, 95% CI 0.597–1.409, log-rank P = 0.69) alterations compared with lung tumors that are EGFR wild-type (n = 212). HR, hazard ratio; CI, confidence interval.

Figure 2.

Clinicopathologic features associated with response, progression-free survival, and overall survival of EGFR^L858Rand EGFR^Δ19 mutant tumors. (A) Response rate of tumors with 0–2 (n = 47) or ≥3 (n = 73) prior lines of therapy, P = 0.01. (B) Progression-free survival with 0–2 (n = 46) or ≥3 (n = 75) prior lines of therapy (HR 2.267, 95% CI 1.499–3.427, log-rank P < 0.001). (C) Overall survival with 0–2 (n = 48) or ≥3 (n = 74) prior lines of therapy (HR 1.845, 95% CI 1.204–2.826, log-rank P = 0.004). (D) Response rate in tumors harboring EGFR^T790M (T790M+, n = 52) or negative for EGFR^T790M (T790M−, n = 56) that had prior EGFR tyrosine kinase inhibitor (EGFR TKI) treatment, P = 0.21. (E) Progression-free survival in tumors harboring EGFR^T790M (n = 52) or negative for EGFR^T790M (n = 57) that had prior EGFR TKI treatment (HR 1.348, 95% CI 0.905–2.007, log-rank P = 0.15). (F) Overall survival in tumors harboring EGFR^T790M (n = 50) or negative for EGFR^T790M (n = 60) that had prior EGFR TKI treatment (HR 0.878, 95% CI 0.574–1.343, log-rank P = 0.55). (G) Response rate in tumors with <1% PD-L1 expression (n = 28) or ≥1% PD-L1 expression (n = 23), P > 0.99. (H) Progression-free survival in tumors with <1% PD-L1 expression (n = 29) or ≥1% PD-L1 expression (n = 22) (HR 1.370, 95% CI 0.761–2.466, log-rank P = 0.29). (I) Overall survival in tumors with <1% PD-L1 expression (n = 30) or ≥1% PD-L1 expression (n = 21) (HR 1.747, 95% CI 0.913–3.342, log-rank P = 0.084). Statistical analysis for response rate used Fisher’s exact test and statistical analysis for Kaplan–Meier plots used the log-rank test. CI, confidence interval; CR, complete response; HR, hazard ratio; PD-L1, programed death-ligand 1; PR, partial response; SD, stable disease; PD, progressive disease.

Figure 3.

Characterization of EGFR allele-specific tumor mutation burden (TMB) and smoking history. (A) TMB was calculated for EGFR mutant tumors harboring deletions in exon 19 [(Δ19) (n = 174)], mutations in exon 21 [L858R (n = 117) and L861Q (n = 12)], insertions in exon 20 [(20ins) (n = 21)], mutations in exon 18 [(G719) (n = 21)], or co-mutations at positions G719 and L861Q (n = 1). Data were combined from Memorial Sloan Kettering Cancer Center, the Yale Cancer Center, and The Cancer Genome Atlas cohorts. Data were transformed within each cohort to within-cohort percentile rank to permit unified analysis, and median TMB percentile rank is indicated. (B) TMB in EGFR^L858Rand EGFR^Δ19 mutant tumors from patients with ever (n = 99) or never (n = 129) smoking status (median 3.8 versus 3.1, P = 0.37). (C) Percentage of ever and never smokers within the EGFR^L858Rand EGFR^Δ19 mutant tumors groups (P = 0.14). (D) Pack-years in EGFR^L858Rand EGFR^Δ19 mutant tumors groups (P = 0.58). Statistics were calculated using the Fisher’s exact test.