Abstract
Optimal treatment of EGFR-mutated lung cancer after progression on osimertinib is unknown. We retrospectively evaluated outcomes in patients who received platinum doublet chemotherapy alone or in combination with immunotherapy or bevacizumab. We found that the addition of immunotherapy to chemotherapy was associated with worse survival; statistically significant differences in survival could not be detected with the addition of bevacizumab to chemotherapy.
Introduction:
Patients with EGFR-mutant lung cancer who have had disease progression on osimertinib commonly receive platinum doublet chemotherapy, but whether adding immunotherapy or bevacizumab provides additional benefit is unknown.
Materials and Methods:
This was a retrospective analysis at 2 university-affiliated institutions. Patients with EGFR-mutant lung cancer who had progression on osimertinib and received next-line therapy with platinum doublet chemotherapy (chemo), platinum doublet chemotherapy plus immunotherapy (chemo-IO), or platinum doublet chemotherapy plus bevacizumab (chemo-bev), were identified; patients who continued osimertinib with these regimens were included. Efficacy outcomes including duration on treatment (DOT) and overall survival (OS) from the start of chemotherapy were assessed. Associations of treatment regimen with outcomes were evaluated using adjusted Cox regression models, using pairwise comparisons between groups.
Results:
104 patients were included: 57 received chemo, 12 received chemo-IO, and 35 received chemo-bev. In adjusted models, patients who received chemo-IO had worse OS than did those who received chemo (hazard ratio (HR) 2.66, 95% CI 1.25–5.65; P= .011) or those who received chemo-bev (HR 2.37, 95% CI 1.09–5.65; P= .030). A statistically significant difference in OS could not be detected in patients who received chemo-bev versus those who received chemo (HR 1.50, 95% CI 0.84–2.69; P= .17).
Conclusion:
In this retrospective study, giving immunotherapy with platinum doublet chemotherapy after progression on osimertinib was associated with a worse OS compared with platinum doublet chemotherapy alone. Platinum doublet chemotherapy without immunotherapy (with consideration of continuation of osimertinib, in selected cases) is a reasonable choice in this setting, while we await results of clinical trials examining optimal next-line chemotherapy-based regimens in EGFR-mutant lung cancer.
Keywords: Targeted therapy, Driver mutations, Pembrolizumab, Atezolizumab, Chemo-immunotherapy
Introduction
Osimertinib is a third-generation EGFR tyrosine kinase inhibitor (TKI) that is now standard-of-care first-line therapy for patients with advanced EGFR-mutated non-small-cell lung cancer (NSCLC), based on the results of the FLAURA trial,1,2 which demonstrated improved progression-free survival and overall survival (OS) with osimertinib compared with first-generation EGFR TKIs. However, most patients will eventually experience systemic disease progression on osimertinib and require a change in systemic therapy. A recent review article provides an evidence-based algorithm to help guide therapy decisions after progression on osimertinib.3 Depending on the mechanism of resistance to osimertinib, another molecularly targeted therapy is sometimes an option, though this remains an experimental strategy. Clinical trials can also be considered, when available. However, the current standard-of-care next-line therapy after progression on first-line osimertinib is platinum doublet chemotherapy-based regimens. Sometimes osimertinib is continued when chemotherapy is started; this is another experimental strategy.4
Given the widespread use of chemoimmunotherapy combinations as first-line therapy for advanced NSCLC without oncogenic drivers, immunotherapy is sometimes given with platinum doublet chemotherapy to patients after progression on osimertinib. Similarly, older studies suggested benefit to chemotherapy and bevacizumab combinations in metastatic NSCLC. However, neither chemotherapy plus immunotherapy nor chemotherapy plus bevacizumab are well-studied in EGFR-mutant lung cancer.
The strongest data supporting the addition of bevacizumab to platinum doublet chemotherapy in NSCLC come from the phase III trial ECOG 4599, which showed improved survival with the combination of carboplatin/paclitaxel plus bevacizumab compared with carboplatin/paclitaxel alone (12.3 months vs. 10.3 months; HR 0.79, P= .003) 5; however, because this study was conducted before EGFR mutations were routinely tested, the study population was molecularly unselected, thus the efficacy of this regimen in EGFR-mutant NSCLC specifically is unknown. The addition of immunotherapy to platinum doublet chemotherapy gained immediate traction in the frontline treatment of metastatic NSCLC following the publication of the results of the phase III trial KEYNOTE-189, in which the addition of pembrolizumab to carboplatin/pemetrexed dramatically improved overall survival compared with carboplatin/pemetrexed alone in metastatic nonsquamous NSCLC (22.0 vs. 10.7 months; HR 0.56; 95% CI 0.45–0.70)6,7; however, this trial excluded patients with sensitizing EGFR or ALK mutations. The similar phase III trial IMpower130, which compared atezolizumab given in combination with carboplatin/nab-paclitaxel versus carboplatin/nab-paclitaxel alone, did include patients with EGFR mutations (provided they had experienced disease progression on TKI therapy), but they only comprised 6% of the population. The trial showed overall survival benefit in the intention-to-treat wild-type population (ie, those without EGFR or ALK mutations) (18.6 vs. 13.9 months; HR 0.79; 95% CI 0.64–0.98), but not in patients with EGFR or ALK mutations (HR 0.98; 95% CI 0.41–2.31).8 The randomized phase III trial IMpower150 trial, which included patients with EGFR mutations, found that the quadruplet regimen of carboplatin and paclitaxel plus atezolizumab and bevacizumab improved outcomes compared with carboplatin and paclitaxel plus bevacizumab, including in the EGFR mutant subset, and in fact the magnitude of benefit was especially impressive for the EGFR mutant subset (HR 0.31; 95% CI 0.11–0.83).9,10
In the present retrospective analysis, we examined outcomes in patients who were treated with platinum doublet chemotherapy alone or in combination with immunotherapy or bevacizumab after experiencing disease progression on osimertinib.
Methods
Study Population
Patients were identified by retrospective chart review at 2 participating institutions: Stanford Cancer Institute (n = 75) and Massachusetts General Hospital (n = 29). This study had institutional review board approval at both institutions. Patients were eligible for inclusion if they had a diagnosis of Stage IV or recurrent metastatic NSCLC with a sensitizing EGFR mutation identified by polymerase chain reaction (PCR) testing or next generation sequencing (NGS). They had to have received osimertinib and gone on to experience either disease progression or toxicity on osimertinib requiring a change in systemic therapy. Patients with biopsy-proven small cell transformation after progression on osimertinib were excluded. Immediately following osimertinib, patients had to have received platinum doublet chemotherapy, either alone or in combination with immunotherapy or bevacizumab as next line treatment. Patients who went on to switch to another targeted therapy (against EGFR or another target) immediately following osimertinib, prior to starting on chemotherapy, were excluded. Patients who continued taking osimertinib along with initiation of chemotherapy-based regimens were included.
Patients could have received prior neoadjuvant or adjuvant chemotherapy and/or radiation. They could have received prior EGFR TKI before osimertinib. They could have previously received a nonchemotherapy agent in combination with osimertinib (eg, necitumumab). They could have received up to 2 cycles of platinum-based chemotherapy in the metastatic setting – eg, if this was given before EGFR testing returned – assuming they did not have disease progression immediately following the 1 to 2 cycles. Patients could have received prior radiation in the early stage setting or in the metastatic setting for oligo-metastatic or oligo-progressive disease.
We collected patient demographics, tumor characteristics, treatment history, and disease course, through April 9, 2021. Only patients started on platinum doublet chemotherapy-based regimens prior to July 5, 2020 at Stanford, or December 4, 2020 at MGH, were included in this study. We recorded OS from the start date of first post-osimertinib chemotherapy, as well as duration on treatment (DOT, time from first dose to last dose) of first post-osimertinib chemotherapy.
Statistical Analysis
We calculated descriptive statistics for patient characteristics, stratified by treatment group: chemo, chemo-IO, and chemo-bev. We compared characteristics between groups using absolute standardized differences (ASD), a measure of the difference in means or proportions between groups expressed in units of standard deviations.11 Values of 0.2, 0.5, and 0.8 correspond to small, moderate, and large differences, respectively. We also compared characteristics between the 3 groups using ANOVA and χ2 tests for continuous and categorical variables, respectively. We used univariate analyses to assess associations between baseline factors (patient and tumor characteristics) and DOT and OS in the full cohort; associations were considered statistically significant if P value was < .05.
We used Kaplan-Meier curves to display the outcomes of DOT and OS from the start of first post-osimertinib chemotherapy, stratified by treatment group. To determine the association between treatment and outcomes, we fit multivariable Cox regression models, adjusting for ECOG performance status and brain metastases; these variables were selected based on having statistically significant associations with either DOT or OS or both in the univariate analyses. We performed pairwise comparisons between groups: chemo-IO versus chemo alone, chemo-bev versus chemo alone, and chemo-IO versus chemo-bev. We additionally performed sensitivity analyses of our Cox regression models, excluding patients who were continued on osimertinib in combination with their chemotherapy-based regimen. All analyses were conducted using R statistical software, Version 1.3.1056.
Results
We identified 104 patients who were eligible for study inclusion: 57 patients (55%) received chemotherapy alone (54 of whom received carboplatin/pemetrexed; 1 received carboplatin/paclitaxel; 1 received carboplatin/nab-paclitaxel; 1 received carboplatin/gemcitabine), 12 patients (12%) received chemotherapy plus immunotherapy (carboplatin/pemetrexed/pembrolizumab in all cases), and 35 patients (34%) received chemotherapy plus bevacizumab (carboplatin/pemetrexed/bevacizumab in all cases). One single additional patient between our 2 institutions received chemotherapy plus both immunotherapy and bevacizumab (carboplatin/paclitaxel/atezolizumab/bevacizumab) and was not included in our analyses. Of the 104 patients, 29 (28%) continued osimertinib when starting chemotherapy; this included 26 of 57 patients (46%) who received chemo, 1 of 12 patients (8%) who received chemo-IO, and 2 of 35 patients (6%) who received chemo-bev.
Baseline characteristics are summarized in Table 1. There were moderate absolute standardized differences between groups for ECOG performance status, PD-L1 expression (< 1% vs. ≥ 1%), and current anticoagulation use. There were small absolute standardized differences between groups for age, sex, race, Charlson Comorbidity Index, EGFR mutation subtype, prior therapy (osimertinib only vs. other EGFR TKI followed by osimertinib), brain metastases, liver metastases, and current steroid use. The only statistically significant between-group difference was for PD-L1 expression (P< .001) (P values not shown in Table 1). The mean ages by group were 62.9 years for the chemo group, 56.5 years for the chemo-IO group, and 60.9 years for the chemo-bev group. The majority of patients in all 3 treatment cohorts had ECOG scores of 0 or 1; however, in the chemo group 27% of patients had ECOG score of 2 to 3, while 0% of the patients in the chemo-IO group and 14% of the patients in the chemo-bev group had ECOG scores 2 to 3. The vast majority of all patients had the classic EGFR exon 19 deletions or L858R mutations. Four patients in the chemo group (7%) and 1 patient in the chemo-bev group (3%) had less common mutations, as listed in Table 1; all of these are thought to be sensitizing mutations.12,13 Out of the full cohort of 104 patients, 63 (61%) had PD-L1 testing. Among those tested, PD-L1 expression was ≥ 1% in 30% of the chemo group, 75% of the chemo-IO group, and 50% of the chemo-bev group. With respect to metastatic sites of disease prior to the start of chemotherapy, brain metastases had occurred in 67% of the chemo group, 50% of the chemo-IO group, and 54% of the chemo-bev group; liver metastases had occurred in 35% of the chemo group, 42% of the chemo-IO group, and 17% of the chemo-bev group.
Table 1.
Baseline Characteristics by Subgroup (n = 104 Patients)
| Characteristic | Chemo(n = 57) | Chemo + IO(n = 12) | Chemo + Bev(n = 35) | ASD |
|---|---|---|---|---|
| Mean age (SD) | 62.9 (10.6) | 56.5 (8.9) | 60.9 (11.1) | 0.427 |
| Male sex – n (%) | 21 (37) | 6 (50) | 10 (29) | 0.298 |
| Race – n (%) | 0.419 | |||
| Asian | 20 (35) | 6 (50) | 20 (57) | |
| White | 30 (53) | 4 (33) | 11 (31) | |
| Other | 7 (12) | 2 (17) | 4 (11) | |
| Any smoking history – n (%) | 17 (30) | 3 (25) | 7 (20) | 0.152 |
| ECOG PS – n (% among those tested) | 0.582 | |||
| Not documented | 1 | 1 | 0 | |
| 0–1 | 41/56 (73) | 11/11 (100) | 30 (86) | |
| 2–3 | 15/56 (27) | 0 | 5 (14) | |
| Mean Charlson Comorbidity Index (SD) | 8.1 (1.7) | 7.3 (0.9) | 7.8 (1.3) | 0.403 |
| Stage at initial diagnosis – n (%) | 0.010 | |||
| Stage I-III | 5 (9) | 1 (8) | 3 (9) | |
| Stage IV | 53 (91) | 11 (92) | 32 (91) | |
| EGFR mutation – n (%) | 0.363 | |||
| Exon 19 deletion | 33 (58) | 7 (58) | 16 (46) | |
| L858R | 20 (35) | 5 (42) | 18 (51) | |
| Other EGFR mutationa | 4 (7) | 0 | 1 (3) | |
| PD-L1 expression – n (% among those tested) | 0.646 | |||
| Not tested/not available | 34 | 0 | 7 | |
| < 1% | 16/23 (70) | 3 (25) | 14/28 (50) | |
| ≥ 1% | 7/23 (30) | 9 (75) | 14/28 (50) | |
| 1%–49% | 3/23 (13) | 4 (33) | 11/28 (39) | |
| 50%–100% | 4/23 (17) | 5 (42) | 3/28 (11) | |
| Prior therapy in metastatic setting – n (%) | 0.310 | |||
| Osimertinib only | 20 (35) | 7 (58) | 13 (37) | |
| Other EGFR TKI followed by osi | 37 (65) | 5 (42) | 22 (63) | |
| History of brain metastases – n (%) | 38 (67) | 6 (50) | 19 (54) | 0.228 |
| History of liver metastases – n (%) | 20 (35) | 5 (42) | 6 (17) | 0.371 |
| History of bone metastases – n (%) | 40 (70) | 8 (67) | 23 (66) | 0.064 |
| Current steroid use – n (%) | 6 (11) | 0 | 4 (11) | 0.341 |
| Current anticoagulation use – n (%) | 16 (28) | 2 (17) | 1 (3) | 0.500 |
Abbreviations: ASD = absolute standardized difference; PS = performance status; SD = standard deviation; TKI = tyrosine kinase inhibitor.
Other EGFR mutations: In Chemo group: L861Q; L861Q and L833F compound mutation; L833V and H835L compound mutation; G719A and E709K compound mutation. In Chemo + Bev group: G719X and S7681 compound mutation.
Of the 104 patients, only 2 discontinued their chemotherapy-based regimens due to toxicity, and both of these were in the chemotherapy cohort. A total of 73 patients (70%) had reached the overall survival endpoint at the time of data cut-off; this included 38 of 57 patients (67%) who received chemo, 12 of 12 patients (100%) who received chemo-IO, and 23 of 35 patients (66%) who received chemo-bev. Supplemental Table S1 shows lines of subsequent therapies and rates of subsequent immunotherapy and anti-angiogenic agents in later line therapies (beyond immediate post-osimertinib chemotherapy regimen), among these 73 patients.
In univariate analyses, 2 baseline characteristics – ECOG performance status and brain metastases – were significantly associated with outcomes. Patients with ECOG PS 2 to 3 at the start of chemotherapy had both shorter DOT (HR 1.81, 95% CI 1.07–3.05; P= .027) and shorter OS (HR 2.65, 95% CI 1.51–4.65; P < .001) compared with those with ECOG PS 0 to 1. Patient with brain metastases at the start of chemotherapy also had both shorter DOT (HR 2.01, 95% CI 1.29–3.12; P= .0019) and shorter OS (HR 3.12, 95% CI 1.83–5.31; P < .001) than did those without brain metastases Table 2.
Table 2.
Univariate Analyses: Associations of Baseline Characteristics With DOT and OS in Full Cohort (n = 104 Patients)
| Characteristic | HR for DOT (95% CI) | P value | HR for OS (95% CI) | Pvalue |
|---|---|---|---|---|
| Age (per year) | 1.01 (0.98–1.03) | .64 | 1.00 (0.98–1.02) | .94 |
| Male sex (ref = female) | 0.85 (0.55–1.31) | .45 | 0.75 (0.46–1.21) | .24 |
| Race | ||||
| Asian | Ref | Ref | ||
| White | 1.17 (0.75–1.82) | .50 | 0.87 (0.53–1.45) | .60 |
| Other | 0.95 (0.49–1.86) | .89 | 1.05 (0.51–2.14) | .90 |
| Any smoking history (ref = no smoking history) | 0.99 (0.61–1.62) | .97 | 0.87 (0.50–1.51) | .62 |
| ECOG PS 2–3 (ref = ECOG PS 0–1) | 1.81 (1.07–3.05) | .027 | 2.65 (1.51–4.65) | < .001 |
| Charlson Comorbidity Index (per 1 unit) | 1.01 (0.88–1.17) | .87 | 1.00 (0.86–1.17) | .99 |
| Early stage recurrent disease (ref = stage IV) | 0.86 (0.42–1.79) | .69 | 1.34 (0.64–2.81) | .44 |
| EGFR mutation | ||||
| Exon 19 deletion | Ref | Ref | ||
| L858R mutation | 1.32 (0.87–2.01) | .20 | 1.32 (0.82–2.12) | .25 |
| Other EGFR mutation | 0.49 (0.12–2.04) | .33 | 0.34 (0.05–2.47) | .29 |
| PD-L1 expression ≥ 1% (ref = PD-L1 < 1%) | 1.20 (0.71–2.02) | .51 | 1.36 (0.74–2.50) | .32 |
| Prior osi only (ref = other EGFR TKI, then osi) | 1.01 (0.65–1.56) | .98 | 1.27 (0.77–2.09) | .36 |
| History of brain metastases | 2.01 (1.29–3.12) | .0019 | 3.12 (1.83–5.31) | < .001 |
| History of liver metastases | 1.15 (0.74–1.80) | .54 | 0.96 (0.58–1.60) | .88 |
| History of bone metastases | 1.19 (0.76–1.87) | .44 | 1.31 (0.79–2.17) | .30 |
| Current steroid use | 1.38 (0.66–2.88) | .39 | 1.64 (0.70–3.88) | .26 |
| Current anticoagulation use | 0.79 (0.45–1.38) | .41 | 0.67 (0.34–1.32) | .25 |
Chemotherapy + Immunotherapy Versus Chemotherapy
The median DOT was 5.22 months (95% CI 2.47-NE) for chemo-IO versus 5.03 months (95% CI 3.68–7.82 months) in the chemo group (Figure 1A). In the unadjusted regression model comparing DOT of chemo-IO to chemo, the hazard ratio for treatment discontinuation was 1.67 (95% CI 0.86–3.25); P= .13. In the multivariable Cox regression model adjusted for ECOG PS and brain metastases, the hazard ratio was 1.92 (95% CI 0.93–3.94); P= .076 (Table 3A).
Figure 1.
Chemo + IO versus chemo: (A) duration on treatment, (B) overall survival.
Table 3A.
Chemo + IO Versus Chemo: Multivariable Cox Models for DOT and OS
| Duration on Treatment | Overall Survival | |||
|---|---|---|---|---|
| Factor | Hazard Ratio (95% CI) | Pvalue | Hazard Ratio (95% CI) | Pvalue |
| Chemo + IO (ref = chemo) | 1.92 (0.93–3.94) | .076 | 2.66 (1.25–5.65) | .011 |
| ECOG PS 2–3 (ref = ECOG PS 0–1) | 1.80 (0.95–3.40) | .070 | 2.32 (1.14–4.70) | .020 |
| Brain metastases | 1.77 (1.01–3.13) | .048 | 3.63 (1.77–7.48) | < .001 |
The median OS was 10.9 months (95% CI 9.4-NE) for chemo-IO versus 12.0 months (95% CI 9.56–19.4 months) for chemo (Figure 1B). In the unadjusted regression model comparing OS of chemo-IO to chemo, the hazard ratio for death was 1.67 (95% CI 0.86–3.23); P= .13. In the multivariable Cox regression model adjusted for ECOG PS and brain metastases, the hazard ratio was 2.66 (95% CI 1.25–5.65); P= .011 (Table 3A). In sensitivity analyses in which the patients who continued on osimertinib when starting next-line therapy (n = 1 of 12 patients in chemo-IO group, n = 26 of 57 patients in chemo group) were excluded, the hazard ratio for death remained similar and statistically significant (HR 2.64, 95% CI 1.10–6.34; P= .029) (Table S2A).
Chemotherapy + Bevacizumab Versus Chemotherapy
The median DOT was 6.01 months (95% CI 4.47–11.7) for chemo-bev versus 5.03 months (95% CI 3.68–7.82 months) in the chemotherapy group (Figure 2A). In the unadjusted regression model comparing DOT of chemo-bev to chemo, the hazard ratio for treatment discontinuation was 0.90 (95% CI 0.57–1.42); P= .64. In the multivariable Cox regression model adjusted for ECOG PS and brain metastases, the hazard ratio was 1.11 (95% CI 0.69–1.80); P= .67 (Table 3B).
Figure 2.
Chemo + Bev versus chemo: (A) duration on treatment, (B) overall survival.
Table 3B.
Chemo + Bev Versus Chemo: Multivariable Cox Models for DOT and OS
| Duration on Treatment | Overall Survival | |||
|---|---|---|---|---|
| Factor | Hazard Ratio (95% CI) | Pvalue | Hazard Ratio (95% CI) | Pvalue |
| Chemo + bev (ref = chemo) | 1.11 (0.69–1.80) | .67 | 1.50 (0.84–2.69) | .17 |
| ECOG PS 2–3 (ref = ECOG PS 0–1) | 1.84 (1.07–3.19) | .029 | 2.79 (1.50–5.18) | .0011 |
| Brain metastases | 2.08 (1.26–3.41) | .0039 | 3.66 (1.89–7.10) | < .001 |
The median OS was 15.2 months (95% CI 10.55-NE) for chemo-bev versus 12.0 months (95% CI 9.56–19.4 months) for chemotherapy (Figure 2B). In the unadjusted regression model comparing OS of chemo-bev to chemotherapy, the hazard ratio for death was 0.87 (95% CI 0.51–1.46); P= .59. In the multivariable Cox regression model adjusted for ECOG PS and brain metastases, the hazard ratio was 1.50 (95% CI 0.84–2.69); P= .17 (Table 3B). In sensitivity analyses in which the patients who continued on osimertinib when starting chemotherapy (n = 2 of 35 patients in chemo-bev group, n = 26 of 57 patient in chemo group) were excluded, the hazard ratio for death remained similar and not statistically significant (HR 1.36 (95% CI 0.66–2.94); P= .39 (Tables S2B).
Chemotherapy + Immunotherapy Versus Chemotherapy + Bevacizumab
The median time to treatment discontinuation was 5.22 months (95% CI 2.47-NE) for chemo-IO versus 6.01 months (95% CI 4.47–11.7) for chemo-bev (Figure 3A). In the unadjusted regression model comparing DOT of chemo-IO to chemo-bev, the hazard ratio for treatment discontinuation was 2.17 (95% CI 1.05–4.48); P= .037. In the multivariable Cox regression model adjusted for ECOG PS and brain metastases, the hazard ratio was 2.22 (95% CI 1.03–4.81); P= .043 (Table 3C).
Figure 3.
Chemo + IO versus chemo + Bev: (A) duration on treatment, (B) overall survival.
Table 3C.
Chemo + IO Versus Chemo + Bev: Multivariable Cox models for DOT and OS
| Duration on Treatment | Overall Survival | |||
|---|---|---|---|---|
| Factor | Hazard Ratio (95% CI) | Pvalue | Hazard Ratio (95% CI) | Pvalue |
| Chemo + IO (ref = chemo + bev) | 2.22 (1.03–4.81) | .043 | 2.37 (1.09–5.14) | .030 |
| ECOG PS 2–3 (ref = ECOG PS 0–1) | 2.42 (0.80–7.35) | .12 | 7.37 (2.18–24.85) | .0013 |
| Brain metastases | 2.10 (1.07–4.11) | .031 | 4.08 (1.68–9.89) | .0018 |
The median OS was 10.9 months (95% CI 9.4-NE) for chemo-IO versus 15.2 months (95% CI 10.55-NE) for chemo-bev (Figure S1B). In the unadjusted regression model comparing OS of chemo-IO to chemo-bev, the hazard ratio for death was 1.94 (95% CI 0.95–3.99); P= .071. In the multivariable Cox regression model adjusted for ECOG PS and brain metastases, the hazard ratio was 2.37 (95% CI 1.09–5.14); P= .030 (Table 3C). In sensitivity analyses in which the patients who continued on osimertinib when starting chemotherapy (n = 1 of 12 patients in chemo-IO group, n = 2 of 35 patients in chemo-bev group) were excluded, the hazard ratio for death remained similar and statistically significant (HR 2.26, 95% CI 1.04–4.91; P= .041) (Table S2C).
Discussion
To our knowledge, this is the first study to examine outcomes with the addition of either immunotherapy or bevacizumab to the platinum chemotherapy backbone for EGFR-mutant lung cancer in the post-osimertinib setting. We found, in multivariable models adjusted for ECOG performance status and brain metastases, that in our cohort the patients who received immunotherapy in combination with platinum doublet chemotherapy had significantly worse OS than those who received platinum doublet chemotherapy, with a hazard ratio for death of 2.66 (P= .011), and then those who received platinum doublet chemotherapy plus bevacizumab, with a hazard ratio for death of 2.37 (P= .030). The statistically significant between-group differences in outcomes observed remained significant in sensitivity analyses excluding patients who continued on osimertinib when initiating chemotherapy-based regimens. A statistically significant difference in OS could not be detected in the chemotherapy + bevacizumab cohort versus the chemotherapy cohort.
At our institutions, the most commonly used and our currently preferred regimens to give after progression on osimertinib are platinum doublet chemotherapy with or without the addition of bevacizumab. In our experience, both the platinum doublet and platinum doublet plus bevacizumab regimens allow for the option of continuing osimertinib along with the chemotherapy-based regimen, and this can be considered in select cases as well. Our group has previously published retrospective data on combining osimertinib plus chemotherapy, and this strategy appeared to be safe and to potentially help with ongoing CNS disease control in patients with brain metastases who have maintained CNS disease control on osimertinib, despite progression extracranially.4 This strategy is now being prospectively evaluated in the phase III trial COMPEL (ClinicalTrials.gov identifier: NCT04765059). In the present study, osimertinib was indeed continued on in many patients in our cohort, especially in the chemo group. These patients were included in the main analyses because the decision to continue osimertinib is not independent from the decision of which chemotherapy-based regimen to give in the post-osimertinib setting, as giving osimertinib in combination with immunotherapy has been shown to cause high rates of toxicity, in particular pneumonitis.14 However, because of the potential for continuation of osimertinib to be a confounding factor in our main analyses, we did perform sensitivity analyses excluding those patients, and our findings remained unchanged.
Considering our findings as well as previously published data, we feel that there are not convincing data at this time to support giving the triplet combination of platinum doublet chemotherapy plus immunotherapy to patients with sensitizing EGFR mutations. As discussed earlier, IMpower130 found no benefit from adding immunotherapy to chemotherapy in the EGFR mutant subgroup,8 and KEYNOTE-189 excluded patients with EGFR mutations.6 In the large phase III clinical trials that demonstrated superiority of single-agent immunotherapy over chemotherapy in advanced NSCLC patients in the second line or beyond (which were conducted before immunotherapy was approved in the first-line setting), subset analyses of patients with EGFR mutations suggested that this group does not derive the same benefit – and possibly has worse outcomes – with immunotherapy.15–18 Real world data have also shown that, unlike those without driver mutations, patients with EGFR mutations have very low response to PD-1/PD-L1 inhibitor monotherapy.19,20 A single-arm phase II clinical trial in which patients with EGFR-mutant, PD-L1 positive (≥ 1%; the majority with PD-L1 > 50%) lung cancer, who had not received prior TKI therapy, were treated with single-agent pembrolizumab was terminated early due to lack of efficacy, as 0 of the 10 EGFR-mutant patients who had enrolled thus far on the study had objective responses, and there were also concerns about toxicity given 2 deaths within 6 months of enrollment, 1 of which was from pneumonitis.21 Potential explanations for low responses to immunotherapy include the relatively immunosuppressed tumor microenvironment and low tumor-infiltrating CD8+ lymphocytes in EGFR-mutant lung cancer.20,22,23 Based upon the existing data on immunotherapy in EGFR-mutant NSCLC, we hypothesized that our retrospective analysis would show no additional benefit from the addition of immunotherapy to platinum doublet chemotherapy over platinum doublet chemotherapy alone in the post-osimertinib setting, but were surprised to observe that the combination appeared to potentially worsen outcomes, with significantly worse survival outcomes in patients who received the combination in our cohort. This is not attributable to toxicity with the combination; only 2 of the 104 patients in our cohort discontinued their chemotherapy-based regimens due to toxicity, and both of these were in the chemotherapy cohort (we did not capture therapy dose reductions or delays due to toxicity, however). There are phase III studies currently ongoing, KEYNOTE-789 (ClinicalTrials.gov identifier: NCT03515837) and CheckMate 722 (ClinicalTrials.gov identifier: NCT02864251), to definitively answer the question of whether the addition of immunotherapy to platinum doublet chemotherapy is beneficial in EGFR-mutant lung cancer in the post-TKI setting. Pending prospective data from these trials, our data do not support the use of chemoimmunotherapy in this setting. However, our small retrospective study, with only 12 patients who received chemo-IO, should be considered hypothesis-generating and not conclusive evidence that chemo-IO is detrimental in this setting. It is possible that there is a subgroup of patients amongst those with EGFR-mutant lung cancer that could benefit from chemo-IO. For example, level of PD-L1 expression could be a factor, and type of EGFR mutation could play a role; there are some retrospective data suggesting that EGFR L858R mutant lung cancers may have better response to immunotherapy than EGFR exon 19 deletion lung tumors.24 The number of patients who received chemo-IO in our cohort was too small to evaluate effect of these variables within the chemo-IO group.
Another option for therapy in the post-osimertinib setting, which was not studied in the present analysis, is the IMpower150 regimen, consisting of carboplatin/paclitaxel plus bevacizumab plus immunotherapy (atezolizumab).9 While at our institutions we do not standardly use the IMpower150 regimen in the immediate post-osimertinib setting (we identified only 1 such patient between our 2 institutions; this patient was not included in the present analysis), we do feel that this is a reasonable option to consider and, based on available data, may be the best way in which to give immunotherapy to a patient with EGFR-mutant disease. We do have patients at both of our institutions who were treated with this regimen in subsequent lines of therapy. The only positive data from randomized phase III trials that support the addition of immunotherapy to any chemotherapy-based regimen in EGFR-mutant lung cancer is from the IMpower150 study, in which the quadruplet appeared to improve OS versus chemotherapy plus bevacizumab in the EGFR positive subset. In fact, the magnitude of effect appeared larger for the EGFR mutant subset (HR 0.31; 95% CI 0.11–0.83) than for the wild-type intention-to-treat population (HR 0.76; 95% CI 0.63–0.93).9,10 The lack of benefit in the EGFR-mutant subgroup in the IMpower130 trial,8 which was a similar study but did not include bevacizumab in either arm, raises the question of whether the presence of bevacizumab was key to the findings seen with IMpower150. Indeed, there are growing data suggesting possible synergy between bevacizumab and immunotherapy, and there is a plausible biological explanation for this. Bevacizumab is known to have immunomodulatory effects, and EGFR-mutant tumors tend to have more immune-suppressive tumor microenvironments; thus, the reprogramming from a more immune-suppressive to a more immune-permissive microenvironment by bevacizumab may be especially important for EGFR-mutant tumors, and prime them to derive benefit from immunotherapy.25,26 However, the EGFR-mutant cohorts in both IMpower150 and IMpower130 were small subsets, and because the trials began enrolling prior to the approval of osimertinib in the first-line setting, very few of the patients had received osimertinib. One concern with the IMpower150 regimen is the toxicity with a 4-drug regimen, particularly one that includes a taxane. The POINTBREAK trial compared carboplatin plus pemetrexed plus bevacizumab against carboplatin plus paclitaxel plus bevacizumab and found comparable survival but lower toxicity with the pemetrexed-based regimen,27 thus pemetrexed is generally preferred over paclitaxel as a platinum partner. However, pemetrexed is not approved in combination with bevacizumab and immunotherapy. There is an ongoing phase II randomized trial, TH-138 (ClinicalTrials.gov identified: NCT03786692), evaluating carboplatin/pemetrexed/bevacizumab with or without atezolizumab, enrolling patients with EGFR-mutant lung cancer post-TKI progression. This is an important study as it will evaluate a quadruplet regimen with both bevacizumab and immunotherapy in EGFR-mutant disease specifically, and it includes pemetrexed as the platinum partner rather than a taxane.
Most of the limitations of our study are inherent to retrospective analyses. There were limited numbers of patients, particularly in the chemo-IO group, and all patients were treated at 2 academic institutions. Thus, the patient population may not be representative of patients with metastatic EGFR-mutant lung cancer in general. We did not include patients who received another targeted therapy or went on a clinical trial immediately following osimertinib, thus this also results in a patient population that is different from the overall group of patients who have previously progressed on osimertinib. There are likely many factors beyond those listed in Table 1 that influence an oncologist’s decision about which regimen to select in the post-osimertinib setting that potentially could be confounding variables that were not adjusted for in our regression models. One example of such a factor is burden of disease: higher disease burden may prompt physicians to consider adding immunotherapy or bevacizumab to chemotherapy. In our cohort we captured history of brain metastases, liver metastases, and bone metastases as binary variables, but not extent of disease in these organs, in the thorax, or elsewhere. Location and other radiographic features of the tumor can also be considerations in therapy selection; eg, in patients with central or cavitary tumors, the bleeding risk with bevacizumab may outweigh potential benefit. Another limitation of this study is that we did not compare outcomes with chemotherapy + bevacizumab + immunotherapy, since the quadruplet (IMpower 150 regimen) is not standardly used at our institution immediately following osimertinib. It would be helpful to see similar analyses repeated in larger multi-center cohorts that include more patients who received chemo-IO, as well as patients who received chemotherapy plus immunotherapy plus bevacizumab.
In conclusion, in our small single-center cohort, the addition of immunotherapy to platinum doublet chemotherapy in the post-osimertinib setting worsened survival outcomes compared with platinum doublet chemotherapy alone or platinum doublet chemotherapy plus bevacizumab. There was not a statistically significant difference in survival outcomes between platinum doublet chemotherapy plus bevacizumab versus platinum doublet chemotherapy alone. Thus, we favor using platinum doublet chemotherapy without immunotherapy; the addition of bevacizumab to chemotherapy could be considered in selected patients, and continuation of osimertinib with initiation of chemotherapy is an option to consider as well. Ongoing studies in EGFR-mutant lung cancer in the post-EGFR TKI setting will help more definitely answer the question of best therapy after osimertinib; these include Keynote-789 and Checkmate 722, both investigating platinum doublet chemotherapy with or without immunotherapy (pembrolizumab and nivolumab, respectively), and the TH-138 study of carboplatin/pemetrexed plus bevacizumab with or without immunotherapy (atezolizumab). It is important to note that there are also numerous currently ongoing phase III trials investigating combinations of third-generation EGFR TKIs with other agents, including but not limited to platinum doublet chemotherapy and anti-angiogenic agents.28 Thus, while osimertinib is the widely accepted first-line therapy of choice in metastatic EGFR-mutant lung cancer currently, this could change in the future, opening further questions about choice and sequencing of subsequent therapies. As the treatment landscape continues to evolve, both prospective studies and ongoing evaluation of real-world data will be crucial to continue to improve outcomes in patients with EGFR mutant lung cancer.
Clinical Practice Points
In patients with metastatic EGFR-mutated non-small-cell lung cancer, optimal next-line therapy after progression on osimertinib is not well-established. Patients typically go on to receive platinum doublet chemotherapy-based regimens, but whether the addition of other agents such as immunotherapy or bevacizumab to the chemotherapy backbone provides additional benefit is unknown.
Given the widespread use of chemoimmunotherapy combinations as first-line therapy for advanced NSCLC without targetable driver mutation, immunotherapy is sometimes given with the chemotherapy backbone. Similarly, older studies suggested benefit to chemotherapy and bevacizumab combinations in metastatic NSCLC. However, neither chemotherapy plus immunotherapy nor chemotherapy plus bevacizumab are well-studied in EGFR-mutant lung cancer.
This study retrospectively assessed survival outcomes in 104 patients who had had progression on osimertinib and received next-line therapy with platinum doublet chemotherapy, platinum doublet chemotherapy plus immunotherapy, or platinum doublet chemotherapy plus bevacizumab. Patients who had continued osimertinib with the initiation of their chemotherapy regimens were included. · We found in adjusted models that patients in our cohort who received chemotherapy plus immunotherapy had worse overall survival than did those who received chemotherapy or chemotherapy plus bevacizumab. There was not a statistically significant difference in survival in patients who received chemotherapy plus bevacizumab versus those who received chemotherapy alone.
Our findings suggest that platinum doublet chemotherapy without immunotherapy (and with consideration of continuation of osimertinib, in selected cases) is a reasonable choice in the post-osimertinib setting, while we await results of clinical trials examining optimal next-line chemotherapy-based regimens in EGFR-mutant lung cancer.
Supplementary Material
Supplementary materials
Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cllc.2021.11.001.
Acknowledgments
This research used data or services provided by STARR, “STAnford medicine Research data Repository,” a clinical data warehouse containing live Epic data from Stanford Health Care (SHC), the Stanford Children’s Hospital (SCH), the University Healthcare Alliance (UHA) and Packard Children’s Health Alliance (PCHA) clinics and other auxiliary data from Hospital applications such as radiology PACS. STARR platform is developed and operated by Stanford Medicine Research IT team and is made possible by Stanford School of Medicine Research Office.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Footnotes
Disclosure
Dr Neal has received research funding from Genentech/Roche, Merck, Novartis, Boehringer Ingelheim, Exelixis, ARIAD/Takeda Pharmaceuticals, Nektar, Adaptimmune, and GlaxoSmithKline; has served in a consulting role for ARIAD/Takeda Pharmaceuticals, Amgen, Calithera Biosciences, Eli Lilly, AstraZeneca, Genentech/Roche, Exelixis, Iovance Biotherapeutics, Loxo Oncology, and Jounce Therapeutics; and has received honoraria from Biomedical Learning Institute, CME Matters, Medscape, MJH CME, MLI Peerview, Prime Oncology, Research to Practice, and Rockpointe. Dr Das has received research funding from Novartis, Varian, Genzyme, Verily, AbbVie, and United Therapeutics and had a consulting role with Jazz Pharma, Bristol-Myers Squibb, and AstraZeneca. Dr Padda has received research funding from EpicentRx, Forty Seven Inc, Bayer Pharmaceuticals, and Boehringer Ingelheim; has received honoraria from CME Solutions and PER; and has participated in the advisory boards for AstraZeneca, Blueprint, Pfizer, AbbVie, G1 Therapeutics, Clovis Oncology, and Janssen Pharmaceuticals. Dr Ramchandran has participated in the advisory boards for Dhristi Inc, GTX, and Varian. Dr Sequist has served in a consulting role for AstraZeneca, Genetech, Janssen, Pfizer and Takeda and has received research funding from: AstraZeneca, BI, Genentech, and Novartis. Dr Piotrowska has received consulting honoraria from AstraZeneca, Eli Lilly, InCyte, Medtronic, C4 Therapeutics, Blueprint Medicines, Janssen, Cullinan, Jazz, and Takeda and receives institutional research support from Novartis, Takeda, Spectrum, AstraZeneca, Tesaro, Cullinan Oncology, AbbVie, Janssen and Daiichi-Sankyo. Dr Wakelee reports receiving honoraria from Novartis, AstraZeneca, Janssen, Daiichi-Sankyo, and Xcovery and research funding from Genentech, Roche, Pfizer, Eli Lilly, Celgene, AstraZeneca, MedImmune, Exelixis, Novartis, Clovis Oncology, Xcovery, Bristol-Myers Squibb, Gilead Sciences, Pharmacyclics, and ACEA Biosciences. The remaining authors declare no conflict of interest.
References
- 1.Ramalingam SS, Vansteenkiste J, Planchard D, et al. Overall survival with osimertinib in untreated, EGFR-mutated advanced NSCLC. N Engl J Med. 2020;382:41–50. [DOI] [PubMed] [Google Scholar]
- 2.Soria JC, Ohe Y, Vansteenkiste J, et al. Osimertinib in untreated EGFR-mutated advanced non-small-cell lung cancer. N Engl J Med. 2018;378:113–125. [DOI] [PubMed] [Google Scholar]
- 3.Piper-Vallillo AJ, Sequist LV, Piotrowska Z. Emerging treatment paradigms for EGFR-mutant lung cancers progressing on osimertinib: a review. J Clin Oncol. 2020;38:2926–2936. [DOI] [PubMed] [Google Scholar]
- 4.White MN, Piotrowska Z, Stirling K, et al. Combining osimertinib with chemotherapy in EGFR-mutant NSCLC at progression. Clin Lung Cancer. 2021;3:201–209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sandler A, Gray R, Perry MC, et al. Paclitaxel-carboplatin alone or with bevacizumab for non-small-cell lung cancer. N Engl J Med. 2006;355:2542–2550. [DOI] [PubMed] [Google Scholar]
- 6.Gandhi L, Garassino MC. Pembrolizumab plus chemotherapy in lung cancer. N Engl J Med. 2018;379:e18. [DOI] [PubMed] [Google Scholar]
- 7.Gadgeel S, Rodríguez-Abreu D, Speranza G, et al. Updated analysis from KEYNOTE-189: pembrolizumab or placebo plus pemetrexed and platinum for previously untreated metastatic nonsquamous non-small-cell lung cancer. J Clin Oncol. 2020;38:1505–1517. [DOI] [PubMed] [Google Scholar]
- 8.West H, McCleod M, Hussein M, et al. Atezolizumab in combination with carboplatin plus nab-paclitaxel chemotherapy compared with chemotherapy alone as first-line treatment for metastatic non-squamous non-small-cell lung cancer (IMpower130): a multicentre, randomised, open-label, phase 3 trial. Lancet Oncol. 2019;20:924–937. [DOI] [PubMed] [Google Scholar]
- 9.Socinski MA, Jotte RM, Cappuzzo F, et al. Atezolizumab for first-line treatment of metastatic nonsquamous NSCLC. N Engl J Med. 2018;378:2288–2301. [DOI] [PubMed] [Google Scholar]
- 10.Reck M, Mok TSK, Nishio M, et al. Atezolizumab plus bevacizumab and chemotherapy in non-small-cell lung cancer (IMpower150): key subgroup analyses of patients with EGFR mutations or baseline liver metastases in a randomised, open-label phase 3 trial. Lancet Respir Med. 2019;7:387–401. [DOI] [PubMed] [Google Scholar]
- 11.Austin PC. Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Commun Stat Simul Comput. 2009;38:1228–1234. [Google Scholar]
- 12.O’Kane GM, Bradbury PA, Feld R, et al. Uncommon EGFR mutations in advanced non-small cell lung cancer. Lung Cancer. 2017;109:137–144. [DOI] [PubMed] [Google Scholar]
- 13.Hata A, Yoshioka H, Fujita S, et al. Complex mutations in the epidermal growth factor receptor gene in non-small cell lung cancer. J Thorac Oncol. 2010;5:1524–1528. [DOI] [PubMed] [Google Scholar]
- 14.Oxnard GR, Yang JC, Yu H, et al. TATTON: a multi-arm, phase Ib trial of osimertinib combined with selumetinib, savolitinib, or durvalumab in EGFR-mutant lung cancer. Ann Oncol. 2020;31:507–516. [DOI] [PubMed] [Google Scholar]
- 15.Rittmeyer A, Barlesi F, Waterkamp D, et al. Atezolizumab versus docetaxel in patients with previously treated non-small-cell lung cancer (OAK): a phase 3, open-label, multicentre randomised controlled trial. Lancet. 2017;389:255–265. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Borghaei H, Paz-Ares L, Horn L, et al. Nivolumab versus docetaxel in advanced nonsquamous non-small-cell lung cancer. N Engl J Med. 2015;373:1627–1639. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Herbst RS, Baas P, Kim DW, et al. Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-010): a randomised controlled trial. Lancet. 2016;387:1540–1550. [DOI] [PubMed] [Google Scholar]
- 18.Lee CK, Man J, Lord S, et al. Checkpoint inhibitors in metastatic EGFR-mutated non-small cell lung cancer-a meta-analysis. J Thorac Oncol. 2017;12:403–407. [DOI] [PubMed] [Google Scholar]
- 19.Mazieres J, Drilon A, Lusque A, et al. Immune checkpoint inhibitors for patients with advanced lung cancer and oncogenic driver alterations: results from the IMMUNOTARGET registry. Ann Oncol. 2019;30:1321–1328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Gainor JF, Shaw AT, Sequist LV, et al. EGFR mutations and ALK rearrangements are associated with low response rates to PD-1 pathway blockade in non-small cell lung cancer: a retrospective analysis. Clin Cancer Res. 2016;22:4585–4593. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Lisberg A, Cummings A, Goldman JW, et al. A phase II study of pembrolizumab in EGFR-mutant, PD-L1+, tyrosine kinase inhibitor naïve patients with advanced NSCLC. J Thorac Oncol. 2018;13:1138–1145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Soo RA, Lim SM, Syn NL, et al. Immune checkpoint inhibitors in epidermal growth factor receptor mutant non-small cell lung cancer: current controversies and future directions. Lung Cancer. 2018;115:12–20. [DOI] [PubMed] [Google Scholar]
- 23.Dong ZY, Zhang JT, Liu SY, et al. EGFR mutation correlates with uninflamed phenotype and weak immunogenicity, causing impaired response to PD-1 blockade in non-small cell lung cancer. Oncoimmunology. 2017;6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hastings K, Yu HA, Wei W, et al. EGFR mutation subtypes and response to immune checkpoint blockade treatment in non-small-cell lung cancer. Ann Oncol. 2019;30:1311–1320. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Manegold C, Dingemans AC, Gray JE, et al. The potential of combined immunotherapy and antiangiogenesis for the synergistic treatment of advanced NSCLC. J Thorac Oncol. 2017;12:194–207. [DOI] [PubMed] [Google Scholar]
- 26.Lin A, Wei T, Meng H, Luo P, Zhang J. Role of the dynamic tumor microenvironment in controversies regarding immune checkpoint inhibitors for the treatment of non-small cell lung cancer (NSCLC) with EGFR mutations. Mol Cancer. 2019;18:139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Patel JD, Socinski MA, Garon EB, et al. PointBreak: a randomized phase III study of pemetrexed plus carboplatin and bevacizumab followed by maintenance pemetrexed and bevacizumab versus paclitaxel plus carboplatin and bevacizumab followed by maintenance bevacizumab in patients with stage IIIB or IV nonsquamous non-small-cell lung cancer. J Clin Oncol. 2013;31:4349–4357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Passaro A, Janne PA, Mok T, Peters S. Overcoming therapy resistance in EGFR-mutant lung cancer. Nat Cancer. 2021;2:377–391. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.