The different overall survival between single-agent EGFR-TKI treatment and with bevacizumab in non-small cell lung cancer patients with brain metastasis

Comparison of epidermal growth factor receptor tyrosine kinase inhibitor (EGFR-TKI) monotherapy or with bevacizumab in real-world non-small cell lung cancer (NSCLC) patients was lacking. 310 patients of advanced NSCLC with common EGFR mutation receiving first-generation EGFR-TKI monotherapy or with bevacizumab were included and propensity-score matched. Progression-free survival (PFS), overall survival (OS) and secondary T790M mutation were analysed. Patients receiving EGFR-TKI and bevacizumab were significantly younger, had better performance status and with high incidence of brain metastasis (55.8%). In the propensity-score matched cohort, PFS (13.5 vs. 13.7 months; log-rank p = 0.700) was similar between the two groups. The OS (61.3 vs. 34.2 months; log-rank p = 0.010) and risk reduction of death (HR 0.42 [95% CI 0.20–0.85]; p = 0.017) were significantly improved in EGFR-TKI plus bevacizumab group. Analysis of treatment by brain metastasis status demonstrated EGFR-TKI plus bevacizumab in patients with brain metastasis was associated with significant OS benefit compared to other groups (log-rank p = 0.030) and these patients had lower early-CNS and early-systemic progressions. The secondary T790M did not significantly differ between EGFR-TKI plus bevacizumab and EGFR-TKI monotherapy groups (66.7% vs. 75.0%, p = 0.460). Forty-one (31.1%) and 31 (23.5%) patients received subsequent osimertinib and chemotherapy, respectively. The post-progression OS of osimertinib and chemotherapy were 22.1 and 44.9 months in EGFR-TKI plus bevacizumab group and were 10.0 and 14.1 months in EGFR-TKI monotherpay group, respectively. First-generation EGFR-TKI with bevacizumab improved treatment efficacy in real-world patients of NSCLC with EGFR mutation. Patients with brain metastasis received additional OS benefit from this treatment.


Scientific Reports
| (2022) 12:4398 | https://doi.org/10.1038/s41598-022-08449-w www.nature.com/scientificreports/ demonstrated PFS benefit of erlotinib plus bevacizumab compared to erlotinib monotherapy in the front-line treatment of advanced NSCLC with sensitizing EGFR mutation 9,10 . However, the OS benefit from bevacizumab add-on was not observed in these studies which may be partly related to the cross-over design of the trials. The superior efficacy of erlotinib and anti-angiogenesis combination over single-agent erlotinib did not seem to be consistent across different clinical trials. Stinchcombe et al. conducted a phase II randomized study comparing erlotinib monotherapy and with bevacizumab in a Caucasian cohort of EGFR-mutant NSCLC patients 11 . In this study, PFS and OS were similar in the two groups. Ichihara et al. reported another phase II single-arm study on gefitinib and bevacizumab combination. The study did not meet the primary endpoint of 1-year PFS though a seemingly longer PFS than historical data of gefitinib monotherapy was observed 12 .
Previously, administration of bevacizumab has demonstrated clinical benefit for both primary brain tumor and brain metastasis from advanced NSCLC, likely as a result of suppressing tumor angiogenesis and reducing intracranial vasogenic edema 13,14 . Lately, significantly improved OS of bevacizumab treatment in advanced NSCLC patients with brain metastasis compared to those without was also observed in a real-world US cancer registry and claim database 15 . Thus, the survival benefit of bevacizumab treatment in EGFR-mutant NSCLC patients with brain metastasis warrants further investigation.
Currently, the third-generation EGFR-TKI osimertinib has arisen as a preferred first-line treatment option for advanced EGFR-mutant NSCLC 16,17 . Recently, combination of osimertinib and bevacizumab has been investigated in both second-line and first-line setting. Akamatsu et al. demonstrated that efficacy of osimertinib plus bevacizumab is statistically equivalent to, yet numerically lower than, osimertinib monotherapy in patients of acquired EGFR T790M-positive NSCLC 18 . In contrary, a phase I/II single arm study reported by Yu et al. showcased a positive result in which osimertinib plus bevacizumab in treatment-naïve EGFR-mutant NSCLC patients met the pre-specified effectiveness end point 19 . On the other side, treatment of osimertinib plus ramucirumab has also been investigated in two phase I studies whereas the efficacy data was premature 20,21 . Overall, whether combination of third-generation EGFR-TKI and anti-angiogenesis agents will be a new standard of treatment remains unsettled.
In this study, we analysed a real-world, brain metastasis-enriched cohort of NSCLC patients with EGFRsensitizing mutation who received first-line erlotinib/gefitinib monotherapy or with bevacizumab. The treatment efficacy and the development of secondary T790M were compared between the two groups.

Methods
Patients and treatment. Patients who received a first-generation EGFR-TKI (gefitinib or erlotinib) monotherapy or with bevacizumab as the first-line treatment of advanced NSCLC with common EGFR mutation (exon 19 deletion or exon 21 L858R) were retrospectively included from January 2014 to December 2019. The dose of EGFR-TKI administered were gefitinib 250 mg/day and erlotinib 150 mg/day, respectively; and bevacizumab was administered at a dose of 7.5 mg/kg every 3 weeks. The lower dose strength was chosen because bevacizumab is not reimbursed by National Health Insurance of Taiwan for NSCLC treatment and this dosage has been demonstrated to achieve an equivalent efficacy and numerically lower adverse events compared to the 15 mg/kg dose strength 22,23 Patients were excluded if the first dose of bevacizumab was given 3 weeks behind the first dose of EGFR-TKI and those who received an EGFR-TKI monotherapy less than 14 days were also excluded. The progression-free survival (PFS) was defined as the interval between the date of starting EGFR-TKI treatment and the date of radiologically or clinically determined progression or death. The treatment response, including complete response (CR), partial response (PR), stable disease, and progressive disease, was evaluated according to the Response Evaluation Criteria in Solid Tumors (version 1.1). The study used data from the Chang Gung Research Database and the study protocol and the waiver of informed consent form were approved by the Ethics Committee of Chang Gung Memorial Hospital.
Statistical analysis. The Mann-Whitney test was used to determine the statistical significance of continuous variables between the two groups and Fisher exact test was used for evaluating the categorical variables. The Kaplan-Meier survival curves were generated using the R package survival, and the hazard ratio (HR) was analysed using the Cox regression model. The propensity-score-matched analysis was used to balance the clinical characteristics between the treatment groups as previously described, in which the distance measure was defined by generalized linear model, the matching method was nearest neighbor matching and the caliper was 0.1 in standard deviation unit 6 . Briefly, the EGFR-TKI plus bevacizumab and EGFR-TKI monotherapy groups served as the dependent variables and the covariates used included age, ECOG PS, EGFR mutation subtypes, brain metastasis and type of EGFR-TKI administered. Paired patients treated with EGFR-TKI plus bevacizumab or EGFR-TKI alone with equivalent propensity scores were selected in a 1:2 manner using the R package MatchIt. The disease progression patterns (CNS progression alone without death versus systemic progression without death) were treated as competing risk events of which the cumulative incidence functions were calculated. The modified Cox regression model for the subdistribution hazard of the cumulative incidence function was applied to calculate the disease progression hazard from a given pattern in the presence of competing events by using the R package cmprsk. All reported p-values are two sided; p < 0.05 was considered statistically significant. Data were also analysed using SPSS (version 10.1; SPSS, Chicago, IL, USA).    Fig. 3A) were noted in EGFR-TKI plus bevacizumab and single-agent EGFR-TKI groups. A significant OS difference was observed among the four treatment groups (Fig. 3A; log-rank test p = 0.030). The pattern of disease progression in patients with baseline brain metastasis was analysed in terms of the cumulative incidence of systemic or CNS progression. The rates of CNS progression (cause-specific HR, 1.77; 95% CI 0.41-7.56; p = 0.858) and of systemic progression over time (cause-specific HR, 0.88; 95% CI 0.49-1.60; p = 0.692, Fig. 3B) did not differ significantly between EGFR-TKI plus bevacizumab and single-agent EGFR-TKI groups whereas treatment of EGFR-TKI plus bevacizumab was associated with lower emergence of early-CNS and early-systemic progressions which were both delayed to occur beyond 8 to 9 months of treatment (Fig. 3B). Table 3. Propensity-score matched cohort. WBRT whole brain radiotherapy, SRS stereotactic radiosurgery.     Fig. 4A), respectively. The post-progression chemotherapy OS were 44.9 and 14.1 months and post-progression chemotherapy 1-year OS rates were 80.0% (95% CI 51.6-100.0%) and 61.6% (95% CI 44.7-84.9%; Fig. 4B) in EGFR-TKI plus bevacizumab and EGFR-TKI monotherpay groups, respectively.

Discussion
The present study demonstrated a superior OS of first-generation EGFR-TKI plus bevacizumab compared to first-generation EGFR-TKI monotherapy in a real-world cohort of EGFR-mutant NSCLC patients with sensitizing mutation. Patients with brain metastasis received additional OS benefit from EGFR-TKI plus bevacizumab. The development of secondary T790M was similar between the two groups. The OS of post-progression treatment, in terms of osimertinib or chemotherapy, in EGFR-TKI plus bevacizumab group were significantly improved compared to single-agent EGFR-TKI group.  www.nature.com/scientificreports/ Although the present study failed to observe PFS benefit of EGFR-TKI plus bevacizumab in EGFR-mutant NSCLC, the OS outcome was significantly improved in patients receiving the regimen. This result was mainly associated with improved efficacies of post-progression treatments after the exposure of EGFR-TKI plus bevacizumab in patients with brain metastasis. Previously, a number of studies have demonstrated satisfactory efficacies of bevacizumab in selected patients with heavily pre-treated symptomatic brain metastasis from advanced solid tumors [24][25][26] . Studies which primarily focused on NSCLC-related brain metastasis also revealed bevacizumab as a viable option that improved intracranial response, stabilized neurological symptoms and reduced systemic corticosteroid use [27][28][29] . Recently, a retrospective study analysed a cohort of EGFR-mutant NSCLC patients with multiple brain metastasis who received either EGFR-TKI plus bevacizumab or single-agent EGFR-TKI where significant PFS and OS benefits were both observed in the EGFR-TKI plus bevacizumab arm 30 . In line with these findings, the present study also demonstrated EGFR-TKI plus bevacizumab provided significant OS benefit in brain metastasis patients who have consisted of approximately 60% of subjects of the propensity-score matched cohort. Overall, patients with brain metastasis represented the major source of OS benefit generated by EGFR-TKI plus bevacizumab in present study. In contrary, these patients were excluded in the JO25567 study and consisted of 32% of the overall population in the NEJ 026 study which may be partly associated with the negative OS results in these studies.
Interestingly, the benefit of bevacizumab treatment in patients with brain metastasis did not seem to be CNS-restricted. Tao et al. specifically investigated a cohort of EGFR-mutant NSCLC patients with multiple brain metastasis and thereby demonstrated that EGFR-TKI plus bevacizumab, compared to EGFR-TKI monotherapy, showed improvement of both intracranial and systemic tumor responses 30 . In present analysis, although the cumulative incidence of CNS and systemic progression over time did not statistically differ between the EGFR-TKI plus bevacizumab and EGFR-TKI monotherapy groups, the early emergence of CNS and systemic progressions were both suppressed in EGFR-TKI plus bevacizumab group and thereby the both types of progression were delayed to occur beyond 8 to 9 months of the treatment. This finding, in association with previous study, likely suggested an improved extracranial tumor control in bevacizumab-treated patients who had intracranial metastasis. Altogether, the bevacizumab treatment-associated CNS and CNS-sparing benefits in patients of brain metastasis may jointly foster a condition that favors the efficacy of subsequent treatments. This was demonstrated in the present analysis that the OS of second-line osimertinib and chemotherapy were both improved in EGFR-TKI plus bevacizumab group compared to single-agent EGFR-TKI.
In addition, other factors related to post-progression treatment may also have impact on OS outcome. Because a cross-over design was allowed in the NEJ 026 study, 28.6% patients of the single-agent erlotnib arm have received subsequent bevacizumab treatment upon disease progression whereas, in present study, only 3.4% of patients who received monotherapy EGFR-TKI have undergone post-progression bevacizumab treatment. Of note, in this analysis, 18.6% of patients who received EGFR-TKI plus bevacizumab continued bevacizumab treatment beyond progression whereas the rate was only 4.5% of the erlotinib plus bevacizumab arm in the NEJ 026 study. The differential rate of bevacizumab treatment beyond progression may also have some impact on OS as previous AvaALL trial has demonstrated a minor benefit of bevacizumab treatment beyond progression compared to standard-of-care alone 31 .
As a new recommended first-line treatment for advanced EGFR-mutant NSCLC, third-generation EGFR-TKI osimertinib provided an improved OS compared to gefitinib/erlotinib. In the post-hoc analysis, cohort of EGFR exon 19 deletion patients was the major source that contributed to the OS benefit whereas cohort of EGFR L858R patients generally received minimal treatment benefit from osimertinib. In contrary, the post-hoc analysis of NEJ 026 study revealed that bevacizumab likely offered additional benefit to the cohort of EGFR L858R patients and this finding has also been reported in real-world patients 32 . The present study also observed a trend of OS benefit in EGFR L858R patients who received combination of EGFR-TKI and bevacizumab. Hence, these findings may warrant further investigation of bevacizumab treatment in patients of EGFR L858R genotype.
Previously, pre-clinical study demonstrated activity of bevacizumab to EGFR T790M mutant and a randomized phase II trial also showcased an improved PFS in patients with de novo T790M mutation treated by erlotinib plus bevacizumab 33,34 . However, the potential activity of anti-angiogenesis agent to EGFR T790M mutant observed previously did not seem to suppress the development of secondary T790M mutation when it was administered simultaneously with EGFR-TKI. This has been observed in the NEJ 026 and RELAY studies in which the rates of secondary T790M were similar between the anti-angiogenesis agent plus erlotinib and monotherapy erlotinib groups 35,36 . Similarly, only a numerically lower secondary T790M rate by 8.3% was observed in EGFR-TKI plus bevacizumab group in the present study. Nevertheless, more real-world data is still required to clarify the impact regarding treatment of anti-angiogenesis agent and the development of secondary T790M. The limitation of the present study, firstly, is the potential bias due to retrospective nature per se. Therefore, the treatment-related complications of bevacizumab such as hypertension, peripheral edema and bleeding which may be associated with poor prognosis could not be recorded as comprehensive as a prospective study. Secondly, a small portion (9.8%) of patients of the propensity-score matched cohort received gefitinib, instead of erlotinib, as the first-generation EGFR-TKI and thus increased the data heterogeneity. However, an earlier large scale meta-analysis has suggested similar efficacy of the two drugs 37 and thus the heterogeneity is likely acceptable.
In conclusion, this study showcased an improved OS of EGFR-TKI plus bevacizumab in a real-world cohort of patients with advanced EGFR-mutant NSCLC in which patients with brain metastasis received most benefit from the regimen. Further study of bevacizumab treatment in brain metastasis-specific cohort of advanced NSCLC patients is warranted.