Background: There are many controversies concerning the best management of epidermal growth factor receptor (EGFR)-mutant non-small-cell lung cancer (NSCLC) patients with brain metastases (BMs). The use of upfront EGFR tyrosine kinase inhibitors (TKIs) and the withholding of local therapies or upfront radiation therapies (RTs) remain controversial. Available treatment options include local therapies such as whole-brain radiation therapy (WBRT), stereotactic radiosurgery (SRS) and surgery, EGFR-TKIs, and chemotherapy. However, the optimal management of combination therapies is still under consideration.
Patients and methods: A total of 45 EGFR-mutated NSCLC patients with BMs were included. All patients successively received EGFR-TKIs, RT (WBRT or SRS), and chemotherapy between 2010 and 2015 at Zhejiang Cancer Hospital. Patient follow-up was conducted by telephone until February 2017. The treatment response was evaluated, and survival data were collected and analyzed by Kaplan–Meier analysis and the Cox regression method.
Results: The median overall survival (OS) was 28 months. Patients with the exon 19 deletion showed the strongest trend toward a longer median OS compared to patients with the exon 21 L858R mutation (not reached vs 26.5 months, P=0.0969). There was no difference in OS between the upfront RT group and the deferral group (26.5 vs 28 months, P=0.57), and similar results were found between the first-line chemotherapy group and the EGFR-TKI group (28 vs 23.2 months, P=0.499). In multivariate analysis, the prognosis correlated with EGFR mutation type (P=0.017).
Conclusion: EGFR-mutant NSCLC patients with BM benefited from the combination and sequential therapies of EGFR-TKIs, chemotherapy, and RTs. Patients with the EGFR exon 19 deletion may have a better OS. However, the optimal timing of RT interval remains to be explored.
Keywords: epidermal growth factor receptor, tyrosine kinase inhibitors, brain metastases, non-small-cell lung cancer, pemetrexed, whole-brain radiation therapy
Brain metastases (BMs) are a common cause of morbidity and mortality in patients with non-small-cell lung cancer (NSCLC), and BMs develop in ~25%–40% of patients with advanced adenocarcinomas; moreover, the incidence of BMs is still increasing.1,2 Patients with epidermal growth factor receptor (EGFR)-mutant NSCLC may have a higher likelihood of being diagnosed with BMs because of prolonged survival from targeted systemic agents and the increased quality of central nervous system imaging.3The median overall survival (OS) of an unselected population of EGFR-mutant and non-EGFR-mutant NSCLC patients with BMs reportedly ranged from 3 to 15 months,4 whereas the median OS after BMs of 19–58 months in patients with EGFR-mutant NSCLC was observed.5,6 Historically, therapeutic options for BMs have been limited to local therapies such as whole-brain radiation therapy (WBRT), stereotactic radiosurgery (SRS), surgery, or a combination of the above. Due to concerns about inadequate central nervous system penetration, chemotherapy is not typically a standard primary treatment for BMs.7 However, previously published studies describing the use of combined cisplatin and pemetrexed therapy confirmed good tolerability and efficiency in managing NSCLC patients with inoperable BMs.8,9 During the last decade, EGFR-tyrosine kinase inhibitors (TKIs) have been successfully employed in NSCLC patients based on the identification of EGFR gene mutations; however, EGFR-TKIs have also been demonstrated to be a potential treatment of choice for BMs from NSCLC patients harboring an activating EGFR mutation.10–16 In addition, some studies showed that the combination of RT and EGFR-TKIs produced superior outcomes for patients with EGFR mutations and BMs.5,6,17,18
There are still several controversies concerning the management of EGFR-mutant NSCLC patients with BMs. The use of upfront EGFR-TKIs and the withholding of local therapies or upfront radiation therapies (RTs) remain controversial. Available treatment options include local therapies such as WBRT, SRS and surgery, EGFR-TKIs, and chemotherapy. To evaluate the efficacy of EGFR-mutant NSCLC patients with BM receiving multiple regimens and to analyze the prognostic factors, we retrospectively investigated 45 patients with EGFR-mutant NSCLC who developed BM between 2010 and 2015 and were successively treated with EGFR-TKIs, pemetrexed-based chemotherapy and radiotherapy.
PATIENTS AND METHODS
In this study, we retrospectively enrolled and analyzed 45 EGFR-mutated NSCLC patients with BMs who systematically received EGFR-TKIs (icotinib, gefitinib, erlotinib, or osimertinib), pemetrexed-based chemotherapy, and local therapies (WBRT or SRS) between 2010 and 2015 at Zhejiang Cancer Hospital. All patients were histologically diagnosed with NSCLC, and EGFR mutations were detected by the amplification refractory mutation system analysis, which identifies tumor lesions with EGFR mutations. BM in these patients was confirmed by magnetic resonance imaging. All patients completed clinical and follow-up evaluations (Table 1). The research was approved by the ethical committee of Zhejiang Cancer Hospital, including verbal informed consent being obtained from all participants. We confirm that patient data confidentiality was maintained.
(To view a larger version of Table 1, click here.)
Evaluation of treatment responses and toxicity
The objective tumor response was assessed according to the Response Evaluation Criteria in Solid Tumors version 1.1, which was divided into complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD). Adverse reactions were evaluated based on the CTCAE 3.0 edition (Common Terminology Criteria for Adverse Events version 3.0) every month.
Using patient hospital records, follow-up registration records, and follow-up phone records, we collected clinical and survival information for the patients. The following data were collected and recorded: age, sex, smoking status, Karnofsky Performance Scale (KPS) at the time of BM, number of BMs and extracranial metastases, primary tumor histological type, EGFR mutation type, first-line treatment, therapy for BM, RT types, and interval between RT and diagnosis. Patients were categorized by age (<65 years, ≥65 years), sex (male, female), smoking status (never, slight, heavy), KPS (≥70, <70), extracranial metastases (yes or no), number of BMs (1 tumor, 2 tumors, more than 2 tumors), primary tumor histology (adenocarcinoma, others), EGFR mutation type (exon 19 deletion, exon 21 L858R mutation), first-line treatment (chemotherapy, EGFR-TKIs), chemotherapy (single agent, dual agents), and RT types (WBRT, SRS, or combination). Finally, a graded prognostic assessment (GPA) was calculated for each patient to determine whether the cohorts shared similar prognostic features.
Subgroup analysis was performed based on first-line treatment regimens; patients were divided into 2 groups: chemotherapy (n=28) and EGFR-TKIs (n=17). Patients were also divided into 2 groups according to whether they received upfront RT (with an interval between RT and diagnosis of ≤3 months), and it was found that 30 patients were treated with upfront RT and 15 received deferral RT instead. A major focus of the study was OS, which was defined as the duration of time from the start of therapy until death or the most recent follow-up.
Kaplan–Meier analysis was used to estimate OS, whereas log-rank testing was used to assess differences. OS was calculated from the date of BM diagnosis until the date of death. Finally, the Cox proportional hazard regression model (forward Wald method) was used for multivariate analysis of the groups to study the effect of prognostic factors from the Kaplan–Meier single-variant test (mutation, GPA, number of intracranial lesions, extracranial metastasis) and clinical factors (first-line treatment, interval of BM) and to evaluate which factors were associated with patient survival as well as to analyze differences in the survival curves for each subgroup. All statistical analyses were performed using SPSS software, version 20.0 (IBM Corporation, Armonk, NY, USA).
Baseline characteristics of patients
Twenty (44.4%) of the 45 NSCLC patients diagnosed with BMs who were enrolled in the study were male. The median age of these patients was 56 years (ranging between 38 and 70 years), and 7 patients (15.6%) had a KPS score of ≥90. A total of 16 patients (35.6%) were smokers, and 11 patients (23.9%) had heavy smoking histories. These NSCLC cases included 41 adenocarcinoma and 4 non-adenocarcinoma cases. The EGFR exon 19 deletion mutation occurred in 22 (48.9%) patients, and 23 (51.1%) patients had the EGFR exon 21 L858R mutation. Thirty-seven patients received icotinib (125 mg/3 times daily), 4 patients gefitinib (250 mg/d), 3 osimertinib (80 mg/d), and 1 erlotinib (150 mg/d). While 14 (31.1%) patients had a single intracranial lesion, 5 (11.1%) patients had 2 lesions, 26 (57.8%) had more than 2 lesions, and 27 (60.0%) had extracranial metastasis. In total, 30 (66.7%) patients received upfront RT. A total of 28 patients (62.2%) were treated with pemetrexed-based chemotherapy as the first-line treatment, while 17 (37.8%) received first-line EGFR-TKIs instead. Thirty-eight patients received WBRT (30Gy/10f/2W), 5 received SRS, and 1 received the combined SRS and WBRT. Twenty-six (57.8%) patients had a GPA score between 0 and 2.0, and 19 (42.2%) patients had a score between 2.5 and 4.0. The patient characteristics at baseline are detailed in Table 1.
Survival analysis showed that the median OS from diagnosis of BM was 28 months for the whole cohort (95% CI, 17.3–38.7 months) in Figure 1. Patients who were younger than 65 years had a significantly longer OS (32.8 months; 95% CI, 23.7–41.9) than that of elderly patients (15.6 months; 95% CI, 8.1–23.1; P=0.001). Patients with the exon 19 deletion showed a trend toward a longer median OS (not reached) than that of patients with the exon 21 L858R mutation (23.2 months; 95% CI, 0.22–1.13; P=0.0969) in Figure 2A. There was no difference in OS between the upfront RT group and the deferral group (28.0 vs 26.5 months, P=0.74) in Figure 2B. Similar results were found between the first-line chemotherapy group and the EGFR-TKI group (28.0 vs 23.2 months, P=0.25) in Figure 2C. And we also found no statistical significance between the GPA score of 0–2 group and 2.5–4 group (P=0.69) in Figure 2D.
(To view a larger version of Figure 2, click here.)
Furthermore, the median intercranial progression-free survival (iPFS) rate of the whole cohort was 17.7 months (95% CI, 12.4–23.0 months). However, the iPFS of 8 patients could not be assessed because the intercranial lesions were not evaluable. The first-line chemotherapy group showed an iPFS of 17.78 months, and the EGFR-TKI group showed and iPFS rate of 13.00 months; however, there was no statistically significant difference (P=0.489). In the chemotherapy group, 25 patients were evaluable, and 15 received upfront RT.
As for intracranial lesions, 9 (19.6%) patients had CR, 21 (45.7%) had PR, 15 (32.6%) had SD, and 1 (2.2%) had PD in the EGFR-TKI group. The objective response rate was 65.2%. In the chemotherapy group, although the intracranial response could not be evaluated in 7 patients, 7 (19.4%) patients had CR, 20 (55.6%) had PR, 10 (27.8%) had SD, and 1 (2.8%) had PD. The objective response rate was 75.0%. Patients who were younger than 65 years had a significantly longer OS (32.8 months; 95% CI, 23.7–41.9) than that of elderly patients (15.6 months; 95% CI, 8.1–23.1; P=0.001).
In the multivariate analysis, the prognosis independently correlated with the EGFR mutation type (adjusted hazard ratio, 3.27; 95% CI, 1.24–8.62; P=0.017). Notably, gender, age, smoking history, KPS, number of metastatic lesions in the brain, absence of extracranial disease, first-line treatment, and RT types were not found to significantly influence the prognosis. Although patients with extracranial metastases showed the strongest trend toward decreased OS, it was not statistically significant (P=0.070). Patients with a higher GPA score had a better survival rate than that of patients with a lower score; however, this also was not statistically significant (P=0.206) (Table 2).
(To view a larger version of Table 2, click here.)