Efficacy of EGFR-TKI therapy in patients with brain metastases from non-small-cell lung cancer : A meta-analysis

Purpose: This meta-analysis aimed to systematically evaluate the efficacy of tyrosine kinase inhibitors of epidermal growth factor receptor (EGFR-TKIs) for patients with brain metastases (BM) from nonsmall-cell lung cancer (NSCLC), and to compare this treatment modality to chemotherapy or radiotherapy. Methods: PubMed, EMBASE (OvidSP), the Cochrane Central Register of Controlled Trials (CENTRAL), Web of Science and ASCO Annual Meeting Abstracts were searched. Controlled clinical studies that compared the efficacy of EGFR-TKIs with chemotherapy or radiotherapy for NSCLC patients with BM were included in the analysis. Efficacy indicators included overall survival (OS), local progression-free survival (LPFS) and objective response rate (ORR). Results: The final sample consisted of 24,637 NSCLC patients with BM from 11 clinical studies. In primary efficacy analysis, it was found that EGFR-TKIs were significantly superior to chemotherapy or radiotherapy in terms of ORR (odds ratio (OR) = 2.10, p = 0.035), OS (hazard ratio (HR) = 0.78, p = 0.011) and LPFS (HR = 0.60, p < 0.001). Conclusion: Among the patients with BM from NSCLC, EGFR-TKIs exhibit a therapeutic advantage over chemotherapy or radiotherapy, which is reflected in the elevation of ORR and improvement in OS and LPFS.


INTRODUCTION
Lung cancer is the most common cancer and is the major cause of cancer-related deaths, accounting for 13.0 % of new cancer cases and 19.4 % of cancer-related deaths in men and women according to the 2012 statistics [1].About 85 % of the lung cancer cases are classified as NSCLC [2], which has a relatively high risk of BM, ranging from 20 to 40 % [3,4].The prognosis of NSCLC patients with BM is very unsatisfactory, with median OS time being < 3 months if untreated [5].
The therapies for NSCLC patients with BM include surgery, stereotactic radiosurgery (SRS), chemotherapy, whole brain radiation therapy (WBRT), EGFR-TKIs or any combinations thereof.However, chemotherapy is not a routine treatment because of the inability of many systemically active chemotherapeutic drugs to cross the blood-brain barrier (BBB).Surgery and SRS are not ideal options for NSCLC patients with BM considering the poor outcome.WBRT has been regarded as the standard treatment plan for multiple BM; it can stabilize or shrink tumors in at least half of the patients, but many patients experience tumor recurrence either at the original sites or at new sites [6].
As micro-molecularly-targeted drugs, EGFR-TKIs, such as gefitinib or erlotinib, can penetrate the BBB in a certain proportion [7,8].This is an important characteristic to consider when the treatment options of patients with BM from NSCLC are evaluated.In addition, EGFR-TKIs are believed to possess the ability to reinforce radioactive sensitivity through multiple means [9,10].
Many studies have shown the responses of NSCLC patients with BM to EGFR-TKIs [11][12][13][14], but most of the studies involved small sample sizes, contained only case reports, or were single-arm.Moreover, negative conclusions were reached based on the findings from some studies.For example, OS, PFS and toxicity in the WBRT + SRS + erlotinib group were completely less satisfactory when compared to those of the WBRT + SRS group in a phase III clinical trial.Three meta-analyses on the benefit of EGFR-TKIs among patients with BM from NSCLC have also been published, but only one included 16 single-arm studies [15] and the other two [16,17] respectively evaluated the efficacy and safety of TKIs and radiotherapy in relatively few participants.
In 2015, Lin [18] found extended period of survival in NSCLC patients with BM who were receiving gefitinib or gamma knife radiosurgery by analyzing medical records of 23,874 target patients based on the National Health Insurance Research Database of Taiwan from 2004 to 2010.Results from such large sample size studies provide more dependable data.Therefore, in the present study, we performed this meta-analysis to systematically examine the efficacy of EGFR-TKIs for NSCLC patients with BM, and the outcome of this treatment modality was compared with those of chemotherapy and radiotherapy.This pooled evaluation, based on controlled studies can assist doctors to make better clinical decisions for the treatment of NSCLC patients with BM.

Data sources
Our meta-analysis was based on the preferred reporting items for systematic reviews and meta-analyses (PRISMA) Statement, in which PubMed, EMBASE (OvidSP), the Cochrane Central Register of Controlled Trials (CENTRAL, last issue), Web of Science and ASCO (The American Society of Clinical Oncology) annual meeting abstracts which were painstakingly searched for the presence of relevant papers that were published from January 2000 until December 2016.The keywords were as follows: non-small cell lung cancer OR non-small cell lung carcinoma OR NSCLC OR large cell lung cancer OR lung adenocarcinoma OR lung squamous cell carcinoma OR lung squamous carcinoma; epidermal growth factor receptor OR EGFR OR erbB-1 OR erlotinib OR gefitinib OR Tarceva OR Iressa OR ZD1839 OR icotinib OR afatinib OR Gilotrif; and brain metastasis OR central nervous system metastasis OR CNS metastasis OR cerebral metastasis.The subject heading search was also performed in specific databases.Moreover, other conference papers and the reference lists of selected studies were searched for manually.The search for publications was not restricted to type of publication and language in which they were published.

Study selection
Two of the authors (ZW and HLJ) reviewed the titles and abstracts of all papers identified from electronic sources and those from manual retrieval, respectively.Only studies that met the following eligibility criteria were included for the purpose of quantitative synthesis: (a) controlled clinical studies (even retrospective clinical designs in consideration of the small number of eligible publications) in which the experimental group received EGFR-TKIs alone or a combination of EGFR-TKIs with chemotherapy or radiotherapy and in which the control group received only chemotherapy, radiotherapy or a combination of chemotherapy and radiotherapy; (b) the patients in the studies or in the subgroup analyzed and confirmed to have NSCLC with brain metastasis; (c) original papers that contained adequate data to estimate the indicators of efficacy with a 95 % confidence interval (95 %CI).Repeated publication and nonoriginal literature (e.g., comment, letter, review, etc.) were excluded.If no agreement could be reached after a discussion between the two reviewers in terms of study selection, a third investigator (HCH) made the final decision.

Efficacy indicators
The ORR was defined as the percentage of complete response and partial response among the evaluated patients.OS referred to the time from randomization to death regardless of the cause of death.LPFS was thought as the length of time from randomization to central nervous system, neurological or intracranial progression or death.In this meta-analysis, the odds ratio (OR) for the ORR and the hazard ratios (HRs) for OS and LPFS were primary indicators for the assessment of the efficacy of EGFR-TKIs for NSCLC patients with BM.

Data extraction and quality assessment
In the meta-analysis, the following information was collected independently by two authors (ZW and HLJ) from each study: first author, publication year, country of study, study design, treatments in the different groups, number of patients, endpoint types, patient characteristics (age, gender, stage, response status, EGFR mutation status) and outcome measures (response rate, median OS and LPFS, HRs for OS and LPFS and their 95 % CIs).Results that were obtained by multivariate analysis were preferred.If the HRs for OS or LPFS were not provided in the papers, log-transformed HRs and variance were estimated from Kaplan-Meier curves with the Engauge graphic software package [19].Moreover, the risks of bias of the included studies were evaluated independently by two authors (ZW and HLJ) on the basis of The Cochrane Handbook for Systematic Reviews (version 5.1.0).HCH was still the final decisionmaker when no consensus could be reached between the other two reviewers.
This research was approved by the Ethical Committee of Department of Biostatistics, School of Public Health, Fudan University, Shanghai, China according to the Declaration of Helsinki promulgated in 1964 as amended in 1996 [20].

Statistical analysis
The analysis was based on the intention-to-treat (ITT) principle.Potential heterogeneity was first detected among the studies using Cochran's Qstatistic and I 2 score [21], which is thought as statistically significant if the p-value for heterogeneity is less than 0.10 or if I 2 is greater than 50 %.If heterogeneity was found, the random-effects model [22] was applied to pool all effect measures including odds ratio for the objective response rate and HRs for OS and LPFS; otherwise, the fixed-effects model (inverse variance method) [23] was used.To assess the presence of possible publication bias, Egger's regression asymmetry test with a funnel plot was performed [24].Its significant level was 0.10.The sensitivity analysis was performed to examine the influence of the study design, study quality and treatment plan on the evaluation of the study's efficacy.
The statistical software package used was STATA 11.2 (Stata Corporation, College Station, TX, USA).All p-values were two-sided.For the examination of the pooled effect indicators, p < 0.05 was considered as statistical significance.

Characteristics of studies
From the 266 studies that were screened, 216 articles were excluded because of apparent nonconformity with the eligibility criteria, and 50 fulltext papers were accessed to further evaluate their eligibility.Next, 15 studies were enrolled in the qualitative synthesis, and 11 studies (see 18, 25 -34, Table 1) that met all criteria were finally selected for the quantitative synthesis.A flowchart of the selection process is presented in Figure 1.The 11 eligible studies included 2 phase III trials, 3 phase II trials, 1 undefined phase trial and 5 retrospective clinical studies.The final sample consisted of 24,637 patients with BM from NSCLC, which included 3,773 in the EGFR-TKIs arm and 20,864 in the control group.Five of the 11 studies were from the Chinese mainland, 2 from Taiwan, 1 from Germany, 1 from the United Kingdom, 1 from the United States and 1 from Switzerland.All subjects in study 1(one) were EGFR-mutant, and 2 studies provided the subgroup analysis for EGFR-mutant and EGFR wild-type patients, but the mutation status for the rest of the patients could not be verified or analyzed.Of the 11 eligible studies, 3 featured EGFR-TKI combination therapy compared with combined radiotherapy or chemotherapy, 5 featured EGFR-TKI combination therapy compared with radiotherapy alone, and 3 featured EGFR-TKI alone compared with chemotherapy or radiotherapy alone.All studies included were peer-reviewed.The characteristics of the included studies are summarized in Table 1.

Publication bias
The result of Egger's asymmetry test did not show any evidence of publication bias (p = 0.683 for ORR, p = 0.558 for OS, p = 0.427 for LPFS, Figure 7).

DISCUSSION
The meta-analysis carried out in the present study included 24,637 NSCLC patients with BM from 6 randomized controlled trials and 5 retrospective clinical studies.The vast majority of the patients came from East Asia, including the Chinese mainland and Taiwan.This metaanalysis was aimed at assessing the value of EGFR-TKIs therapy in NSCLC patients with BMs.The results indicated that EGFR-TKIs were significantly superior to chemotherapy or radiotherapy in terms of ORR (OR = 2.10, p = 0.035), OS (HR = 0.78, p = 0.011) and LPFS (HR = 0.60, p < 0.001).The data suggested that EGFR-TKIs therapy for advanced NSCLC patients with BM not only could improve the short-term quality of life, but also could prolong the overall lifespan.
Systemic chemotherapy for central nervous system (CNS) metastases has only played a limited role in the treatment of patients with BM from NSCLC because of the existence of the blood-brain barrier (BBB).However, as a micromolecule and highly permeable compound, an EGFR-TKI can pass through the BBB in a certain proportion and target intracranial lesions.A pharmacokinetic study indicated that the penetration rates of erlotinib and gefitinib in the cerebrospinal fluid (CSF) were 2.77 ± 0.45 % and 1.13 ± 0.36 %, respectively [35].In clinical practice, EGFR-TKIs were thought to be an important treatment option for these patients.The present meta-analysis strengthened the evaluation of the therapeutic efficacy of EGFR-TKIs in a target population by pooling all the existing controlled studies, in which three primary measures favored EGFR-TKI therapy.
WBRT is considered as the standard treatment plan for multiple brain metastases because radiation not only kills the tumor cells, but also damages endothelial cells of the BBB and thus increases its permeability.This characteristic can increase the concentration of EGFR-TKIs in the brain without a dose escalation and it also allows for the combination of EGFR-TKI therapy and WBRT [36].In addition, EGFR-TKIs are able to sensitize tumor cells to radiation and improve local control rates.Thus, the combination of EGFR-TKIs with concurrent radiotherapy seems to have a strong theoretical basis for the production of a synergistic action that enhances the therapeutic efficacy and reduces the side effects of each therapy.In a phase II trial of erlotinib with concurrent WBRT for patients with BM from NSCLC that was published in Journal of Clinical Oncology in 2013 [37], the ORR was 86 %, and the median OS was 11.8 months (95 % CI: 7.4 -19.1 months).Meanwhile, no increase in neurotoxicity, or any grade >/= 4 toxicity, was detected.A metaanalysis [16] also suggested, by evaluating the therapeutic effects among 980 NSCLC patients with BM, that EGFR-TKIs combined with radiotherapy produced superior efficacy and safety.The efficacy of combination therapy was similarly confirmed by a sensitivity analysis in this study, which was performed by the inclusion of 24,289 participants from the studies with groups that received a combination of radiotherapy and EGFR-TKIs.In this analysis, the ORR, OS and LPFS of patients who received concurrent EGFR-TKIs and radiotherapy were significantly superior to the same indicators of patients who received radiotherapy alone, chemotherapy alone or a combination of chemotherapy and radiotherapy (OR for ORR: 1.82, p = 0.012; HR for OS: 0.78, p = 0.024; HR for LPFS: 0.59, p = 0.036).
Another sensitivity analysis was performed to explore the efficacy of EGFR-TKIs in current clinical trials with randomization and comparison.The difference in the LPFS was still statistically significant (HR: 0.63, p = 0.009), which demonstrated that EGFR-TKIs could more effectively control local progression of cancer.Although the ORR did not favor EGFR-TKIs (OR: 4.81, p = 0.054), the ORRs in the EGFR-TKIs treatment group were clearly higher than those in the control group in all three clinical trials that reported the ORR (72.9 vs 24.2 %, 54.

Figure 1 :
Figure 1: PRISMA flow diagram of the identification process

Figure 3 :
Figure 3: Forest Plot of the hazard ratio for the overall survival according to primary efficacy analysis

Figure 4 :Figure 5 :Figure 6 :
Figure 4: Forest Plot of the hazard ratio for the LPFS according to primary efficacy analysis

Table 1 :
Characteristics of the studies included in the meta-analysis