Bevacizumab plus erlotinib versus erlotinib alone for advanced EGFR-mutant non-small cell lung cancer: a meta-analysis of randomized clinical trials

Epidermal growth factor receptor (EGFR) mutation represents the most frequent driver-gene alteration in non-small cell lung cancer (NSCLC), and 90% of EGFR mutations involving a deletion within exon 19 (19del) or a leucine-to-arginine substitution in exon 21 (L858R) [1, 2]. Erlotinib, a small-molecule EGFR tyrosine kinase inhibitor (TKI), is recommended as a first-line therapy for patients with advanced NSCLC harboring the two common EGFR mutations [3]. However, most NSCLC patients treated with erlotinib develop therapeutic resistance, with the median progression-free survival (PFS) of 9.7–13.1 months [4, 5]. To overcome this acquired resistance and improve the PFS, several studies [6‐10] have investigated the combination of bevacizumab and erlotinib for the NSCLC patients with EGFR mutations. Most of these studies (ARTEMIS [9], NEJ026 [8], JO25567 [7], BEVERLY [10]) have displayed better PFS in the combination therapy than erlotinib monotherapy, but no overall survival (OS) benefit in the combination therapy. Whether this combination therapy was more effective than monotherapy? Some meta-analyses tried to answer this question. For example, Chen’s meta-analysis [11] and Landre’s meta-analysis [12] both showed that the combination therapy significantly prolonged PFS but no OS compared to monotherapy. However, the conclusions of these meta-analyses were controversial, because the OS results of both ATEMIS study and NEJ026 study were immature at that time. As the OS results of the two studies were published in 2021, we conducted a meta-analysis to provide more reliable and stable evidence to determine the PFS and OS benefit in the combination therapy for advanced EGFR-mutant NSCLC. A secondary objective was to investigate whether efficacy differed between 19del and L858R mutation or between with brain metastasis and without brain metastasis.

Erlotinib, a first-generation reversible TKI, targets the EGFR signaling pathway. Bevacizumab, an anti-vascular endothelial growth factor (VEGF) monoclonal antibody, targets the VEGF signaling pathway. Additional clinical benefits may be conferred by the dual blockade of the EGFR and VEGF pathways that critical to tumor growth, metastasis and angiogenesis [15, 16]. Therefore, quite a few clinical studies [6‐10] had attempted this combination therapy: erlotinib plus bevacizumab. Although some clinical studies had achieved prolongation of PFS, there was no corresponding prolongation of OS. This meta-analysis is motivated by uncertainty about the value of combination therapy compared with monotherapy.

According to the results of this meta-analysis, the combination therapy significantly prolonged PFS (Fig. 2A), but did not prolong OS (Fig. 2B) compared to monotherapy for advanced EGFR-mutant NSCLC patients. Although final survival outcomes from study ATEMIS and NEJ026 were included in this meta-analysis, there was still no prolongation of OS, consistent with the findings of three previous meta-analyses [11‐13]. Based on the advantages of dual blockade of EGFR and VEGF pathways, the prolongation of PFS by combination therapy is currently indisputable, while there are other relevant reasons for the failure to prolong OS. For instance, OS might have been influenced by the subsequent treatments used after disease progression, especially the use of third-generation TKI osimertinib. In the study ATEMIS, more patients from the monotherapy group received osimertinib as subsequent treatment than in the combination therapy group (29.2% vs 17.2%), and this ratio was 28.9% vs 23.3% in another study [6]. In addition, the number of patients receiving subsequent anticancer treatment was also greater in the monotherapy group in the ATEMIS study (50.0% vs 33.8%). Thirdly, Grade ≥ 3 adverse events occurred more frequently in the combination therapy group than in the monotherapy group [6, 9]. Therefore, the combination therapy might be compromised because of the above problems.

In general, the L858R mutation subgroup derives less benefit from first-line EGFR TKIs than the 19del mutation subgroup, which has been repeatedly demonstrated in previous clinical trials and real-world studies [17‐19]. The prognosis of patients with L858R mutation urgently needs to be improved in the era of TKI targeted therapy [20]. Would adding a VEGF inhibitor in combination with the TKI change the subgroup’s outcome? Therefore, subgroup analysis was conducted in this meta-analysis. In 19del mutation subgroup, the HR for PFS was 0.60 with a 95% confidence interval of 0.47 to 0.76 (p < 0.0001), and the HR for OS was 1.00 with a 95% confidence interval of 0.73 to 1.37 (p = 1.00) (Fig. 3A, B). In L858R mutation subgroup, the HR for PFS was 0.59 with a 95% confidence interval of 0.46 to 0.77 (p < 0.0001), and the HR for OS was 0.80 with a 95% confidence interval of 0.58 to 1.11 (p = 0.18) (Fig. 3C, D). The subgroup analysis showed that the combination therapy only prolonged PFS, but not OS, regardless of L858R or 19del mutation, which was consistent with the results of overall population. When comparing the HR value alone, both PFS and OS of L858R mutation were improved more significantly than those with 19del mutation. This suggests that the combination therapy can reverse the poor prognosis of L858R mutation subgroup, and the combination therapy may be one of the treatment options for patients with L858R mutation [21, 22].

Presence of brain metastasis is a well-known adverse prognostic factor and common site of disease progression in EGFR-mutant NSCLC. Previous study showed that the anti-VEGF drug bevacizumab was effective in treating brain metastasis [23], and also could significantly reduce the incidence of brain metastasis [24]. In two retrospective studies [25, 26], bevacizumab combined with TKI prolonged the OS in patients with brain metastasis compared to TKI monotherapy. Of all the 5 RCT studies, only 2 studies [8, 9] conducted further subgroup analysis for patients with brain metastasis, and thus only the 2 studies were included. The subgroup analysis suggested that the combination therapy not only provided significant improvement in PFS (HR = 0.60, 95% CI 0.39–0.90; p = 0.01), but also had a positive trend to prolong OS (HR = 0.69, 95% CI: 0.46–1.02; p = 0.06) in patients with brain metastasis (Fig. 4A, B). The statistical P-value did not reach significant significance, which may be explained by the following reasons. First of all, only two RCTs were included. Secondly, the treatment-related adverse effects associated with the combination therapy [6, 8, 9]. However, it had given us good enlightenment and pointed out the direction of future research. It is the only way to include more large-scale RCT studies.

It is worth mentioning that osimertinib has replaced erlotinib as the preferred first-line treatment for advanced EGFR-mutant NSCLC patients and second-line treatment for patients with EGFR T790M mutation. Can osimertinib combined with bevacizumab further prolong PFS and OS in patients with EGFR-mutant NSCLC? Several studies had presented the results. BOOSTER [27] and WJOG8715L [28] were two studies of second-line treatment, and neither study showed an extension of PFS and OS. WJOG9717L [29] was a study of first-line treatment, and it also showed that PFS was not prolonged in the combination group, and the OS data were immature. Despite only three phase II studies, the prospects for combination therapy are not promising.

This meta-analysis reveals that the combination of erlotinib and bevacizumab can prolong progression-free survival but not overall survival compared to erlotinib alone in advanced EGFR-mutant non-small cell lung cancer. For patients with brain metastasis, the combination therapy not only can prolong progression-free survival, but also has a tendency to prolong overall survival. This combination therapy model deserves to be considered in the future, especially for patients with brain metastasis.