The results of this study demonstrated that bevacizumab biosimilar is equivalent in efficacy to bevacizumab in locally advanced and advanced NSCLC. Patients receiving previous treatment, patients receiving regimens other than in combination with chemotherapy, patients with rare genetic mutations, and patients with brain metastases also benefit clinically from both products. The AE spectra and incidence rates of the two products were similar.
Bevacizumab is a humanized monoclonal antibody targeting VEGF and is prepared by recombinant DNA technology. By combining with VEGF, it can inhibit the binding of VEGF to its receptor and block the signaling pathway of angiogenesis in tumor tissues. Bevacizumab has become an important component of systemic therapy for advanced NSCLC without genetic mutations. A biosimilar is an approved biological product that is highly similar to the original drug with no clinically remarkable difference in safety, purity, or potency. Biosimilar therapy is an important avenue to reduce patient expenditure and the financial burden of national healthcare systems while maintaining therapeutic effect (Science, Accessed January 2021). However, a biosimilar is not an exact copy of its reference product and requires vigilance and concern in clinical application.
Current prospective clinical trials have confirmed that bevacizumab biosimilar in combination with platinum-containing two-drug chemotherapy has similar efficacy and safety compared with bevacizumab in patients with untreated advanced non-squamous NSCLC. In 2019, China National Medical Products Administration approved the marketing of bevacizumab biosimilar developed by Qilu Pharmaceutical Co., Ltd. (trade name: Encoda) with all indications approved for bevacizumab in China. However, no clinical data has verified the efficacy and safety of the biosimilar in clinical application. Existing trials have excluded patients receiving previous treatment, patients receiving combination with targeted therapy or immunotherapy, patients with brain metastases, patients with rare genetic mutations (such as EML4–ALK rearrangement), and patients with ECOG scores greater than 2. No consensus has been reached on the efficacy and safety of bevacizumab biosimilar in these populations. Among the published retrospective studies comparing bevacizumab biosimilar with bevacizumab, most have focused on patient clinical characteristics, cost-effectiveness, and economic impact rather than efficacy and safety. Only one study mentioned the statistical analysis of treatment modalities for 18 patients with NSCLC who used bevacizumab(Yang et al., 2022). Our study is not comparable to this one because of the lack of other information and its limited patient population. Our study greatly expanded the sample size and provided detailed information. It is the first retrospective study to evaluate the efficacy and safety of bevacizumab biosimilar.
Subject demographics, baseline disease characteristics. And exposure to treatment were well balanced and comparable between the two treatment groups. In the biosimilar group, the ORR was 28.9% (95% CI: 25.1–32.7%), and the DCR was 89.5%. The ORR and DCR of the control group were 30.9% (95% CI: 26.3–35.5%) and 87.3%, respectively. The unstratified ORR risk ratio was 0.934 with a 95% CI of 0.767–1.138 and a 90% CI of 0.792–1.103. The unstratified ORR risk difference was − 0.020 with a 95% CI of − 0.118–0.035 and a 90% CI of − 0.105–0.023. The definition of equivalence by FDA, PMDA, and EMA corresponds to the 90% CI of ORR hazard ratio in the range of 0.73–1.37; the 95% CI of the ORR HR is 0.729–1.371, and the 95% CI of the ORR risk difference is − 13–13%(Syrigos et al., 2021; Trukhin et al., 2021). In the present study, the CIs of ORR risk ratio and ORR risk difference were within these predefined equivalence margins. The estimated mPFS values were 6.27 (95% CI: 5.53–7.01) months in the biosimilar group and 4.93 (95% CI: 4.24–5.62) months in the reference group. The estimated HR for bevacizumab biosimilar and bevacizumab comparison was 1.084 (95% CI: 0.932–1.260, p = 0.296). Therefore, bevacizumab biosimilar and reference bevacizumab are equivalent in efficacy.
In different historical clinical trial results for bevacizumab biosimilar, the ORR ranges from 41.5–53.1%, and the mPFS is about 7.5 months(Chu et al., 2021; Reck et al., 2020; Reinmuth et al., 2019). Numerically, our study showed a lower ORR and a shorter PFS than other studies. In view of the differences between clinical application and clinical trials, such as the number of treatment lines, drug combinations, and dose intensity, the above results are not highly comparable. The subgroup analyses were performed based on the following subgroups: sex, age, pathology, stage, number of treatment lines, radiotherapy, combined treatment regimens, dose intensity, genetic mutations, brain metastasis, liver metastasis, and history of hypertension.
In the subgroup analysis of combined treatment regimens, the ORR of the biosimilar group was lower than that of the reference group when combined with chemotherapy and immunotherapy (22.58% vs. 33.64%, p = 0.048). Further logistic regression analysis showed that the influencing factors of ORR in patients combined with chemotherapy and immunotherapy were the number of treatment lines and combination with radiotherapy (p = 0.021, 0.008). Product type (i.e., bevacizumab or bevacizumab biosimilar) was not a main factor (p = 0.604). In a subgroup of patients with a history of hypertension, bevacizumab biosimilar obtained a longer PFS than reference bevacizumab (8.85 vs. 6.63 months, p = 0.047). The mean dose intensity of the biosimilar group was significantly higher than that of the reference group (8.28 vs. 7.37 mg/kg, p = 0.007) in this subgroup. Based on multivariate analysis, the product type was not the main factor affecting PFS (p = 0.595). Among the remaining subgroups, the ORR and PFS analyses support the equivalence between the two groups.
The number of treatment lines, combined treatment regimens, and radiotherapy were the significant factors affecting the PFS of both groups (p < 0.001, p = 0.001, p = 0.039). Different genetic mutations, dose intensity, brain metastases, liver metastases, and a history of hypertension were not the main factors affecting PFS (p = 0.627, 0.946, 0.142, 0.062, 0.030).
The treatments combined with chemotherapy and another targeted therapy (biosimilar group: ORR = 52.94%, mPFS = 15.25 months; reference group: ORR = 65.00%, mPFS = 8.17 months) and combined with another targeted therapy group (biosimilar group: ORR = 33.33%, mPFS = 11.74 months; reference group: ORR = 50.00%, mPFS = 18.20 months) showed better efficacy than other treatment regiments. Among these treatments, approximately 80% of other targeted therapies are EGFR–TKI. Preclinical studies have shown that VEGF and EGFR share a common downstream signaling pathway(Jiang et al., 2019; Larsen et al., 2011; Le et al., 2021; Yang, 2019), but the clinical trial data of EGFR–TKI combined with bevacizumab are still immature and have many uncertainties(Ma et al., 2021). Our study confirmed the benefits of bevacizumab (or bevacizumab biosimilar) combined with EGFR–TKI in ORR and PFS but failed to obtain OS results for all patients because of the short follow-up period. Previous studies have shown that patients with exon 19 deletions have a better prognosis after EGFR–TKI treatment than those with 21 p.L858 mutations(Hsu et al., 2018). However, our study revealed similar ORR and PFS benefits between the two mutation types in both groups. The synergistic antiproliferative effects of EGFR–TKI and antiangiogenic treatment might eliminate the prognostic differences caused by genetic mutations. More prospective studies or in-depth retrospective clinical data are expected to better explore the above conclusions.
In addition to EGFR mutations, studies on patients with other rare mutations receiving targeted therapy combined with bevacizumab are few, and all of which are exploratory studies with small samples. The American Society of Clinical Oncology reported a study from China in which 16 patients with EML4–ALK rearrangement receiving crizotinib and bevacizumab have a mPFS of 13.0 months. Our study included 89 patients that harbor other rare driving gene mutations, including ALK rearrangement, ROS1 rearrangement, RET rearrangement, BRAF mutation, HER2 mutation, KRAS mutation, and MET aberration. These populations have also been proven to benefit from bevacizumab or bevacizumab biosimilar.
In the previous phase III clinical trial studies, only the AVAiL and AVAPERL studies used bevacizumab at 7.5 mg/kg, and the other studies used 15 mg/kg. These studies lacked the ability to directly compare the two doses of bevacizumab. With 7.5 mg/kg as the boundary line, our study showed that no statistical differences in the ORR and PFS between the low- and high-dose groups (biosimilar group: 33.19% vs. 25.85%, 6.70 vs. 5.90 months; reference group: 29.79% vs. 32.50%, 4.80 vs. 5.00 months). In addition, the patients with the brain metastases (BMS) also benefit from bevacizumab or biosimilar (biosimilar group: ORR = 26.67%, mPFS = 6.07 months; reference group: ORR = 23.60%, mPFS = 4.43 months).
Similar incidences of TEAEs at any grade were observed in the biosimilar group and the reference group (421 [76.41%] vs. 283 subjects [71.65%], p = 0.098), and most of them were classified as grade 1 or 2 events. No statistically significant difference in the incidence of grade 3 or 4 TEAEs was observed between the two groups (22.14% vs. 19.49%, p = 0.324). No fatal effects happened. Among the treatment-related AEs in the biosimilar group, neutropenia had the highest incidence (22.87%), followed by anemia (16.15%), alopecia (14.70%), nausea (12.89%), fatigue (10.16%), thrombocytopenia (10.16%), hypertension (9.62%), and fever (8.71%). In the reference group, neutropenia had the highest incidence (24.05%), followed by anemia (16.46%), alopecia (13.67%), fatigue (11.90%), thrombocytopenia (10.89%), nausea (10.63%), loss of appetite (9.37%), and bleeding (7.34%). In general, the AE spectra and AE rates of bevacizumab biosimilar and reference bevacizumab were similar. The overall incidences of grade 3 and 4 TEAEs were low and similar, indicating that bevacizumab biosimilar and bevacizumab have favorable safety profiles.
The incidence of hypertension at any grade in the biosimilar group was higher than that in the reference group (9.62% vs. 5.82%, p = 0.033), but no differences in the incidence of hypertension were observed at grades 3 and 4 (1.63% vs. 0.51%, p = 0.093). We further analyzed the clinical data to understand the origin of the differences. Among patients over 70 years of age who received high-dose treatment, the incidence of hypertension caused by the biosimilar (n = 48) was significantly higher than that of reference bevacizumab (n = 32; (37.5% vs. 15.6%, p = 0.034). Among patients over 70 years of age with a history of hypertension, the incidence of hypertension caused by biosimilar (n = 25) was higher than that of reference bevacizumab (n = 17; 40% vs. 23.5%, p = 0.266). Among these patients, we first hypothesized that differences in the incidence of hypertension might be attributed to some patients in the subgroup over 70 years of age. We then compared the incidence of hypertension in patients over 70 years of age after excluding patients with high-dose therapy and a history of hypertension. The results showed that the incidence of hypertension was similar between the two products (17.3% vs. 10.0%, p = 0.511). After imbalanced factors were eliminated, no statistically significant difference in the incidence of hypertension of any grade was found between the biosimilar group (n = 541) and reference group (n = 392; 7.9% vs. 5.1%, p = 0.087).
We acknowledge the limitations of the retrospective study design, which is inherently affected by selection bias and missing data. Moreover, reliance on electronic health records may mean that some events may be underestimated. Therefore, larger prospective studies are needed to confirm our findings.