The addition of bevacizumab to erlotinib continued to provide a significant progression-free survival benefit over erlotinib alone when used in the frontline treatment of patients with EGFR-mutated, nonsquamous non–small cell lung cancer, according to data from the final analysis of the phase 3 BEVERLY trial.
The addition of bevacizumab (Avastin) to erlotinib (Tarceva) continued to provide a significant progression-free survival (PFS) benefit over erlotinib alone when used in the frontline treatment of patients with EGFR-mutated, nonsquamous non–small cell lung cancer (NSCLC), according to data from the final analysis of the phase 3 BEVERLY trial (NCT02633189).1
The data, which were virtually presented during the 2021 ESMO Congress, showed that at a median follow-up of 36 months, the median investigator-assessed PFS with bevacizumab plus erlotinib was 15.4 months (95% CI, 12.2-18.6) vs 9.6 months (95% CI, 8.2-10.6) with erlotinib alone (adjusted hazard ratio [HR], 0.66; 95% CI, 0.47-0.92; P = .015).
The statistically significant difference was confirmed in a multivariable Cox model that was adjusted by age, gender, ECOG performance status, smoking history, type of mutation, and center size, according to lead study author Maria Carmela Piccirillo, of the Istituto Nazionale Tumori, IRCCS, Fondazione G. Pascale in Napoli, Italy.
These data were confirmed in the blinded independent central review (BICR) analysis, where the doublet was found to result in a median PFS of 14.8 months (95% CI, 12.0-18.3) vs 9.6 months (95% CI, 7.1-10.6) with the monotherapy (adjusted HR, 0.68; 95% CI, 0.48-0.96; P = .027).
“In the BEVERLY trial, with Italian [patients with] EGFR-mutated nonsquamous NSCLC, the addition of bevacizumab to first-line erlotinib significantly prolonged PFS, significantly increased response rate, non-significantly prolonged overall survival [OS], did not modify quality of life [QoL], and did not produce unexpected safety issues,” Piccirillo said in a presentation on the data. “Bevacizumab plus erlotinib might be considered as a first-line therapeutic option in patients who cannot receive osimertinib [Tagrisso].”
Tumor angiogenesis is known to be a potential mechanism of resistance to EGFR TKIs in patients with NSCLC whose tumors harbor EGFR mutations. Findings from prior trials have indicated that the addition of anti-VEGF monoclonal antibodies, like bevacizumab, to EGFR TKIs can result in prolonged PFS.
Data from a phase 2 Japanese trial showed that the combination of bevacizumab and erlotinib resulted in a median PFS of 16.0 months vs 9.7 months with erlotinib alone (HR, 0.54; 95% CI, 0.36-0.79) in this patient population, thus inspiring the launch of the phase 3 BEVERLY trial, Piccirillo explained.
The phase 3 trial enrolled patients who were 18 years of age or older who had stage IV or IIIB nonsquamous NSCLC with an activating EGFR mutation and an ECOG performance status of 0 to 2. If patients had a tumor that harbored an EGFR T790M mutation and exon 20 insertions or had a squamous component, they were excluded. Other exclusion criteria included brain metastases and concomitant pathologies or laboratory alteration or concomitant medication use that prevent or contraindicate the use of either study agent.
Participants were randomized 1:1 to receive oral erlotinib at a once-daily dose of 150 mg plus intravenous bevacizumab at 15 mg/kg every 21 days or erlotinib alone at the same dose. Treatment was administered until progressive disease, intolerable toxicity, or patient or physician decision to discontinue.
Stratification factors included performance score (0 to 1 vs 2), type of mutation (exon 19 deletion vs 21 L858R vs others), and center.
The primary end point of the trial was PFS. Other end points included OS, QoL, objective response rate (ORR) per investigator and BICR assessment and in accordance with RECIST v1.1 criteria, and safety.
The study was designed with an expected median PFS in the erlotinib arm was 10 months2 and an expected median PFS of 16.67 months in the combination arm,3 which corresponds with a HR of 0.60, with a 2-sided alpha of 0.05, and a power of 0.80.
Two interim analyses for futility were planned, as well as 1 efficacy interim analysis. The first interim analysis was done in October 2019, with 58% events and futility was not demonstrated. The second interim analysis was anticipated in 2020 but was not done because of the COVID-19 pandemic. The final analysis was done in July 2021 with 140 events reported.
At a median follow-up of 36 months, a total of 160 patients were enrolled, with 80 assigned to erlotinib and 80 assigned to erlotinib/bevacizumab; these patients comprised the intent-to-treat (ITT) population for efficacy. A total of 79 patients on the control arm and 80 patients on the investigative arm started treatment on the trial, and they comprised the ITT population for safety.
The median age of participants was 67.0 years (range, 59.5-73.0), and 63.8% of patients were female. Moreover, 51.9% of patients were never smokers and the remainder were former or current smokers. Most patients (61.9%) had an ECOG performance status of 0, followed by a1 (34.4%), and 2 (3.8%). Regarding EGFR mutation type, 55.0% had an exon 19 deletion, 41.3% had an exon 21 L858R mutation, and 3.8% had other mutation. Ninety-five percent of patients had stage IV disease, and the remaining 5% had stage IIIB disease.
Additional data from the trial demonstrated a numerically higher median OS with the doublet over the monotherapy, at a median 33.3 months (95% CI, 24.3-45.1) vs 22.8 months (95% CI, 18.3-33.0), respectively (adjusted HR, 0.72; 95% CI, 0.47-1.10; P = .132). However, this benefit was not determined to be of statistical significance, according to Piccirillo.
“The benefit of the addition of bevacizumab was confirmed in all subgroups,” Piccirillo said. “Moreover, we found a significant interaction between treatment effect and smoking history, with a higher effect of the addition of bevacizumab in patients who were former or current smokers at the moment of diagnosis.”
The HR for PFS in this subset was 0.49 (95% CI, 0.28-0.82; P = .0323), and the HR for OS was 0.41 (95% CI, 0.21-0.80; P = .0077).
Bevacizumab plus erlotinib elicited an investigator-assessed ORR of 70.0% (95% CI, 60.0%-80.0%) vs 50.0% (95% CI, 39.0%-60.9%) with erlotinib alone (P = .01). Among the responders in the doublet arm, 1.3% had a complete response, 68.8% experienced a partial response, and 22.5% had stable disease. Moreover, 7.5% had disease progression.
Per BICR, the ORR was still higher with the doublet. In the investigative and control arms, the ORRs were 71.3% (95% CI, 61.3%-81.2%) and 53.8% (95% CI, 42.8%-64.7%), respectively (P = .02).
Regarding QoL, 49.3% of those in the erlotinib-alone arm experienced improved best response of Global Health Status vs 60.0% of those in the bevacizumab/erlotinib arm (P = .14). The P value regarding time to deterioration was .325. Although those in the doublet arm experienced slightly better outcomes, there was no significant difference noted in any of the QoL items examined on the trial.
In terms of safety, the most frequently reported adverse effects on the erlotinib/bevacizumab arm included diarrhea (95.0%, grades 0-2; 5.0%, grade 3 or higher), fatigue (6.3%, grade 3 or higher), aspartate aminotransferase (AST) increased (98.8%, grades 0-2; 1.3%, grade 3 or higher), proteinuria (93.8%, grades 0-2; 6.3%, grade 3 or higher), hypertension (76.3%, grades 0-2; 23.8%, grade 3 or higher), thromboembolic event (95.0%, grades 0-2; 5.0%, grade 3 or higher), and rash (66.3%, grades 0-2; 33.8%, grade 3 or higher).
In the erlotinib-alone arm, the most common grade 0-2 effect was fatigue (100.0%), followed by proteinuria and thromboembolic event (both 98.7%), diarrhea (96.2%), AST increased and hypertension (both 94.9%), and rash (83.5%). Grade 3 or higher effects reported in this arm included rash (16.5%), AST increased and hypertension (5.1% each), diarrhea (3.8%), proteinuria and thromboembolic event (1.3% each).
“Patients on the experimental arm experienced more severe hypertension and skin rash, while proteinuria was significantly more frequent with bevacizumab when considering all grades,” Piccirillo noted.
One toxic death occurred on the combination arm, and it was due to intracranial hemorrhage.
At the time of the final analysis, 99.0% of those on the erlotinib arm and 89% of those on the bevacizumab/erlotinib arm had discontinued treatment. Among those who discontinued, 14% and 11% of patients on the control and investigative arms, respectively, did not receive posttreatment anticancer therapy.
The first posttreatment therapy administered in those who had received the monotherapy or the combination, respectively, included osimertinib (Tagrisso; 57% vs 49%), erlotinib (13% vs 16%), gefitinib (Iressa; 0% vs 6%), platinum-based chemotherapy (29% vs 27%), and single-agent chemotherapy (2% in both arms).
“The BICR analysis confirmed the results of the investigator-assessed analysis,” Piccirillo concluded. “A hypothesis-generating subgroup analysis suggests that bevacizumab might work better among smokers…The combination strategy is worth further investigation.”