Bevacizumab plus erlotinib

A number of studies have evaluated bevacizumab plus erlotinib in advanced NSCLC patients with EGFR wild type tumors (Table 1). The first clinical studies with erlotinib plus bevacizumab were conducted by Herbst et al. [17–18], in cisplatin-refractory NSCLC. A phase I/II study evaluated the dose limiting toxicity (DLT) and activity of bevacizumab plus erlotinib in 40 patients with non-squamous NSCLC, in progression after a platinum first-line treatment, regardless of EGFR status [17]. No DLTs were reported: the recommended treatment was bevacizumab 15 mg/kg every three weeks as intravenous infusion and erlotinib 150 mg/day orally. The most frequent adverse events (AEs) were rash (85% of patients), diarrhea (65%) and infection (29%). Overall, 20% of patients had a partial response (PR), while 65% had stable disease. Notably, only nine patients had enough material to proceed for an EGFR mutational analysis and in two cases a mutation was found. Thereafter, in a randomized phase II study [18], 120 patients with non-squamous NSCLC in progression after a platinum treatment were treated with bevacizumab plus erlotinib, or bevacizumab plus docetaxel/pemetrexed or chemotherapy alone. Bevacizumab plus erlotinib was better tolerated than bevacizumab plus chemotherapy: serious AEs were reported in 55%, 41% and 33% of patients in chemotherapy, bevacizumab-chemotherapy and bevacizumab-erlotinib arm, respectively. Fatigue, nausea and vomiting were commonly reported in all 3 arms. Median progression-free survival (PFS) was 4.8, 4.4 and 3 months for bevacizumab plus chemotherapy, bevacizumab plus erlotinib and chemotherapy alone, respectively. Median overall survival (OS) was 12.6, 13.7 and 8.6 months for the bevacizumab plus chemotherapy, bevacizumab plus erlotinib and chemotherapy alone arm, respectively. The risk of progression or death was 0.66 [95% confidence interval (CI): 0.38–1.16] and 0.72 (95% CI: 0.42–1.23) for patients treated with bevacizumab plus chemotherapy or bevacizumab plus erlotinib against chemotherapy alone arm, respectively. No clinically meaningful differences were found between bevacizumab-chemotherapy and bevacizumab-erlotinib arms. Interestingly, one complete response was observed in a patient with an activating EGFR mutation treated with bevacizumab-erlotinib. These promising results prompted a randomized, double-blind, phase III clinical study (BeTa Trial) to compare bevacizumab plus erlotinib with erlotinib in 636 patients with advanced NSCLC in progression after a first-line treatment [19]. Unfortunately, the addition of bevacizumab to erlotinib did not improve the OS that was 9.3 months for bevacizumab-erlotinib versus 9.2 months for erlotinib group, respectively (P = 0.7583). Furthermore, grade 5 AEs were 6% in the bevacizumab group, including two arterial thromboembolic events, and 4% in the erlotinib group. In 30 patients with EGFR-mutant tumors, the hazard ratio (HR) for OS was lower in the bevacizumab than in the control group (HR 0.44, 95% CI: 0.11–1.67). A phase III trial compared second-line erlotinib plus bevacizumab and panitumumab versus erlotinib in 297 metastatic EGFR unselected NSCLC Chinese patients, pretreated with platinum-based chemotherapy [20]. The trial demonstrated an improvement in median PFS (4.6 versus 1.9 months, in experimental and erlotinib arm, respectively, P = 0.003) and in median OS (10.4 versus 8.9 months, in experimental and erlotinib arm, respectively, P = 0.031). However, severe diarrhea and rash were more common in the experimental than in the control arm (24% and 26% versus 8.5% and 8.5%, respectively). The safety and activity of bevacizumab plus erlotinib in pretreated non-squamous NSCLC patients with asymptomatic untreated brain metastases were evaluated in a phase II study (BRAIN trial) [21]. However, only 24 patients were enrolled and results may have only an exploratory value: response rate (RR) was 12.5%, 6-month PFS rate 57.2% and median OS 12 months. AEs were similar to those reported with previous trials of bevacizumab; one patient had grade 1 intracranial hemorrhage. Finally, the ATLAS trial was a randomized phase III study that evaluated the efficacy of erlotinib plus bevacizumab versus bevacizumab plus placebo as maintenance therapy in 745 patients with metastatic non-squamous NSCLC, without progression after four cycles of first-line chemotherapy plus bevacizumab [22]. The bevacizumab-erlotinib combination improved median PFS (4.8 versus 3.7 months, HR 0.71, 95% CI: 0.58–0.86), but not OS (14.4 in bevacizumab-erlotinib versus 13.3 months in the bevacizumab arm, respectively, HR 0.92, 95% CI: 0.7–1.21). Moreover, adverse events, severe AEs (mainly rash and diarrhea), and AEs leading to erlotinib/placebo discontinuation were observed more frequently with bevacizumab-erlotinib combination. On this basis, due to the modest benefit in PFS, the lack of benefit in OS and the worst toxicity profile, erlotinib plus bevacizumab didn’t become a new standard of care as second-line or maintenance therapy.

Several studies have also been conducted with bevacizumab plus erlotinib as first-line therapy in advanced NSCLC patients unselected for EGFR status. A phase II trial in 47 patients naïve for any treatment met the primary end-point that was the non-progression-rate (NPR) at 6 weeks, with a 75% NPR at 6 weeks [23]. However, OS was disappointing (6.9 months) and some patients experienced serious AEs, including a kidney thrombosis, a sigmoid colon perforation and thromboembolic events. The SAKK 19/05 trial evaluated the activity and feasibility of bevacizumab plus erlotinib followed at disease progression by platinum-based chemotherapy in 101 metastatic non-squamous NSCLC patients [24]. This trial did not meet the primary end-point, with a disease stabilization rate at twelve weeks of 54.5% (lower than expected). Moreover, only 63% of patients proceeded to second-line chemotherapy. A disease stabilization rate of 83.3% was observed in EGFR mutated patients (13.8% of cases). The most common AEs with the combination treatment were acne, fatigue and diarrhea. The TASK trial compared erlotinib plus bevacizumab versus chemotherapy plus bevacizumab in naïve patients with metastatic non-squamous NSCLC, unselected for EGFR [25]. At an interim analysis, PFS was significantly worse with bevacizumab plus erlotinib compared with bevacizumab plus chemotherapy arm (median PFS 18.4 versus 25 weeks, respectively; HR 2.05, 95% CI: 1.11–3.77). Moreover, a subgroup analysis showed no benefit with erlotinib treatment also in patients with EGFR positive cancers, probably due to the small sample size. AEs occurred in 84.1% and 82% in the experimental and control group, respectively: the most frequent AEs in the bevacizumab-chemotherapy arm were nausea and neutropenia, while in bevacizumab-erlotinib arm diarrhea and cutaneous rash. A randomized phase II study (INNOVATIONS trial) compared erlotinib plus bevacizumab (EB) against cisplatin, gemcitabine and bevacizumab (PGB) in 224 non-squamous, unselected, metastatic NSCLC patients as first-line therapy [26]. The combination of chemotherapy plus bevacizumab showed superior OS (HR 1.41, P = 0.0409), PFS (HR 1.85, P < 0.0001) and RR (36% versus 12%, P < 0.0001) compared with erlotinib plus bevacizumab. In patients with EGFR activating mutations (20%), the response rate was 25% versus 17% for for EB versus PGB, respectively. No difference in PFS (4.2 versus 4.6 months for EB and PGB, respectively) and only a positive trend for OS (17.0 versus 10.0 months, for EB and PGB, respectively; HR 0.45, 95% CI: 0.18–1.16, P = 0.092) were observed. The most common AEs in EB group were diarrhea and skin rash, while in the control group were thrombocytopenia and thromboembolic events.

Oral VEGF inhibitors and erlotinib

Erlotinib has been combined with oral VEGF inhibitors, including sorafenib and sunitinib in NSCLC patients (Table 2). Sorafenib is a small molecule targeting VEGFR-2 and VEGFR-3, rapidly accelerated fibrosarcoma (Raf), rearranged during transfection (RET), tyrosine-protein kinase KIT, FMS-like tyrosine kinase 3 (FLT-3), and PDGFR-β, with limited anti-tumor activity in NSCLC when used as a single-agent [27–28]. A phase II study compared erlotinib plus sorafenib versus erlotinib plus placebo in 168 pretreated NSCLC patients [29]. In both treatment arms the majority of patients had a non-squamous histology, was female, and already received bevacizumab in a prior line. An EGFR and kirsten rat sarcoma (KRAS) mutation test was done in 65% of the patients and most of them were WT. The primary endpoints [overall response rate (ORR) and PFS] were similar in the two groups. Sixty-seven patients with EGFR wild-type tumors had a median PFS of 3.38 versus 1.77 months (P = 0.018) and a median OS of 8 versus 4.5 months (P = 0.019) with erlotinib plus sorafenib versus erlotinib plus placebo, respectively. These data suggest a potential benefit for the combination in EGFR-WT tumors, more dependent on other signalling pathways, such as PDGFR, VEGFR or Raf inhibited by sorafenib. Sunitinib is a kinase inhibitor targeting VEGFR-1, -2 and -3, PDGFR alpha and beta, stem-cell factor receptor (KIT), FLT3 and RET that has demonstrated antitumor activity as a single agent in phase II studies, in patients with refractory NSCLC, with RR of 2.1–11.1% [30–31]. In a phase III study, erlotinib 150 mg and blinded sunitinib 37.5 mg or placebo were administered orally once daily every 4-week in patients with recurrent NSCLC [32]. Nine hundred-sixty patients were randomized; the majority of them had a diagnosis of lung adenocarcinoma, with unknown EGFR status. OS (primary endpoint) was similar between the two treatment arms (9.0 versus 8.5 months in erlotinib plus sunitib versus erlotinib, respectively). Patients treated in the combination arm had a modest benefit in terms of PFS (3.6 versus 2.0 months, HR 0.807, 95% CI: 0.695–0.937, P = 0.0023), and ORR (10.6% versus 6.9%, P = 0.0471), respectively. Negative results were also reported by a phase II study that compared erlotinib plus sunitinib versus erlotinib plus placebo, in patients with metastatic NSCLC pretreated with one or two prior chemotherapy lines [33]. The combination of sunitinib and erlotinib did not improve PFS (2.8 versus 2.0, HR 0.898, P = 0.321), OS (8.2 versus 7.6 months, HR 1.066, P = 0.617) or objective response rates (4.6% versus 3.0%, respectively). A biomarker analysis showed a better PFS for patients with low-tumor PDGFR-alpha RNA in the combination arm, suggesting the relevance of tumor vascular microenvironment when combining angiogenesis and EGFR inhibitors. In both trials rash, diarrhea, fatigue and nausea were more frequent with the combined treatment.

Angiogenesis plus EGFR inhibitors in “EGFR mutated” patients

Several clinical trials evaluated the activity of the dual inhibition of angiogenesis and EGFR in NSCLC patients with EGFR mutation, including combination of erlotinib plus bevacizumab, gefitinib plus bevacizumab, afatinib plus bevacizumab, erlotinib plus ramucirumab (Table 3).

Bevacizumab and erlotinib

Retrospective analyses in clinical trials conducted with erlotinib plus bevacizumab in unselected NSCLC patients suggested an advantage of the combination compared with erlotinib as a single agent in patients with EGFR activating mutations. On this basis, a number of phase II-III trials prospectively evaluated the efficacy of the combination in this setting of patients. In the first randomized phase II study (JO25567), Seto et al. [34], compared erlotinib plus bevacizumab versus erlotinib in 154 naïve Japanese NSCLC patients with EGFR activating mutations. Erlotinib plus P = 0.0015), but no OS (47 versus 47.4 months, HR 0.81, 95% CI: 0.53–1.23) [35]. A greater benefit from the combination was observed in patients with EGFR exon 19 deletion (median PFS 18 versus 10.3 months, HR 0.41, 95% CI: 0.24–0.72). On the contrary, in patients with EGFR point mutation L858R in exon 21, the benefit seemed to be modest and not clinically significant (HR 0.67, 95% CI: 0.38–1.18). Severe AEs were 91% with erlotinib and bevacizumab and 53% with erlotinib single agent. The most common AEs of any grade were rash, diarrhea, hypertension in the combination group; rash, diarrhea and paronychia in the erlotinib group. These findings were confirmed by the NEJ026 phase III study [36]. In this trial, 112 Japanese patients with EGFR-positive NSCLC were randomized to erlotinib plus bevacizumab and 114 to erlotinib single agent. The results from the preplanned interim analysis showed an improvement in PFS with the combination strategy compared with erlotinib single agent (16.9 compared with 13.3 months, HR 0.605, 95% CI: 0.417–0.877, P = 0.016), with a worse toxicity profile (grade 3–5 AEs in 88% versus 46% of patients, respectively). Moreover, erlotinib plus bevacizumab was associated with a higher frequency of hypertension (46% versus 10%), proteinuria (32% versus 5%), and non pulmonary haemorrhage (26% versus 3%). The BELIEF trial evaluated the activity of first-line erlotinib plus bevacizumab in patients with activating EGFR mutations, according to the expression of T790M resistance mutation [37]. Since the T790M mutation is linked to a high level of signal transducer and activator of transcription 3 (STAT3) protein which up-regulates the VEGF expression, this study was intended to be a proof of concept that adding bevacizumab to erlotinib could improve the treatment of de-novo T790M mutated patients. Among 109 patients treated with erlotinib and bevacizumab, 37 (34%) were T790M positive. The overall median PFS was 13.2 months (95% CI: 10.3–15.5), with a 12 month PFS survival rate of 55% (95% CI: 45–64). The primary end point was met only in T790M-positive patients, with a median PFS of 16 months and a 12 month PFS rate of 68%. Severe AEs were reported in 4.7% of patients included in the safety analysis (including one acute coronary syndrome and a colonic perforation). Inconclusive results were reported by a small phase II trial conducted in U.S. comparing erlotinib plus bevacizumab against erlotinib alone in 88 EGFR mutated patients (75% Caucasian) [38]. The combination failed to result in a significant difference in PFS (17.9 versus 13.5 months, respectively; P = 0.39) and OS (32.4 versus 50.6 months, respectively; P = 0.33). However, these results were affected by the small number of enrolled patients resulting in a low power of the study. The more frequent serious AEs in the combination and erlotinib arms were skin eruption (26% versus 16% of patients), diarrhea (9% versus 13% of patients), hypertension (40% versus 20% of patients), and proteinuria (12% versus 0% of patients).

Bevacizumab and other EGFR TKIs

The Okoyama Lung Cancer Study Group 1001 evaluated the activity of the combination of gefitinib and bevacizumab in metastatic NSCLC patients harbouring the EGFR mutation [39]. Forty-two Japanese patients received gefitinib 250 mg/day plus bevacizumab 15 mg/kg every three weeks. Median PFS was 14.4 months and 1-year PFS rate was 56.7% (95% CI: 39.9–70.5), but the primary end point of the study was not met, because the lower limit of the CI reached the 40% threshold. Cutaneous rash, hypertension, proteinuria and diarrhea were the AEs more frequently reported. Negative results were also reported in another Japanese phase II trial that evaluated the activity of gefitinib plus bevacizumab in EGFR mutated patients with metastatic NSCLC cancer, without brain metastases and T790M expression, although the conclusions of this study were invalidated by the very small sample size (16 patients) [40]. A phase II trial (the ABC study) evaluated afatinib plus bevacizumab in metastatic EGFR mutated NSCLC, pretreated with erlotinib or gefitinib [41]. Thirty-two patients were enrolled in the trial; in 14 of them (44%) the T790M expression was detected. Six patients (18.8%, 95% CI: 7.2–36.4%) obtained an objective response. In patients T790M positive the ORR and median PFS were 14.3% and 6.3 months, while in T790M negative they were 22.2% and 7.1 months, respectively. AEs frequently reported were skin rash and diarrhea. Therefore, the combination seemed to be promising in T790M negative patients who were found to be resistant to other EGFR treatments.

Ramucirumab and erlotinib

The Relay study compared ramucirumab plus erlotinib against erlotinib plus placebo in EGFR mutated metastatic NSCLC, without brain metastases [42]. The trial enrolled 449 patients with common EGFR mutations (patients with T790M mutation at baseline were excluded): 224 patients were randomized to erlotinib 150 mg/day plus ramucirumab 10 mg/kg every two weeks, while the remaining 225 patients were randomized to erlotinib plus placebo. The addition of ramucirumab to erlotinib improved PFS (19.4 versus 12.4 months, HR: 0.59, 95% CI: 0.46–0.76, P < 0.0001). OS data were immature. No difference in the percentage of patients who developed T790M mutation between the two groups was reported. One patient died during the treatment due to an event potentially related to the ramucirumab-erlotinib treatment. In the combination group, a serious AE was reported in 29% of patients, including hypertension (24%) and cutaneous rash (15%). In the control group, the 21% of patients had a serious AE, including hypertansaminasemia (8%) and cutaneous rash (9%). Recently, the combination of erlotinib and ramucirumab has been approved by FDA for the treatment of EGFR mutated naïve patients.