Bevacizumab in the Treatment of Non-Small-Cell Lung Cancer

Home , Targeted therapy, Treatment of lung cancer

TE Stinchcombe and MA Socinski

Multidisciplinary Thoracic Oncology Program, Lineberger Comprehensive Cancer Center, University of North Carolina Chapel Hill, Chapel Hill, NC, USA

Lung cancer is the leading cause of cancer death in the et al., 2006). Approximately 85% of patients will have United States. The majority of patients present with the non-small-cell histology, and approximately 55% of advanced disease, and treatment with standard cytotoxic patients will present with locally advanced or metastatic chemotherapy improves survival and quality of life in disease (Bulzebruck et al., 1992; Govindan et al., 2006). patients with a preserved functional status. However, the The current standard treatment for patients with prognosis is poor with the majority of patients dying in preserved functional status is double-agent cytotoxic less than a year. Treatment with standard cytotoxic chemotherapy (Pfister et al., 2004). The median survival chemotherapy has reached a therapeutic plateau, and new with chemotherapy treatment is 8–10 months, and the therapeutic approaches have investigated therapies that 1-year survival rate is 20–30% (Schiller et al., 2002). The target the specific molecular pathways involved in addition of a third cytotoxic agent does not appear to carcinogenesis and angiogenesis. The most promising improve survival, and a therapeutic plateau has been strategy for inhibiting angiogenesis involves agents that reached with cytotoxic therapy (Delbaldo et al., 2004). either target the proangiogenesis growth factor, vascular Because of the relatively poor prognosis of patients endothelial growth factor A (VEGF) by preventing with advanced-stage disease, there has been great binding to the receptor or inhibiting the downstream interest in developing new therapeutic strategies, espe- signaling of the vascular endothelial growth factor cially the development of antiangiogenesis therapy. receptor. The only therapeutic agent approved for the Angiogenesis, the process where tumors develop new treatment of lung cancer is bevacizumab, a monoclonal blood supply, is essential for solid tumor growth beyond antibody that binds to VEGF. A recent phase III trial 2–3 mm3 (Folkman, 1992). Complex interactions invol- revealed a statistically significant improvement in re- ving numerous inhibitory or stimulatory factors con- sponse rate, progression free and overall survival with tribute to the careful regulation of angiogenesis. When combination of bevacizumab with chemotherapy over the stimulatory factors exceed the inhibitory factors an chemotherapy alone. Attempts to identify surrogate ‘angiogenic switch’ occurs, and a cascade of events is markers of antiangiogenesis activity are currently on- initiated, including the release of proangiogenic stimu- going, and may assist in the selection of patients for latory factors, endothelial cell activation, endothelial cell antiangiogenesis therapy and the development of this class migration, proliferation and capillary tube formation of agents. (Herbst et al., 2005b). Solid tumors develop the ability Oncogene (2007) 26, 3691–3698. doi:10.1038/sj.onc.1210366 to secrete proangiogenic factors which lead to continued angiogenesis and metastases (Folkman, 1990; Herbst Keywords: monoclonal antibody; targeted therapy; et al., 2005b). angiogenesis The vascular endothelial growth factors, VEGF-A (also known as VEGF), VEGF-B and VEGF-C, are proangiogenic factors that play a critical role in the development of non-malignant and malignant angiogenesis, and the proliferation and maintenance of Introduction vasculature (Robinson and Stringer, 2001). VEGF is the most active in the development of tumor angio- In the United States, lung cancer is the leading cause of genesis and metastasis. VEGF binds to VEGFR-1 (also cancer deaths among both men and women, and more known as Flt-1), which mediates angiogenesis, and to patients will die of lung cancer than breast, colon and VEGFR-2 (also known as KDR), which mediates prostate combined (Jemal et al., 2006). An estimated lymphangiogenesis and angiogenesis (Ferrara et al., 174 000 new diagnoses and 162 000 deaths are expected 2003). VEGF has been associated with the development from lung cancer in the United States in 2006 (Jemal hyperpermeable, immature vasculature, which is typical of malignancy-related angiogenesis (Dvorak et al., 1995; Correspondence: Dr MA Socinski, Multidisciplinary Thoracic Oncol- Bates et al., 2001; Ferrara et al., 2003). VEGFR ogy Program, Lineberger Comprehensive Cancer Center, University of North Carolina Chapel Hill, 3009 Old Clinic Building CB 7305, receptors are expressed on lung cancer cells and mediate Chapel Hill, NC 27599-7305, USA. antiapoptotic and metastatic signals; therefore, antian- E-mail: [email protected] giogenic therapy may have direct tumor effects as well Bevacizumab in the treatment of NSCLC TE Stinchcombe and MA Socinski 3692 (Dy and Adjei, 2006). VEGF may play a significant role treatment arms: carboplatin (area under the curve in the development of inflammation and other processes (AUC) ¼ 6 mg/ml/min) using the Calvert equation frequently associated with malignancy. VEGF is also (Calvert et al., 1989) and paclitaxel (200 mg/m2) every known as vascular permeability factor, and can induce 3 weeks alone or with low or high bevacizumab (7.5 or vascular leakage and may be involved brain edema 15 mg/kg every 3 weeks, respectively) (Johnson et al., (Ferrara et al., 2003). Malignant pleural effusions from 2004). After completing the planned chemotherapy on lung cancer have high levels of VEGF, and VEGF the bevacizumab containing arms, non-progressing secretion has also been associated with the development patients could continue the bevacizumab at the same of ascites (Senger et al., 1983; Yanagawa et al., 1999; dose and schedule for up to a maximum of 18 doses. Yano et al., 2000a). The direct role VEGF has in Patients on all three arms received a median of six cycles angiogenesis and other malignant processes made it an of carboplatin and paclitaxel, and patients on the low- attractive target for the development of new therapeutic dose and high-dose bevacizumab arm received a median agents. of eight and 10 doses of bevacizumab, respectively. The role of angiogenesis in the progression of lung Treatment with bevacizumab (15 mg/kg) resulted in a cancer is well established, with multiple studies revealing higher response rate, a longer median time to progres- that high microvessel density is associated with the sion and a trend towards increased survival (Table 1). development of metastases and poor survival (Fontanini The difference in median time to tumor progression et al., 1995; Lucchi et al., 1997; D’Amico et al., 1999; was statistically significant between the high-dose Yano et al., 2000b). High vascularity has also been bevacizumab arm and control arm (7.4 versus 4.2 associated with tumor progression (Ushijima et al., months, respectively; P ¼ 0.023), but no difference was 2001). The preclinical data on the role of angiogenesis in seen between the low-dose bevacizumab and control the behavior of lung cancer supported the development treatment arm (4.3 versus 4.2 months, respectively; of clinical trials of antiangiogenesis agents. There are P ¼ nonsignificant). currently multiple different agents, with antiangiogenic The combination of bevacizumab with chemotherapy activity being investigated in non-small-cell lung cancer was notable for several bevacizumab-associated adverse (NSCLC), however, the only agent currently approved events, including hypertension, proteinuria and bleed- by the United States Food and Drug Administration is ing. Two distinct clinical patterns of bleeding were bevacizumab. Bevacizumab is a recombinant humanized observed: minor mucocutaneous hemorrhage, predomi- monoclonal antibody that binds to the circulating nantly epistaxis and major hemoptysis. Six patients VEGF-A ligand. experienced a major life-threatening bleeding described Bevacizumab as a single agent has limited activity, as hemoptysis or hematemesis, and four events were and in the clinical phase I trial in patients with solid fatal. Four of the 13 patients (30%) with squamous tumors, and no responses were seen in 23 patients histology treated with bevacizumab experienced severe evaluable for response; however, 12 patients had stable hemorrhage versus two of 54 (4%) patients with non- disease over the duration of the study (Gordon et al., squamous histology. Five of the six cases had cavitation 2001). Thus, bevacizumab as a single agent appears to at baseline or tumor necrosis at baseline or during be cytostatic; however, treatment with bevacizumab is therapy. Pulmonary hemorrhage may have been asso- synergistic with chemotherapy. In a pivotal phase III ciated with central location. However, many squamous trial in patients with metastatic colorectal cancer, cell carcinomas are centrally located; thus location may patients treated with chemotherapy in combination with have been a surrogate marker for histology. The bevacizumab had superior survival to patients treated episodes of pulmonary hemorrhage did not appear to with the same chemotherapy (Hurwitz et al., 2004). The be dose dependent because five of the six patients were synergistic effects of bevacizumab with chemotherapy have been attributed to a reduction in vascular perme- Table 1 Efficacy results of a randomized phase II trial of ability with normalization of the vasculature resulting in bevacizumab in combination with carboplatin and paclitaxel (Johnson a reduction of interstitial pressure within the tumor et al., 2004) (Jain, 2001). These changes are believed to enhance the Carboplatin/Paclitaxel Bevacizumab delivery of cytotoxic therapy to the malignant cells (Jain, 2001). 7.5 mg/kg 15 mg/kg All patients n ¼ 32 n ¼ 32 n ¼ 34

Bevacizumab in the treatment of lung cancer Response (%) 18.8 28.1 31.5 Median TTP (months) 4.2 4.3 7.4 A phase I trial in patients with solid tumors indicated Survival (months) 14.9 11.6 17.7 that bevacizumab could be safely combined with full- Patients with non-squamous histology dose carboplatin and paclitaxel, and the efﬁcacy and n ¼ 25 n ¼ 22 N ¼ 32 safety of bevacizumab was investigated in a randomized phase II trial in patients with advanced or recurrent Response (%) 20 31.8 50 NSCLC (Margolin et al., 2001; Johnson et al., 2004). Median TTP (months) 4.0 6.3 7.1 Survival (months) 12.2 14.0 17.8 Ninety-nine patients were randomized to one of three

Oncogene Bevacizumab in the treatment of NSCLC TE Stinchcombe and MA Socinski 3693 on the low-dose arm, or time dependent because the Table 2 ECOG 4599 trial efficacy results (Sandler, 2006a) X events occurred early (p60 days) and late ( 180 days). Parameter CP CPB HR (95% CI) P-value If a subset analysis is performed excluding the squamous histology patients, it suggests that bevacizumab therapy Response (%) 15% 35% Not reported o0.001 improves response, time to tumor progression and PFS (median) 4.5 months 6.2 months 0.66 (0.57–0.77) o0.001 overall survival (Table 1). Overall survival 10.3 12.3 months 0.79 (0.67–0.92) 0.003 Based on the results of this phase II trial, the Eastern Abbreviations: CP, carboplatin and paclitaxel; CPB, carboplatin, Cooperative Oncology Group (ECOG) initiated a phase paclitaxel and bevacizumab; PFS, progression-free survival. II/III trial, ECOG 4599, comparing carboplatin, paclitaxel versus carboplatin, paclitaxel, bevacizumab (15 mg/kg) in chemotherapy-naı¨ ve patients with non- Table 3 ECOG 4599 trial efficacy results and gender (Brahmer et al., squamous histology and advanced-stage IIIB/IV disease 2006) (Sandler, 2006). Patients with a history of gross Males Females hemoptysis (defined as >1/2 a teaspoon of bright red blood), central nervous system metastases or history of Parameter CP CPB P-value CP CPB P-value thrombotic disorder or hemorrhagic disorders were Number of patients 230 191 162 190 excluded. Patients were required to have a computed Response (%) 16 29 0.001 14 41 o0.0001 tomography (CT) scan of the head within 4 weeks of PFS (months) 4.3 6.3 o0.0001 5.3 6.2 0.002 enrollment to evaluate for brain metastases. Patients Survival (months) 8.7 11.7 0.001 13.1 13.3 0.87 with central tumors were not excluded. Abbreviations: CP, carboplatin/paclitaxel; CPB, carboplatin, paclitaxel, The bevacizumab-containing treatment arm in com- bevacizumab; PFS, progression free survival. parison to the standard chemotherapy arm had a statistically significant higher rate of National Cancer Institute (NCI) Common Terminology Criteria for patients should not be denied therapy with bevacizu- Adverse Events (CTCAE) grade 4 neutropenia (25.5 mab. A retrospective analysis of the episodes of versus 16.8%, respectively; P ¼ 0.002), thrombocytope- pulmonary hemorrhage occurring o150 days from nia (1.6 versus 0.2%, respectively; P ¼ 0.04), and febrile initiation of treatment on the phase II and III trials neutropenia (5.2 versus 2%, respectively; P ¼ 0.02) was performed. This analysis used a match case-control (Sandler et al., 2006a). There was a higher rate of method to identify potential radiographic and clinical non-hematologic toxicity on the bevacizuamab arm risk factors for severe pulmonary hemorrhage (Sandler in comparison to the standard chemotherapy arm et al., 2006b). This analysis identified hemoptysis at as well including grade 3–5 hypertension (7.7 versus baseline, and tumor cavitation as probable risk factors 0.7%, respectively; Po0.001), hemorrhage (4.4 versus for pulmonary hemorrhage; however, central location 0.7%, respectively; Po0.001) and proteinuria (3.1 was not a risk factor. The number of episodes of versus 0%, respectively; Po0.001) (Sandler, 2006). hemoptysis included in this analysis was small (n ¼ 6), There were seven hemorrhage-related deaths on the and a larger analysis which includes patients from bevacizumab arm versus one on the carboplatin and ongoing clinical trials will be needed to better determine paclitaxel arm. The hemorrhage-related deaths on the the pretreatment and treatment-related risk factors and bevacizumab-related arm consisted of hemoptysis the relative risk associated with each factor. (n ¼ 5) and gastrointestinal bleeding (n ¼ 2). There were In addition to these completed clinical trials, an 15 treatment-related deaths on the bevacizumab-treat- ongoing phase III trial in advanced NSCLC, the ment arm versus two on the standard treatment arm; the AVAiL, is investigating bevacizumab in combination difference was significant (P 0.001). ¼ with cisplatin and gemcitabine. This three-arm trial will There was a statistically significant improvement in enroll 1050 patients: the control arm is cisplatin and the response, progression free survival, and overall gemcitabine, and the two experimental arms are survival in the bevacizumab-containing arm (Table 2). cisplatin and gemcitabine in combination with bevaci- An exploratory subgroup analysis was performed zumab 7.5 mg/kg, and 15 mg/kg every 3 weeks. The investigating the efficacy of bevacizumab with regard primary end point will be progression-free survival, and to age, stage, prior weight loss, race, performance status and gender. Survival was seen across all groups except this trial will provide valuable safety and efficacy for gender (Sandler, 2006). No survival benefit was seen information about bevacizumab at two different doses, among women despite a fourfold increase in response and in combination with a different chemotherapy and statistically significant difference in progression-free regimen. The safety and efficacy data from this trial survival (Table 3) (Brahmer et al., 2006). The test for should be available in the next year. treatment sex interaction in a proportional hazards model for survival was statistically significant (P ¼ 0.04). The reasons for this difference are unclear, and there Second-line trials were no survival differences seen in the previous trial of bevacizumab in metastatic colorectal cancer (Hurwitz Two phase II trials have investigated the role of et al., 2004). Until the potential relationship between bevacizumab in patients with non-squamous histology gender and overall survival can be better clarified female who have progressed after initial chemotherapy

Oncogene Bevacizumab in the treatment of NSCLC TE Stinchcombe and MA Socinski 3694 treatment. Neither of these trials included patients who design, it cannot be definitively determined if treatment had received bevacizumab therapy in the first-line with a bevacizumab-containing treatment results in therapy. A phase I/II trial investigated the toxicity and superior survival to standard therapy. efficacy of bevacizumab in combination with erlotinib, an oral epidermal growth factor (EGFR) tyrosine kinase inhibitor (TKI) in patients who had progressed after Patient populations excluded from the ECOG 4599 trial initial chemotherapy for NSCLC (Herbst et al., 2005a). Forty patients were enrolled in the trial, and 22 of the Owing to the inclusion and exclusion criteria used, a patients (55%) had received Xtwo prior chemotherapy significant percentage of the advanced NSCLC popula- regimens, and nine patients were never smokers (22.5%), which has been associated with higher response tion was excluded from the ECOG 4599 trial. On a to EGFR TKI therapy (Pao et al., 2004). No dose- recent North American trial of advanced NSCLC, 17% limiting toxicities were observed in the phase I portion, of patients presented with brain metastases, and there and the phase II dose was established as erlotinib 150 mg is evidence that the prevalence of brain metastases is daily and bevacizumab 15 mg/kg every 3 weeks. The increasing on cooperative group trials (Treat et al., 2005; most common toxicities observed were rash, diarrhea, Wakelee et al., 2006a). Previously, brain imaging was infection, hematuria and proteinuria. No treatment- only performed in patients with advanced disease if related deaths were reported, and the only grade 3 or 4 there was a clinical suspicion of brain metastases (Pfister toxicities seen were rash (6%) and infection (6%). et al., 2004). Since the presence of brain metastases will This combination demonstrated significant activity. determine if a patient is a candidate for bevacizumab Eight patients (20%) had a partial response (PR), and therapy, brain imaging will probably be performed more 26 patients (65%) had stable disease (SD) as their best frequently, and the prevalence of brain metastases may response. The median overall survival for the 34 patients increase. Magnetic resonance imagining (MRI) is more treated at the phase II dose was 12.6 months, and the sensitive than CT scans, and increase use of MRI may progression-free survival was 6.2 months. Two (22%) of lead to an increase in the prevalence of brain metastases the nine patients who had never smoked had a PR. as well (Schellinger et al., 1999). EGFR mutations have been associated with response to The exclusion of patients with brain metastases was EGFR TKI therapy (Lynch et al., 2004; Paez et al., based on a single episode of central nervous hemorrhage 2004; Pao et al., 2004), and nine patients had sufficient related to brain metastases on the phase I trial (Gordon paraffin-embedded tissue for EGFR mutational analy- et al., 2001). Given the prevalence of brain metastases sis. These nine patients included three patients with a PR and the limited safety data of bevacizumab, there is and three patients with SD, and three with progressive significant interest in investigating the safety of beva- disease (PD). Mutations were detected in one of the cizumab in this patient population. A phase II trial of three patients with a PR, and one of the three with SD. bevacizumab in combination with first- or second-line A second randomized phase II trial was performed in therapy in patients with treated brain metastases from the second-line setting to investigate the safety and non-squamous NSCLC (known as the PASSPORT efficacy of bevacizumab with chemotherapy or erlotinib trial) has been initiated. The interval between comple- versus standard chemotherapy (Fehrenbacher et al., tion of radiation therapy for the brain metastases and 2006). Patients were required to have an ECOG initiation of the bevacizumab therapy cannot be less performance status of 0–2 and non-squamous histology. than 4 weeks. The systemic therapy used will be at the Patients with central lesions, cavitation or lesion discretion of the investigator. The primary end point for proximal to a major blood vessel were excluded. Other this study is the incidence of NCI grade X2 central inclusion and exclusion criteria were similar to the nervous system hemorrhage, and secondary objectives criteria used for ECOG 4599. One hundred and twenty will be overall survival and toxicity. This trial will enroll patients were randomized to one of three treatment approximately 100 patients at 30 centers. arms: chemotherapy (docetaxel or pemetrexed) alone, There is also interest in determining, if patients with bevacizumab and chemotherapy (docetaxel or peme- squamous histology can be risk stratified for the trexed) or bevacizumab and erlotinib. There was a trend development of hemoptysis based on clinical or radio- towards improved progression-free survival with the graphic factors so that not all patients with squamous addition of bevacizumab and chemotherapy (HR ¼ 0.66; histology will be excluded. The prevalence of the 95% CI 0.38–1.16) and bevacizumab and erlotinib squamous histology on recent ECOG cooperative group (HR ¼ 0.72; 95% CI 0.42–1.23) in comparison to trials is estimated to be approximately 20% of patients chemotherapy alone, however, the progression-free (Wakelee et al., 2006a). The current hypothesis is that survival and overall survival did not reach statistical bevacizumab therapy induces central necrosis or causes significance. The preliminary safety data indicate a rate enlargement of a pre-existing cavitation. This event in of grade 3–5 hemorrhage 5.1%, and the overall rate of combination with poorly developed and immature neutropenia was similar between the chemotherapy vessels of the tumor may result in hemorrhage into the treatment arms. This randomized phase II trial provides tumor cavity (Johnson et al., 2004). Thoracic radiation valuable safety data about bevacizumab in the second- therapy (TRT), a standard therapy for hemoptysis, may line setting, and the trend towards improved progres- reduce the risk of patients developing serious pulmonary sion-free survival is intriguing. However, due to the trial hemorrhage. A clinical trial to investigate treatment

Oncogene Bevacizumab in the treatment of NSCLC TE Stinchcombe and MA Socinski 3695 with bevacizumab in patients with squamous histology setting. The activity of bevacizumab in combination after TRT is in development. Another phase II trial, the with standard second-line therapies after previous BRIDGE trial, will treat patients with squamous cell treatment with bevacizumab is currently unknown. histology with two cycles of carboplatin and paclitaxel ECOG has proposed a randomized phase II trial that for two cycles, and then initiate treatment with will randomize patients upon disease progression to carboplatin, paclitaxel and bevacizumab for an addi- second-line chemotherapy alone or second-line che- tional four cycles. Bevacizumab will be continued until motherapy and bevacizumab. Patients will be required disease progression, unacceptable toxicity or a maxi- to have received ﬁrst-line chemotherapy with four to six mum of 12 months of therapy. The hypothesis being cycles of platinum-based therapy and bevacizumab until tested is that tumor cavitation from rapid response to disease progression. bevacizumab in combination with chemotherapy may Based on the promising results of phase II trials with have been a contributing factor for the development of the combination of bevacizumab and erlotinib, the pulmonary hemorrhage, and that prior treatment with integration of the combination into ﬁrst-line therapy is chemotherapy or TRT will result in a more gradual being investigated in a randomized phase III trial, the response and patients will be at a lower risk of ATLAS trial. All patients will receive four cycles of pulmonary hemorrhage. Until the safety data from platinum-based therapy, carboplatin in combination these clinical trials are available, patients with squamous with paclitaxel, docetaxel or gemcitabine, in combina- histology should not receive bevacizumab therapy. One tion with bevacizumab, and patients will be randomized clinical scenario where patients with squamous histology to therapy with bevacizumab alone versus bevacizumab may have acceptable risk with bevacizumab therapy is if and erlotinib until disease progression. Patients, who the patient has had surgical resection of all their progress on the bevacizumab alone arm, will receive intrathoracic disease, and upon relapse the metastatic erlotinib. The primary end point will be progression-free disease is entirely extrathoracic. survival, and the secondary end points will be overall survival, response rate and safety.

Duration of therapy and integration into current treatment paradigms Adjuvant therapy and therapy for unresectable stage III disease The optimal duration of bevacizumab therapy is another clinical question as patients continued bevaci- Given the improvement in survival seen in advanced- zumab until disease progression on the ECOG 4599 stage disease, there is interest in integrating bevacizu- trial. Treatment with single-agent bevacizumab may mab into the treatment of early or locally advanced cause regression of vasculature, and the inhibition of stage disease, and potentially improving the overall neo-vascularization which theoretically would have survival. Recent trials have revealed a survival benefit to cytostatic effects on tumor growth and increase the cisplatin-based adjuvant therapy, and there is significant progression-free survival. Clinical evidence of activity of interest in improving the efficacy of adjuvant therapy single-agent bevacizumab in NSCLC is limited, how- (Arriagada et al., 2004; Douillard et al., 2005; Winton ever, some data are available from the randomized et al., 2005). A United States Intergroup trial, ECOG phase II trial (Johnson et al., 2004). Patients who were 1505, has been developed that will compare cisplatin- treated with carboplatin and paclitaxel could receive based therapy versus cisplatin-based therapy in combi- single-agent bevacizumab upon disease progression. nation with bevacizumab in patients with resected stage Nineteen patients received second-line therapy with IB (with tumors X4 cm), stage II and stage III disease single agent bevacizumab, and five patients had stable (Wakelee et al., 2006b). The chemotherapy must be disease. The median survival in this subgroup of patients cisplatin-based, however, the second agent will be was 10 months, and the 1-year survival rate was 47% docetaxel, gemcitabine or vinorelbine at the discretion from the time of crossing over. Some of the concerns of the treating physician, and patients on the bevacizu- about continued therapy with bevacizumab are the mab treatment arm will receive bevacizumab for 1 year. potential development of resistance to bevacizumab, The target accrual for this trial will be 1500 patients. and the longer duration of treatment potentially A multi center phase I/II trial investigating the role increases the risk for developing treatment-related of bevacizumab in the treatment of patients with unre- complications. It is also possible that the majority of sectable stage IIIA/B disease who are appropriate the benefit of bevacizumab therapy may be obtained in a candidates for combined modality therapy is being defined duration of therapy, and continuous therapy is coordinated by the University of North Carolina. not necessary. The inconvenience, the expense and the Patients with squamous histology are eligible, but fact that the single agent response rate of bevacizumab is patients with squamous histology that abut or invade low are also factors that need to be considered. major blood vessels will be excluded. Other inclusion Another vexing clinical question is whether patients and exclusion criteria are similar to ECOG 4599. who progressed on single-agent bevacizumab therapy Patients will receive induction therapy with two cycles should be considered ‘resistant’ to bevacizumab, and of carboplatin, paclitaxel and bevacizumab, and then should not receive bevacizumab in the second-line will receive weekly carboplatin, paclitaxel and bi-weekly

Oncogene Bevacizumab in the treatment of NSCLC TE Stinchcombe and MA Socinski 3696 bevacizumab with concurrent TRT to 74 Gy. If the or radiological parameters of antiangiogenesis activity initial cohorts of patients have a tolerable toxicity (Davis et al., 2003). Plasma VEGF levels are one profile, intermittent erlotinib with the TRT will be potential surrogate marker. A correlative science com- integrated in the treatment. The schedule of erlotinib is panion study of ECOG 4599 investigated elevated Tuesday through Friday during TRT, and the initial VEGF levels, basic fibroblast growth factor (bFGF), cohort will receive 50 mg daily. The final planned cohort and soluble intracellular adhesion molecule-1 (ICAM) on the trial will receive bevacizumab 10 mg/kg every 2 in NSCLC, and if there were changes in the levels in weeks and erlotinib 150 mg. After completion of the response to endothelial apoptosis (Dowlati et al., 2006). TRT, patients will receive consolidation therapy with One hundred and sixty-six patients had pretreatment bevacizumab 15 mg/kg every 3 weeks for six cycles and levels, and 112 patients had posttreatment levels. Only erlotinib 150 mg daily. The erlotinib will be discontinued baseline ICAM level was significantly associated with at the end of the consolidation therapy. This trial will response. Patients with low ICAM in both arms had a provide valuable information about the toxicity of higher response rate than patients with a high ICAM (29 bevacizumab and erlotinib in combination with TRT, versus 13%, respectively; P ¼ 0.03), and better overall additional toxicity data about patients with squamous survival (P ¼ 0.00005), and 1 year survival (65 versus histology, who are selected based on clinical and 25%, respectively). No other factor predicted survival. radiographic criteria, and the response rate of carbo- Tests for treatment interaction were significant for platin, paclitaxel and bevacizumab in stage III disease. progression-free survival in the patients with low ICAM The Southwest Oncology Group (SWOG) is also levels. This suggests baseline ICAM levels are prognos- performing a phase I/II trial in patients with unresect- tic for survival, and that improvements in progression- able stage IIIA/B disease. All patients will receive free survival from the addition of bevacizumab occur cisplatin, etoposide and with TRT and docetaxel and mainly in patients with low ICAM levels. bevacizumab consolidation therapy. Patients will receive The use of non-invasive radiological imaging has been growth stimulating factor (G-CSF) support during investigated in a number of phase I trials of antiangio- the consolidation therapy. Two cohorts of patients genesis agents. A phase I trial of recombinant human will receive bevacizumab, on different schedules, with endostatin (rh-Endo) investigated tumor blood flow and cisplatin, and etoposide during the TRT. Patients with metabolism as measured on positron emission tomo- squamous histology and tumor cavitation or located graphy (PET) (Herbst et al., 2002). This trial found that within 1 cm of a major blood vessel will be excluded. The tumor blood flow generally decreased with increasing other inclusion and exclusion criteria are similar to the doses of rh-Endo, however, the effects were complex ECOG 4599 trial. The primary end point will be the rate and nonlinear. On a phase I trial with tetrathiomolyb- of grade 4 or 5 hemorrhage, and the secondary end date, an anticopper, antiangiogenic agent, a patient with points will be progression-free survival, overall survival rib metastases from renal cell carcinoma was evaluated and response rate. with three-dimensional ultrasound which revealed a 4.4- fold decrease in tumor blood flow (Brewer et al., 2000). A dynamic CT scan confirmed the decrease in blood Predictors of response to Bevacizumab flow (Brewer et al., 2000). Another non-invasive method of assessing changes in tumor blood flow is dynamic The selection of the optimal dose of antiangiogenesis contrast-enhanced magnetic resonance imaging (DCE- therapy and the predictors of response to therapy are MRI). A significant correlation between changes in areas of active investigation. In preclinical models, tumor vasculature and clinical outcome using this antiangiogenesis therapies generally result in disease technology was detected on a phase I clinical trial of stabilization or delay in disease progression (Ferrara PTK787 (Morgan et al., 2003). At this point, the et al., 2003). The fact that these therapies are cytostatic preliminary data provide valuable information about means that a dose–response curve is unlikely to provide the potential applications of these new technologies. assistance in determining the optimal dose. Further However, these radiological markers will have to be complicating the issue is that the traditional paradigm of correlated with clinical outcomes and validated in larger establishing the maximum tolerated dose based on dose– trials before they can be used in clinical decision making. limiting toxicities may not be applicable as many of these agents have limited acute toxicities (Adjei, 2006). The most direct and accurate method of determining the Conclusion optimal dose of antiangiogenetic therapy would be serial biopsies, however, given the risks associated with repeat The ECOG 4599 trial demonstrated that the addition of biopsies, this approach is unlikely to be employed in bevacizumab to standard chemotherapy, carboplatin lung cancer. The one clinical situation where this may be and paclitaxel resulted in a statistically significant and feasible is on a clinical trial investigating the use of clinically relevant improvement in response rate, pro- preoperative antiangiogenesis therapy in early-stage gression-free survival and overall survival. This pivotal disease. trial provided confirmation of the concept that inhibi- This has led to the interest of in developing surrogate tion of angiogenesis would be an effective target for pharmacokinetic, pharmacodynamic, correlative science inhibiting disease progression in lung cancer. The

Oncogene Bevacizumab in the treatment of NSCLC TE Stinchcombe and MA Socinski 3697 carefully selected inclusion and exclusion criteria limited bevacizumab in patient populations that were excluded enrollment to patients who were at a lower risk for from the ECOG 4599 trial. There is also significant bevacizumab-related bleeding complications. However, interest in integrating bevacizumab therapy into earlier even in this select patient population, there continued to stage disease and second-line therapy. Bevacizumab is be a statistically significant increased risk of grade 3–5 the first of several new antiangiogenic agents that will hemorrhage. Ongoing trials are evaluating the safety of have activity in NSCLC.

References

Adjei AA. (2006). What is the right dose? The elusive factor, microvascular hyperpermeability, and angiogenesis. optimal biologic dose in phase I clinical trials. J Clin Oncol Am J Pathol 146: 1029–1039. 24: 4054–4055. Dy GK, Adjei AA. (2006). Angiogenesis inhibitors in lung Arriagada R, Bergman B, Dunant A, Le Chevalier T, Pignon cancer: a promise fulfilled. Clin Lung Cancer 7(Suppl 4): JP, Vansteenkiste J. (2004). Cisplatin-based adjuvant S145–S149. chemotherapy in patients with completely resected non- Fehrenbacher L, O’Neill V, Belani C, Bonomi P, Hart L, small-cell lung cancer. N Engl J Med 350: 351–360. Melnyk O et al. (2006). A phase II, multicenter, randomized Bates DO, Heald RI, Curry FE, Williams B. (2001). Vascular clinical trial to evaluate the efficacy and safety of endothelial growth factor increases Rana vascular perme- bevacizumab in combination with either chemotherapy ability and compliance by different signalling pathways. (docetaxel or pemetrexed) or erlotinib hydrochloride com- J Physiol 533: 263–272. pared with chemotherapy alone for treatment of recurrent Brahmer JR, Gray R, Schiller JH, Perry M, Sandler A, or refractory non-small cell lung cancer. J Clin Oncol (2006 Johnson DH. (2006). ECOG 4599 phase III Trial of ASCO Annual Meeting Proceedings Part I) 24, Abstract Paclitaxel plus Carboplatin +/À Bevacizumab: subset 7062, 379S. analysis of survival by gender. J Clin Oncol (2006 ASCO Ferrara N, Gerber HP, LeCouter J. (2003). The biology of Annual Meetings Proceedings) 14: abstract 7036, 373S. VEGF and its receptors. Nat Med 9: 669–676. Brewer GJ, Dick RD, Grover DK, LeClaire V, Tseng M, Folkman J. (1990). Endothelial cells and angiogenic growth Wicha M et al. (2000). Treatment of metastatic cancer with factors in cancer growth and metastasis. Introduction. tetrathiomolybdate, an anticopper, antiangiogenic agent: Cancer Metastasis Rev 9: 171–174. phase I study. Clin Cancer Res 6: 1–10. Folkman J. (1992). The role of angiogenesis in tumor growth. Bulzebruck H, Bopp R, Drings P, Bauer E, Krysa S, Probst G Semin Cancer Biol 3: 65–71. et al. (1992). New aspects in the staging of lung cancer. Fontanini G, Bigini D, Vignati S, Basolo F, Mussi A, Lucchi M Prospective validation of the International Union Against et al. (1995). Microvessel count predicts metastatic disease Cancer TNM classification. Cancer 70: 1102–1110. and survival in non-small cell lung cancer. J Pathol 177: Calvert A, Newell D, Gumbrell L, O’Reilly S, Burnell M, 57–63. Boxall F. (1989). Carboplatin dosage: prospective evalua- Gordon MS, Margolin K, Talpaz M, Sledge Jr GW, Holmgren E, tion of a simple formula based on renal function. J Clin Benjamin R et al. (2001). Phase I safety and pharmacokinetic Oncol 7: 1748–1756. study of recombinant human anti-vascular endothelial growth D’Amico TA, Massey M, Herndon II JE, Moore MB, Harpole factor in patients with advanced cancer. J Clin Oncol 19: Jr DH. (1999). A biologic risk model for stage I lung cancer: 843–850. immunohistochemical analysis of 408 patients with the use Govindan R, Page N, Morgensztem D, Read W, Tiemey R, of ten molecular markers. J Thorac Cardiovasc Surg 117: Vlahiotis A et al. (2006). Changing epidemiology of small- 736–743. cell lung cancer in the United States over the last 30 years: Davis DW, McConkey DJ, Abbruzzese JL, Herbst RS. (2003). analysis of the surveillance, epidemiologic, and end results Surrogate markers in antiangiogenesis clinical trials. Br J database. J Clin Oncol 24: 4539–4544. Cancer 89: 8–14. Herbst RS, Johnson DH, Mininberg E, Carbone DP, Delbaldo C, Michiels S, Syz N, Soria JC, Le Chevalier T, Henderson T, Kim ES et al. (2005a). Phase I/II trial Pignon JP. (2004). Benefits of adding a drug to a evaluating the anti-vascular endothelial growth factor single-agent or a 2-agent chemotherapy regimen in advanced monoclonal antibody bevacizumab in combination with non-small-cell lung cancer: a meta-analysis. JAMA 292: the HER-1/epidermal growth factor receptor tyrosine kinase 470–484. inhibitor erlotinib for patients with recurrent non-small-cell Douillard JY, Rosell R, Delena M, Legroumellec A, Torres A, lung cancer. J Clin Oncol 23: 2544–2555. Carpagnano F. (2005). ANITA: Phase III adjuvant vinor- Herbst RS, Mullani NA, Davis DW, Hess KR, McConkey DJ, elbine (N) and cisplatin (P) versus observation (OBS) Charnsangavej C et al. (2002). Development of biologic in completely resected (stage I-III) non-small cell lung markers of response and assessment of antiangiogenic cancer (NSCLC) patients. final results after 70-month activity in a clinical trial of human recombinant endostatin. median follow-up. On behalf of the Adjuvant Navelbine J Clin Oncol 20: 3804–3814. International Trialist Association. J Clin Oncol 23: abstract Herbst RS, Onn A, Sandler A. (2005b). Angiogenesis and 7013, 624S. lung cancer: prognostic and therapeutic implications. J Clin Dowlati A, Gray R, Johnson DH, Schiller JH, Brahmer J, Oncol 23: 3243–3256. Sandler AB. (2006). Prospective correlative assessment of Hurwitz H, Fehrenbacher L, Novotny W, Cartwright T, biomarkers in E4599 randomized phase II/III trial of Hainsworth J, Heim W et al. (2004). Bevacizumab plus carboplatin and paclitaxel {+/À} bevacizumab in advanced irinotecan, fluorouracil, and leucovorin for metastatic non-small cell lung cancer (NSCLC). J Clin Oncol (Meeting colorectal cancer. N Engl J Med 350: 2335–2342. Abstract) 24: 7027–702, 370S. Jain RK. (2001). Normalizing tumor vasculature with anti- Dvorak HF, Brown LF, Detmar M, Dvorak AM. (1995). angiogenic therapy: a new paradigm for combination Vascular permeability factor/vascular endothelial growth therapy. Nat Med 7: 987–989.

Oncogene Bevacizumab in the treatment of NSCLC TE Stinchcombe and MA Socinski 3698 Jemal A, Siegel R, Ward E, Murray T, Xu J, Smigal C radiographic risk factors associated with early onset, severe et al. (2006). Cancer statistics, 2006. CA Cancer J Clin 56: pulmonary hemorrhage in bevacizumab-treated patients 106–130. with advanced non-small cell lung cancer (NSCLC). Journal Johnson DH, Fehrenbacher L, Novotny WF, Herbst RS, of Clinical Oncology, 2006, ASCO, Annual Meeting Proceed- Nemunaitis JJ, Jablons DM et al. (2004). Randomized phase ings Part I 24: Abstract 7068. II trial comparing bevacizumab plus carboplatin and Schellinger PD, Meinck HM, Thron A. (1999). Diagnostic paclitaxel with carboplatin and paclitaxel alone in previously accuracy of MRI compared to CCT in patients with brain untreated locally advanced or metastatic non-small-cell metastases. J Neurooncol 44: 275–281. lung cancer. J Clin Oncol 22: 2184–2191. Schiller J, Harrington D, Belani C, Langer C, Sandler A, Lucchi M, Fontanini G, Mussi A, Vignati S, Ribechini A, Krook J et al. (2002). Comparison of four chemotherapy Menconi GF et al. (1997). Tumor angiogenesis and biologic regimens for advanced non-small cell lung cancer. N Engl J markers in resected stage I NSCLC. Eur J Cardiothorac Surg Med 346: 92–98. 12: 535–541. Senger DR, Galli SJ, Dvorak AM, Perruzzi CA, Harvey VS, Lynch TJ, Bell DW, Sordella R, Gurubhagavatula S, Dvorak HF. (1983). Tumor cells secrete a vascular perme- Okimoto RA, Brannigan BW et al. (2004). Activating ability factor that promotes accumulation of ascites fluid. mutations in the epidermal growth factor receptor under- Science 219: 983–985. lying responsiveness of non-small-cell lung cancer to Treat J, Belani CP, Edelman MJ, Socinski MA, Ansari RH, gefitinib. N Engl J Med 350: 2129–2139. Obasaju CK et al. (2005). A randomized phase III trial of Margolin K, Gordon MS, Holmgren E, Gaudreaul J, Novotny W, gemcitabine in combination with carboplatin or paclitaxel Fyfe G et al. (2001). Phase Ib trial of intravenous recombinant versus paclitaxel plus carboplatin in advanced (Stage IIIB, humanized monoclonal antibody to vascular endothelial growth IV) non-small cell lung cancer (NSCLC): update of the factor in combination with chemotherapy in patients with Alpha Oncology trial (A1-99002L). J Clin Oncol 23: 1096S- advanced cancer: pharmacologic and long-term safety data. abstract 7025, 627s. J Clin Oncol 19: 851–856. Ushijima C, Tsukamoto S, Yamazaki K, Yoshino I, Sugio K, Morgan B, Thomas AL, Drevs J, HennigJ, Buchert M, Jivan A Sugimachi K. (2001). High vascularity in the peripheral et al. (2003). Dynamic contrast-enhanced magnetic resonance region of non-small cell lung cancer tissue is associated with imaging as a biomarker for the pharmacological response of tumor progression. Lung Cancer 34: 233–241. PTK787/ZK 222584, an inhibitor of the vascular endothelial Wakelee HA, Bernardo P, Johnson DH, Schiller JH. (2006a). growth factor receptor tyrosine kinases, in patients with Changes in the natural history of nonsmall cell lung cancer advanced colorectal cancer and liver metastases: results from (NSCLC) – comparison of outcomes and characteristics in two phase I studies. JClinOncol21: 3955–3964. patients with advanced NSCLC entered in Eastern Paez JG, Janne PA, Lee JC, Tracy S, Greulich H, Gabriel S et al. Cooperative Oncology Group trials before and after 1990. (2004). EGFR mutations in lung cancer: correlation with Cancer 106: 2208–2217. clinical response to gefitinib therapy. Science 304: 1497–1500. Wakelee HA, Schiller JH, Gandara DR. (2006b). Current Pao W, Miller V, Zakowski M, Doherty J, Politi K, Sarkaria I status of adjuvant chemotherapy for stage IB non-small-cell et al. (2004). EGF receptor gene mutations are common in lung cancer: implications for the New Intergroup Trial. Clin lung cancers from ‘never smokers’ and are associated with Lung Cancer 8: 18–21. sensitivity of tumors to gefitinib and erlotinib. Proc Natl Winton T, Livingston R, Johnson D, Rigas J, Johnston M, Acad Sci USA 101: 13306–13311. Butts C et al. (2005). Vinorelbine plus cisplatin versus Pfister DG, Johnson DH, Azzoli CG, Sause W, Smith TJ, observation in resected non-small-cell lung cancer. N Engl J Baker S et al. (2004). American Society of Clinical Oncology Med 352: 2589–2597. Treatment of Unresectable Non-Small Cell Lung Cancer Yanagawa H, Takeuchi E, Suzuki Y, Ohmoto Y, Bando H, Guideline: Update 2003. J Clin Oncol 22: 330–353. Sone S. (1999). Vascular endothelial growth factor in Robinson CJ, Stringer SE. (2001). The splice variants of malignant pleural effusion associated with lung cancer. vascular endothelial growth factor (VEGF) and their Cancer Immunol Immunother 48: 396–400. receptors. J Cell Sci 114: 853–865. Yano S, Shinohara H, Herbst RS, Kuniyasu H, Bucana CD, Sandler A. (2006). Oral presentation clinical science session: Ellis LM et al. (2000a). Production of experimental angiogenesis inhibitors in lung cancer. 42nd Annual Meeting of malignant pleural effusions is dependent on invasion of the American Society of Clinical Oncology,Altanta,Georgia. the pleura and expression of vascular endothelial growth Sandler A, Gray R, Perry MC, Brahmer J, Sc toiler JH, factor/vascular permeability factor by human lung cancer Dowlati A et al. (2006a). Paclitaxel-carboplatin alone or cells. Am J Pathol 157: 1893–1903. with bevacizumab for non-small-cell-lung cancer. N Engl J Yano T, Tanikawa S, Fujie T, Masutani M, Horie T. (2000b). Med 355: 2542–2550. Vascular endothelial growth factor expression and neovas- Sandler A, Johnson D, Brahmer J, Schiller J, Ostland M, cularisation in non-small cell lung cancer. Eur J Cancer 36: Gray R et al. (2006b). Retrospective study of clinical and 601–609.

Oncogene