There is an urgent need for new therapies to treat non-small cell lung cancer (NSCLC) because current chemotherapy regimens are of limited effectiveness. The role of vascular endothelial growth factor in promoting tumor angiogenesis, in maintaining existing vasculature, and in resistance to traditional therapies, together with its negative prognostic significance in NSCLC, make it an appropriate target for therapy. Bevacizumab (Avastin), a monoclonal antibody directed against vascular endothelial growth factor, has shown promise in treating a number of different cancers. In a recent Phase II trial in patients with advanced metastatic NSCLC, the addition of bevacizumab to standard carboplatin/paclitaxel chemotherapy significantly increased the time to progression and increased the response rate when compared with chemotherapy alone. This was particularly impressive in the subset of patients with non-squamous histology. Bevacizumab is generally well tolerated and did not appear to increase the incidence or severity of nausea/vomiting, neuropathy and renal toxicity, which are typically associated with carboplatin/paclitaxel chemotherapy. Adverse events in Phase I and II studies included hypertension, thrombosis, proteinuria (with occasional nephrotic syndrome), and epistaxis. Serious tumor-related bleeding episodes (hemoptysis/hematemesis) seem to be the main safety concern in patients with NSCLC, with squamous cell histology as a possible risk factor. Present ongoing studies are under way in NSCLC including (a) a Phase II neo-adjuvant study in combination with paclitaxel and carboplatin in patients with stage IB–IIA NSCLC; (b) a Phase I/II study of bevacizumab in combination with the epidermal growth factor receptor tyrosine kinase inhibitor agent, Tarceva, in patients with previously treated NSCLC; and (c) an Eastern Cooperative Group randomized Phase III study of paclitaxel and carboplatin with/without bevacizumab in patients with previously untreated IIIB (malignant pleural effusion) or metastatic NSCLC. These studies will help to establish the role of bevacizumab in NSCLC.
Non-small cell lung cancer (NCLSC) constitutes >75% of lung cancers and has proved to be especially intractable to treatment. Surgery provides the only curative treatment, but resection is possible in only 30% of patients at diagnosis, with metastatic disease developing in 50% of patients within 5 years (1) . Unfortunately, despite the development of new chemotherapy regimens, including platinum-based regimens, the prognosis for advanced inoperable NSCLC remains poor. In a recent Eastern Cooperative Oncology Group clinical trial comparing a standard regimen of cisplatin/paclitaxel with three other platinum-based regimens in 1155 patients, the overall response rate was 19%, and median survival was only 8.0 months, with a 33% survival rate at 1 year and 11% at 2 years. No significant differences in these parameters were found between the different treatments. Median time to progression was 3.4 months with cisplatin/paclitaxel; only the cisplatin/gemcitabine group showed a significantly better result (4.2 months; Ref. 2 ). Similar results were reported in another recent comparative study involving carboplatin/paclitaxel and vinorelbine/cisplatin (3) . Because these data represent the best care available at present, there is clearly a pressing need for new therapeutic approaches for NSCLC.
WHY TARGET VEGF IN LUNG CANCER?
Treatment for cancer is now moving beyond traditional chemotherapy with the advent of specific targeted therapies, and much research effort is focused on developing treatments based on the inhibition of tumor angiogenesis. Generally, tumors cannot grow beyond ∼2 mm in diameter without developing a vascular supply (4) . Not only does neovascularization permit further growth of the primary tumor, but it also provides a pathway for migrating tumor cells to gain access to the systemic circulation and to establish distant metastases. Angiogenesis, whether physiological or pathological, is governed by a host of proangiogenic and antiangiogenic factors. Among these, vascular endothelial growth factor (VEGF) is the most potent and specific of the endothelial cell mitogens, acting both as an endothelial cell survival factor and as a key factor in mobilizing circulating endothelial cell precursors to nascent blood vessels (5, 6, 7, 8) . Not only does VEGF promote the vascularization and growth of the primary tumor, but it also seems to play a key role in the early stages of establishing new metastatic foci (8) . VEGF mRNA is up-regulated in the majority of human tumors, and this tends to correlate with poor prognosis (9 , 10) . Thus, inhibition of VEGF offers an especially attractive target for antiangiogenic therapy in cancer.
In NSCLC, the majority of studies support a correlation between VEGF expression, microvessel density (MVD) and poor prognosis. However, several studies suggest that angiogenesis has little or no predictive value (11 , 12) , possibly reflecting methodological differences such as the use of different monoclonal antibodies (12) . In a study of the significance of the various VEGF isoforms in NSCLC, Yuan et al. (13) reported that VEGF-189 showed correlations with high intratumoral MVD, short survival, and early postoperative relapse, whereas VEGF-121 was correlated with short survival and relapse. In contrast, expression of two other isoforms, VEGF-165 and VEGF-206, had no predictive value.
Vascularity also varies within a tumor. Ushijima et al. (14) found strong correlation between peripheral, but not central, MVD counts and poor prognosis; moreover, high peripheral MVD values carried an especially poor prognosis when associated with high expression of VEGF. A similar correlation of poor prognosis with localized elevated MVD at the advancing front of NSCLC tumors, accompanied by increased localized staining of VEGF and its receptor, has also been reported by Koukourakis et al. (15) . Furthermore, these workers were able to differentiate between a standard MVD, and an “activated MVD,” with the latter composed of blood vessels that expressed the VEGF/VEGF-receptor complex. Although the standard-MVD level was higher in nontumorous areas of the lung, activated-MVD levels were higher in the tumor, particularly at the advancing tumor front. Multivariate analysis showed that activated-MVD was the most important independent prognostic factor.
Finally, it should be noted that some NSCLC tumors do not appear to require any neoangiogenesis for their progression. Passalidou et al. (16) reported that 9 of a total of 113 tumors showed no evidence of new blood vessel growth but instead filled up the alveoli and appropriated existing blood vessels within the trapped alveolar septa.
A variety of therapeutic strategies, aimed at blocking VEGF or its receptor signaling system, are currently being developed for the treatment of cancer. The best-studied approaches are the VEGF/VEGF-receptor blockade by monoclonal antibodies and the inhibition of receptor signaling by tyrosine kinase inhibitors.
Bevacizumab [rhuMAb-VEGF (Avastin); Genentech Inc., South San Francisco, CA] is a recombinant humanized monoclonal antibody to VEGF composed of human IgG1 framework regions and antigen-binding complementarity-determining regions from a murine antibody (A.4.6.1) that blocks the binding of human VEGF to its receptors (17) . Bevacizumab is being assessed in a range of cancer types, including NSCLC. Promising results have been seen in breast cancer, colorectal cancer, and, in particular, renal cell cancer (18, 19, 20, 21, 22) . Indeed, the renal cancer study was stopped early because of a highly significant increase in time to progression with bevacizumab monotherapy (22) . In a randomized Phase II trial of 104 patients with previously untreated metastatic colorectal carcinoma comparing 5-fluorouracil plus leucovorin with or without bevacizumab, there was improvement in response rate, time to progression, and survival (20) . These results were confirmed in a larger randomized Phase III trial reported at the 2003 annual meeting of the American Society of Clinical Oncology (21) .
BEVACIZUMAB IN NSCLC
Bevacizumab or its parent murine antibody A.4.6.1 inhibit the growth of various human tumor cell types in murine xenograft models (23, 24, 25) , including the CALU-6 NSCLC model (26) . There was no inhibitory effect exerted on the growth of cancer cells in vitro, and treated tumors showed reduced vascularity (27) and reduced interstitial pressure (25) . In addition to the inhibition of the primary tumor, reduction in metastases was also observed in some studies (28, 29, 30) .
Hypoxia-induced production of VEGF is believed to be important in mediating tumor resistance to radiotherapy and chemotherapy (31) , and the antibody A.4.6.1 significantly augments the antitumor effects of both modalities (26) . Administration of bevacizumab was able to reverse the protective effect of VEGF against the antiangiogenic effects of docetaxel in endothelial cells in vitro and in vivo (32) .
Finally, VEGF has been shown to inhibit the differentiation of dendritic cells, and this inhibition can be reversed by an anti-VEGF antibody, with a corresponding increase in antitumor immune responses (33 , 34) .
Phase II Clinical Trial in NSCLC.
A randomized, multicenter, Phase II trial was conducted involving a total of 99 patients with newly diagnosed stage IIIB (with pleural effusion), stage IV, or recurrent NSCLC (35) . Patients were randomized to receive carboplatin [target area under the curve (AUC) 6 mg/ml/min] + paclitaxel (200 mg/m2) chemotherapy every 3 weeks, or carboplatin + paclitaxel chemotherapy with bevacizumab 7.5 mg/kg (low dose) or 15 mg/kg (high dose) every 3 weeks (Fig. 1)<$REFLINK> . Principal efficacy endpoints were time-to-disease progression and best tumor response rates (complete or partial response), as determined by both the investigators and an independent review facility.
The results are shown in Table 1<$REFLINK> . Bevacizumab (15 mg/kg) plus carboplatin/paclitaxel increased the response rate (31.4 versus 18.8%) and median time to progression (7.4 versus 4.2 months) compared with chemotherapy alone. A modest increase in survival from 13.2 to 14.2 months was also evident (35) . Nineteen patients in the control group who showed disease progression were allowed to cross over to receive high-dose bevacizumab monotherapy; no objective responses were observed in these patients, although 6 of the 19 had one measurement of stable disease (Fig. 2<$REFLINK> ; Ref. 35 ).
In Phase I trials, bevacizumab did not display dose-limiting toxicities when administered alone, nor did it lead to synergistic toxicities when combined with standard chemotherapy regimens (36 , 37) . Adverse events in Phase I and II studies, including the Phase II NSCLC trial (see above), included hypertension, thrombosis, proteinuria (with occasional nephrotic syndrome), and epistaxis. In the extension study, no unexpected adverse events were observed after 1 year of therapy. Deep venous thrombosis occurred after 1 year of treatment, but patients were able to remain on bevacizumab with anticoagulant therapy (30) .
The main tolerability concern in this study was the occurrence of bleeding episodes (26) . Classified as hemoptysis/hematemesis, they occurred in six patients (five in the low-dose group), four of whom died. All six cases appeared to be tumor-related, originating from centrally located pulmonary tumors close to major blood vessels. Cavitation or necrosis of the tumor had occurred in five cases. Such bleeding episodes have not been reported in patients receiving bevacizumab in breast, colorectal, prostate, or renal cell cancer trials (37) . Exploratory analysis found squamous cell histology as a possible risk factor in patients receiving bevacizumab, because this characteristic was found in tumors of four (67%) of the six patients with serious bleeding, whereas only 20 (20%) of 99 of the enrolled subjects had squamous cell histology (38) . In the subset of patients with non-squamous cell histology, the median survival time was increased from 12.3 months to 17.9 months, which was not statistically significant given the small number of patients. Accordingly, patients with this tumor type have been excluded from the ongoing trial E4599. Further information as to whether this issue is of particular concern in lung cancer may come from planned trials of bevacizumab in combination with etoposide plus platinum in patients with small cell lung cancer, in whom central tumors are common.
There was an increased incidence of thrombotic events associated with bevacizumab; however, when thrombotic events related to occlusion of central lines were considered excluded, the number of events was similar across the three treatment arms.
Several ongoing trials are assessing bevacizumab in combination with other therapies in NSCLC (37) . These include the following protocols: (a) Protocol NCI-2655 (AVF-2314s), a Phase II study involving patients with stage IB, II, or IIIA resectable tumors, in which bevacizumab, paclitaxel, and carboplatin are administered as neoadjuvant treatment, with two cycles of treatment at 3-week intervals, followed by surgical resection at week 8; (b) Protocol OSI-2486s, a Phase I/II study, in which bevacizumab (15 mg/kg) i.v. every 21 days is administered in combination with erlotinib (150 mg/day) p.o. (Tarceva), an inhibitor of the epidermal growth factor receptor, to patients with recurrent NSCLC. Data from the Phase I component of the trial (n = 12 patients) show the combination produces a 25% response rate, with an additional 42% of patients classified with stable disease; and (c) Protocol E4599 (AVF-2366s), a Phase II/III study of paclitaxel + carboplatin, with or without bevacizumab, as first-line therapy for patients with advanced, metastatic, or recurrent, non-squamous cell NSCLC. This study is projected to accrue a total of ∼850 patients. In addition to measuring response rates, time to progression and tolerability, this study will also explore the role of plasma VEGF as a prognostic factor, whether basic fibroblast growth factor is a favorable prognostic factor, and whether other surrogate clinical markers, potentially released by damaged endothelial cells, can be identified. Possible candidates for such markers include E-selectin and vascular cell adhesion molecule.
In targeting VEGF, bevacizumab is directed against one of the key factors that promote tumor growth and spread, not only by mediating tumor angiogenesis but also by promoting development of resistance to standard therapies. The encouraging results achieved by adding bevacizumab to standard combination chemotherapy in patients with advanced metastatic NSCLC additional studies, particularly evaluation in earlier stage disease. Ongoing trials will provide more detailed information as to the most effective use of this new drug, including optimal dosage and scheduling with chemotherapy and other targeted therapies, as well as the minimizing of the risk of bleeding episodes.
There are several arguments for using antiangiogenic therapy in combination with other drugs (20 , 41 , 42) . Firstly, the profusion of angiogenic factors that can be produced by tumors suggests that the inhibition of angiogenesis may require the combined action of several inhibitors. Secondly, traditional chemotherapeutic agents also exert antiangiogenic effects of their own, and these effects can be potentiated by agents such as antibodies to VEGF (5) .
Contrary to the assumption that inhibition of angiogenesis would impede the delivery of chemotherapeutic reagents to tumors by reducing their vascularity, the efficacy of chemotherapy is potentiated by coadministration of antiangiogenic therapy (43) . This is because antiangiogenesis agents act to prune and normalize the tumor vascular supply, which is typically aberrant (38) . Finally, inhibiting VEGF can counter tumor resistance to chemotherapy and radiotherapy, and possibly to other antiangiogenic therapies (44) , as well as reduce the enhanced tumor interstitial pressures that impede drug delivery.
It is possible that we may have reached the limits of what conventional chemotherapy can achieve in patients with advanced NSCLC, and the advent of targeted therapies such as bevacizumab is likely to improve the prospects for such patients.
Dr. Thomas Lynch: In that Phase II trial that you and Dr. Herbst are doing, what would it take to make you think that there was a signal here that was worthy of going forward?
Dr. Sandler: In the present trial, I think there is a signal. We don’t have any time-to-progression data or survival data yet. But we each have two or three patients who are now at a year and beyond. When this study was originally written, we were stopping the treatment at 1 year; we actually had to write an amendment to go beyond a year, so we think there is something there.
Dr. Ramaswamy Govindan: The adjuvant study is a maintenance study, correct? For how long?
Dr. Sandler: The way it will initially be written is up to a year.
Dr. Lynch: And the rationale for up to a year?
Dr. Sandler: There is no good rationale. You could say 2 years.
Dr. Lynch: We’ve been struggling with that issue as well.
Dr. Mark Socinski: How would that adjuvant trial be influenced by the results of the stage IV trial?
Dr. Sandler: Well, that’s the issue, whether you can approve something when the metastatic trial is ongoing. The view was to start discussing this design now and maybe we’ll have some data by the time it is actually in formulation. Obviously, if the metastatic trial were negative, the enthusiasm would go down.
Dr. Socinski: If it were positive, could you really get people to agree to the randomization?
Dr. Sandler: I would think so, because we’ve had positive studies in metastatic settings before, and not everybody was receiving chemotherapy.
Dr. Bruce Johnson: Is the bevacizumab trial going to be presented at the 2004 American Society of Clinical Oncologists meeting?
Dr. Sandler: The data safety committee is going to take a look at the data in the next couple of weeks, because the concept was to potentially extend this study to 800 or 900 patients in order to be able to detect a slightly smaller difference, say 25% instead of 33%. The committee is going to look at toxicity and also take sort of an unblinded look at the outcome. If this study stops at 600 patients, it’s possible it could be presented, but I think not. I don’t think it will be ready for this year’s American Society of Clinical Oncologists meeting.
Dr. Paul Bunn: My bias always was that it didn’t make a lot of difference whether the bevacizumab was given early or late. In that Phase II trial, I would have liked to have seen chemotherapy/bevacizumab given concurrently, compared with chemotherapy followed by bevacizumab. If sequential therapy were just as good, which would be my prediction, you probably wouldn’t have to exclude all of those patients because, if they respond, then they’re much less likely to have the bleeding.
Dr. Roy Herbst: I agree that all of these antiangiogenic agents would probably have some applicability in the maintenance approach. There are data that radiation therapy increases VEGF levels; chemotherapy probably does too because it’s a survival factor. So, it’s a perfect time to do it.
Dr. Lynch: What is the European perspective on bevacizumab? Is there as much enthusiasm for it in Europe as there seems to be in the United States?
Dr. Rafael Rosell: In Europe, we have a lot of enthusiasm about the possibility of performing trials with bevacizumab. In Spain particularly, we are on the verge of organizing a trial with a design similar to what Dr. Sandler has recommended, a small Phase II randomized study, but using a combination from the beginning, for instance, docetaxel plus bevacizumab.
Presented at the First International Conference on Novel Agents in the Treatment of Lung Cancer, October 17–18, 2003, Cambridge, Massachusetts.
Requests for reprints: Dr. Alan B. Sandler, Division of Hematology/Oncology, Vanderbilt University Medical Center, 2220 Pierce Avenue, 777 PRB, Nashville, TN 37232-6307. Fax: 615-343-7602; E-mail: