Demonstration of the clinically significant activity of bevacizumab in advanced-stage ovarian cancer has attracted a great deal of interest. Here, we summarize the two positive phase III trials that led to EMA approval of bevacizumab as first-line therapy and discuss the optimum use of the drug in this disease.
by Susana Banerjee and Stan Kaye
In December 2011, two positive phase III trials1, 2 that assessed bevacizumab in patients with ovarian cancer were reported in the New England Journal of Medicine; these results led to the EMA approval of the drug as first-line treatment in combination with carboplatin and paclitaxel for this disease.3 Bevacizumab is currently the most widely tested antiangiogenic agent for the treatment of cancer. Bevacizumab is a monoclonal antibody that targets the VEGF pathway, which has a critical role in ovarian function as well as in the spread of ovarian cancer.4 Therefore, positive results from clinical trials assessing bevacizumab in this notoriously difficult-to-treat disease have been eagerly anticipated.
The first study (GOG-0218) was reported by Burger et al.1 and was a double-blind, three arm, placebo-controlled study in 1873 patients with newly diagnosed stage III (incompletely resected with residual disease >1 cm) or stage IV epithelial ovarian cancer. Patients were randomly assigned to one of three treatments: combination chemotherapy (carboplatin–paclitaxel), carboplatin–paclitaxel chemotherapy plus concurrent bevacizumab, or carboplatin–paclitaxel chemotherapy plus concurrent and maintenance bevacizumab. The bevacizumab dose was 15 mg/kg for up to 22 cycles (15 months total). After a protocol amendment, stage III patients with macroscopic residual disease of ≤1 cm were also included. Nevertheless, all patients enrolled had advanced-stage disease and their overall outlook was worse than those patients assessed in the second study, ICON7.
Perren et al. published the results from the ICON7 study. The trial randomly assigned patients to one of two arms: 1528 patients received carboplatin–paclitaxel chemotherapy with or without concurrent and maintenance bevacizumab. Bevacizumab was given at 7.5 mg/kg (half the dose used in GOG-0218) for a total of 18 cycles (12 months total). In this trial, 9% of patients had high-risk, early-stage disease (FIGO stage I or IIA, clear cell or grade 3 histology) whereas 30% were at the highest risk for progression (FIGO stage IV, or stage III and >1 cm residual disease).
The primary endpoint in both trials was progression-free survival (PFS), which was evaluated using RECIST and Gynecologic Cancer Intergroup (GCIG) CA125 criteria in GOG-0218; only RECIST criteria were included in the assessment in ICON7. Despite key differences, for both studies the primary endpoint was met for concurrent and maintenance bevacizumab. In GOG-0218, median PFS was extended by 3.8 months (14.1 months vs 10.3 months; P<0.001).1 In the ICON7 trial, the median PFS was 17.3 months in the chemotherapy-alone arm compared to 19.0 months with the addition of bevacizumab (HR 0.81; P=0.004).
In GOG-0218, an additional analysis was carried out that did not take account of CA125 progression (that is, only interpreting the response based on RECIST criteria); in this analysis, the median PFS was six months longer in the group receiving bevacizumab (concurrent and as maintenance) compared to the chemotherapy-alone control arm (12 months vs 18 months; HR 0.645; P<0.001).1 However, this analysis, which was required by the regulatory agencies, has been criticised owing to the bias associated with unequal censoring in the two arms.
In ICON7, the magnitude of PFS improvement is relatively modest (1.7 months);2 however, a preplanned analysis demonstrated that the benefit of bevacizumab is greater in patients defined to be at the highest risk of progression. The 3.6-month improvement in PFS seen in this subgroup using restricted means analysis (restricted means 14.5 months vs 18.1 months; HR 0.73; P=0.002) is similar to the difference in PFS reported in GOG-0218 for the equivalent arms (3.8 months).
For the assessment of the effects of bevacizumab treatment on overall survival, final mature data are awaited. However, in ICON7, an improvement in overall survival with bevacizumab in the high-risk group was particularly noteworthy (28.8 months vs 36.6 months; HR 0.64, 95%CI 0.48–0.85; P=0.002).2 The demonstration of a survival benefit of almost eight months in patients with a poor prognosis is very encouraging.
Toxic effects were as expected, with hypertension grade ≥2 being common (23% of patients in the GOG-0218 study; 18% of patients in the ICON7 study) but generally well controlled. Overall, bevacizumab treatment was well tolerated. Although bowel perforations had been reported in earlier bevacizumab trials,5 these perforations were rare events in GOG-0218 (<3% of the patients) and ICON7 (1% of the patients). However, the incidence was higher with bevacizumab therapy compared to control arms.
Based on these new trial results, is it possible to say that bevacizumab is the new standard of care? To answer this, several questions need to be addressed. First, which patients should be offered bevacizumab? Although both studies met their primary endpoints for the whole trial population, it could be argued that given the overall survival benefit seen in high-risk patients in ICON7,2 women with stage IV or stage III >1 cm residual disease should be considered for first-line treatment. The OCEANS study,6 in which patients with recurrent platinum-sensitive disease were treated with bevacizumab in combination with chemotherapy (carboplatin with gem-citabine), provides a new dimension to this issue. This study reported a significant improvement in PFS with the addition of bevacizumab (8.4 months vs 12.4 months; HR 0.48; P<0.0001) and strongly suggests a role for bevacizumab in this setting of recurrent disease.5 Therefore, a reasonable proposal for patients optimally debulked and thus at a lower risk of early relapse would be to reserve bevacizumab until first recurrence.
The second question is what is the optimal dose of bevacizumab? The licensed dose of bevacizumab, based on the PFS data of GOG-0218, is 15 mg/kg. However, when comparing PFS improvement in a similar patient population (high-risk) in ICON7, there is no difference in PFS improvement between the groups receiving 15 mg/kg and 7.5 mg/kg. The 7.5 mg/kg dose is likely to be more cost-effective and, so far, this is the dose which is associated with an overall survival benefit.
Based on the available data, should bevacizumab maintenance be extended until disease progression? The maximal treatment effect, as indicated by the greatest separation of PFS curves in GOG-0218 and ICON7, coincided with the end of planned bevacizumab treatment. When bevacizumab is discontinued, the impression is that the disease returns promptly and this is in keeping with observations in other cancers. Results from the OCEANS study,6 seemingly superior to the GOG-0218 and ICON7 results, were achieved when bevacizumab was continued until disease progression. Taken together, these findings suggest that bevacizumab therapy until disease progression is warranted.
A fourth question is: should bevacizumab be given in combination with chemotherapy (in addition to maintenance) for first-line therapy? The lack of PFS difference between the chemotherapy-alone control arm and the concurrent bevacizumab arm in GOG-0218 would suggest that the main impact of bevacizumab is as maintenance treatment post chemotherapy. However, the significantly increased response rates (48% vs 67%; P<0.0001) in the subset of patients with measurable disease following debulking surgery in the bevacizumab arm of the ICON7 trial, and in the OCEANS study (57% vs 79%; P<0.0001), indicates clearly that bevacizumab enhances chemosensitivity, and its omission from concurrent treatment may be unwise.
Finally, does the extent of benefit reported so far justify the cost? For those patients with the worst initial outlook, a PFS improvement of four months translates into almost double the time without chemotherapy before the first recurrence. This improvement does represent an important clinical benefit and patient selection is therefore paramount. The identification of a group of patients likely to benefit most from bevacizumab treatment could tip the balance towards a cost-effective therapy.
These important studies by Burger et al.1 and Perren et al. demonstrate that the anti-VEGF strategy has real potential in ovarian cancer. In addition to bevacizumab, other agents targeting this pathway are in active development and future trials will undoubtedly clarify the best strategy to use all these approaches for the benefit of our patients.
Competing interests: Stan Kaye declares an association with Roche. See the article online for full details of the relationship. Susana Banerjee declares no competing interests
Author affiliations: Susana Banerjee, the Royal Marsden NHS Foundation Trust, London, UK; Stan Kaye, the Institute of Cancer Research and Royal Marsden Hospital, Sutton, UK BOX P47,
Practice points: The addition of bevacizumab given concurrently with chemotherapy and continued as maintenance treatment significantly increases progression-free survival as first-line therapy for ovarian cancer, in particular for those patients at high risk of progression.
This article was first published in Nature Reviews Clinical Oncology on 28 February 2012, and is published with permission. © 2012 Nature Publishing Group. doi:10.1038/nrclinonc.2012.28