Originally published Nov. 1.
A retrospective study led by researchers at the University of Bristol has found that the ratio of the VEGF-A splice isoform VEGF165b to total VEGF levels could potentially be used to predict the response of colorectal cancer patients to Avastin.
In the study, published last week in Clinical Cancer Research, researchers from several institutions studied about 100 blinded tumor samples from a Phase III trial of Roche's anti-angiogenic drug Avastin (bevacizumab) with or without FOLFOX4, using immunohistochemistry to score each sample for levels of VEGF isoforms relative to normal tissue.
Using a predictor based on the ratio of VEGF165b to total VEGF, the team was able to show that patients with lower than average relative VEGF165b levels showed “significantly better outcome” based on measurements of progression-free survival — eight months versus five months — when taking Avastin and FOLFOX4 versus FOLFOX4 alone. For patients with higher than average relative levels, there was no difference in PFS between those treated with Avastin and those treated with only FOLFOX4.
Based on the results, the group reported that a low VEGF165b to total VEGF ratio “may be a predictive marker for becavizumab [response] in metastatic colorectal cancer.”
David Bates, the study's lead author, told PGx Reporter in an e-mail that he and his team are now approaching Roche and government funding sources to support a larger trial to quantitatively assess the levels of VEGF isoforms to more precisely stratify which patients are more likely to benefit from Avastin. If the study results could be replicated in a larger trial, the team would hope to develop a test based on VEGF isoform ratios to predict response to the drug.
Bates said that earlier work by the group had shown over-expression of VEGF165b inhibited the effect of Avastin in mice, leading to the hypothesis that patients with low relative VEGF165b levels would do better on Avastin than those with higher levels.
According to the group, Avastin has been shown to increase survival of colorectal cancer in around 10 percent to 15 percent of patients, but it has been impossible thus far to predict who will benefit. In a release announcing the publication of the paper, Bates said that Avastin is not currently approved for colorectal cancer treatment by the UK's National Institute for Health and Clinical Excellence because “the benefit to an unknown minority of patients does not justify the cost of treatment.” With the help of a reliable predictive test to identify best repsonders, NICE could find the cost-benefit profile of the drug more favorable for the CRC population.
Colorectal cancer is not the only area in which Avastin has failed to gain approval based on an inability to identify subsets of responders. Last year, the US Food and Drug Administration revoked Avastin's breast cancer indication "after concluding that the drug has not been shown to be safe and effective for that use." The agency said Roche's subsidiary Genentech was free, however, to move forward with a planned trial focusing on whether patients with high levels of plasma VEGF-A have a superior response to Avastin compared to those with low levels (PGx Reporter 11/30/2011).
In their colorectal cancer study, Bates and his colleagues studied samples from patients in a previous clinical trial, E3200, reassessing 44 subjects from the combined treatment arm and 53 patients in the FOLFOX4-only treatment arm.
The group used a subjective measure — immunohistochemistry scoring — to record ratios of VEGF165b relative to total VEGF levels for each subject relative to normal tissue. Then, the researchers compared patient outcomes in those with ratios above and below the group average.
Overall, subjects with lower than average relative VEGF165b showed a three-month longer PFS with Avastin and FOLFOX4 than with FOLFOX4 alone, while those with a higher than average ratio did the same regardless of treatment arm – meaning they got no benefit from Avastin.
According to Bates and his colleagues, one possible reason for better Avastin response in those with lower levels of VEGF165b is that VEGF165b, itself a growth inhibitor, binds more of the drug than its sister molecule VEGF165b, diminishing the endogenous anti-angiogenic activity of VEGF165b.
Low VEGF165b might also fuel more angiogenic tumors, making patients more responsive to Avastin treatment, the group wrote.
Bates and his colleagues note in the paper that the study is limited by its small size. While the group also looked at overall survival in addition to progression-free survival, the study cohort was not large enough to detect a statistically significant difference in OS in patients treated with Avastin, he told PGx Reporter in his e-mail.
He added that it will be especially important, moving forward, to conduct quantitative analysis of VEGF levels, rather than using a subjective system of imunohistochemical scoring.
Additionally, the group stressed that the results from the study are also unlikely to carry across different disease types. VEGF concentration has been a subject of study in a number of other cancer types as a potential predictor of Avastin response.
In an exploratory study led by the Scripps Research Institute and presented at the American Society of Clinical Oncology's annual meeting this year, researchers compared treatment with Avastin plus paclitaxel to paclitaxel alone in breast cancer patients using a 13-gene VEGF signature developed at the University of North Carolina. The so-called "Hypoxia Signature" has been shown in previous studies to characterize poor outcomes and distant metastases in breast, lung, and brain cancer patients.
The Scripps researchers found that progression-free survival was the same in both treatment arms, despite whether patients had low or high VEGF signatures. However, patients with the high VEGF signature in both treatments arms experienced an overall survival benefit compared to those with a low VEGF signature (PGx 9/5/2012).
Bates said he is currently in conversation with diagnostics companies and with Roche to discuss the possibility of future research into using VEGF165b ratios to predict Avastin response, but “nothing is funded yet.”