Published in the March 2013 Newsletter
Bryan P. Schneider, MD, recipient of the 2012 Young Investigator Award from the ECOG Research and Education Foundation, gave the keynote presentation—Breast Cancer Genomics and Pharmacogenomics: Making it Personal—at the General Session of the ECOG-ACRIN Fall 2012 Group Meeting.
Dr. Schneider is Associate Professor in the Department of Medical and Molecular Genetics, Division of Clinical Pharmacology; Associate Director of the Indiana Institute for Personalized Medicine; and Associate Professor in the Department of Medicine, Division of Hematology/ Oncology, at the Indiana University School of Medicine (Indianapolis, IN). He is also a member of ECOG-ACRIN’s Breast Committee and Vice Chair of the Pharmacogenetics Subcommittee within the Developmental Therapeutics Committee. In 2010, he chaired ECOG’s Education Symposium for Developmental Therapeutics and led the Survivorship Working Group.
We recently caught up with Dr. Schneider and asked him to recap his presentation for our members who may have missed it.
Dr. Schneider, you have been interested in the role of genetic variability to predict effectiveness and toxicity for some time now. Could you provide our readers with some background on your research?
One of the first biomarker studies our group performed was a correlative study to E2100. The E2100 parent trial, which was led by Dr. Kathy Miller—one of my colleagues here at IU and a leader on the ECOG-ACRIN Breast Committee—was a phase III study of first-line metastatic breast cancer that randomized 722 patients to paclitaxel with or without bevacizumab. The results of this trial were published in the New England Journal of Medicine in 2007.
As most of our readers know, bevacizumab is an anti-VEGF therapy. In our correlative work, we discovered that VEGFA SNPs correlated significantly with improved median overall survival in patients who had received bevacizumab, whereas two other VEGFA SNPs correlated with protection from grade 3/4 bevacizumab-induced hypertension. We published this finding in the Journal of Clinical Oncology in 2008.
In addition to E2100, other important trials of bevacizumab in breast cancer include AVADO and RIBBON-1. All trials, with the exception of AVF-2119 (an old refractory trial), have shown that bevacizumab improves both progression-free survival and the overall response rate. That said, none of the trials have demonstrated increased overall survival, and all of them have shown increased toxicity, such as hypertension, headaches, risk of stroke (rarely), thrombosis, proteinuria, or cardiovascular events.
Data from E2100 were responsible for the accelerated approval of bevacizumab by the FDA. Our readers are no doubt aware that the FDA later withdrew its approval of bevacizumab for the treatment of metastatic breast cancer, and also that bevacizumab continues to have approval throughout the European Union for the treatment of metastatic disease. The ambiguity in the risk-to-benefit ratio implied by these actions heightened the need for a biomarker to identify which patients truly benefited from bevacizumab and which patients had toxicity. The biomarker findings from the E2100 correlative study, which I mentioned were published in 2008, served as preliminary data for a proposal that ECOG submitted to the Susan G. Komen for a Cure organization for its Promise Award. The proposal was ultimately funded and helped us move forward with a second correlative study.
Some of our readers may not be familiar with your Komen Promise Award proposal. Could you talk a bit more about it?
Our overarching plan was to identify a signature to determine which patients should receive bevacizumab based on a biased approach (preliminary data) and an unbiased approach (GWAS). The proposal had two very distinct objectives. Aim 1 was to develop a genomic analysis plan to identify biomarkers for efficacy and therapy-induced toxicity in E5103, the phase III FDA registration trial for bevacizumab in the adjuvant breast cancer setting. Aim 2 had several components, including to formally assess the impact of a biomarker on patient and physician decisions; determine clinically meaningful risks and benefits; create decision aids; conduct formalized education programs; and support the advocacy community in educating the patient constituency. Aim 2 is being led by Dr. Ann Partridge, who founded and directs the Program for Young Women with Breast Cancer at the Dana-Farber Cancer Institute and Mary Lou Smith, who is co-founder of the Research Advocacy Network and a patient representative to the ECOG-ACRIN Breast Committee. The hope is to develop, a priori, the definition of a clinically important biomarker–from both patient and physician perspectives.
Additionally, we hope to provide support and education to our patient advocates. For example, patient advocates recently convened in Indianapolis for training, which included a laboratory session where they could see obstacles and challenges at the translational level.
Could you tell us a little more about E5103 and the correlative work that you led?
In the E5103 parent trial, nearly 5000 patients with early-stage breast cancer were randomized to either standard chemotherapy or to one of two arms that combined standard chemotherapy with different durations of bevacizumab. The goal of our correlative work was to define the top genetic variations across the genome and predict effectiveness and toxicity.
We looked at more than one million genetic variations per patient in this large study population. Our final results were based on about 3300 patients, for whom we had DNA. Because we had genome-wide data, we could transcend the bevacizumab question and also look at chemotherapy. The side effects of chemotherapy present a great opportunity to explore genetic variation—for example, why is it the case that when we give the same drug and the same dose, some tolerate it well and others don’t? I was fascinated with the heterogeneity in the severity of side effects in my patients in the clinic. I would walk into a room, and a patient would say that she felt great, but in the next room, the patient would be afraid to walk down the stairs because she couldn’t feel her feet due to taxane-induced peripheral neuropathy. We felt that these differences were connected to genetic predisposition and wanted to investigate them. So, we asked: how does genetic variation affect toxicity and quality of life?
We investigated peripheral neuropathy because it is one of the most common side effects, but it is not specific to bevacizumab. As most readers know, peripheral neuropathy occurs predominately with the taxanes (paclitaxel and docetaxel). In fact, taxane-induced peripheral neuropathy is the most common nonhematologic toxicity. Although several established clinical risk factors for the severity of side effects are known, much of the heterogeneity cannot be explained.
What can you share with us about your findings in E5103?
In E5103, we have found several SNPs that seem to predict toxicity. Variability in several genes may explain why some patients get peripheral neuropathy and some don’t. We concluded that taxane-induced peripheral neuropathy is common and important. The clinical predictors of neuropathy were advanced age (specifically, a 12.9% increase with each decade of life) and African American race (the hazard ratio was 2.1), and the genetic predictors of neuropathy included several SNPS. Importantly, as we move forward, we hope that the findings will be clinically useful for both patients and physicians in determining the risks vs. benefits of treatment and promote the discovery of new treatments for peripheral neuropathy.
Where do you go from here?
We still have a little more work to do to validate E5103 results by replication in another trial. The results of taxane-induced peripheral neuropathy from E5103 were presented at the 2011 ASCO meeting, and those findings served as preliminary data for the receipt of a Conquer Cancer Foundation award from ASCO. Validation of the neuropathy findings from E5103 will be done using data from a completed independent adjuvant randomized phase III trial, which is E1199. Our plan is to study genetic predictors in both Caucasians and African Americans. We anticipate that the optimal genetic predictors will be composed of multiple genes (a signature) as opposed to one SNP.
We also hope that our findings will be more than just an important publication, but ultimately impact clinical care and help patients. Our plan is to continue to work hard through partnerships with our colleagues at the Dana-Farber Cancer Institute and the Research Advocacy Network to really understand what threshold a biomarker has to achieve to be clinically relevant and also to try to strategize alternate approaches in patients who are affected.
We have also looked at markers that predispose patients receiving anthracyclines to congestive heart failure, which is one of the most worrisome toxicities with that category of drugs. If you tell a patient that he or she can’t receive a drug because they are at high risk of congestive heart failure, I think that it is important to also be able to tell them why. We have provocative leads, but more work is being done in this area. The execution of an unbiased genome-wide approach has truly allowed us to optimize the number of important questions we can ask. In addition to evaluating markers for taxane-induced neuropathy and anthracycline-induced congestive heart failure, we can look at many other toxicities, including bevacizumab-induced hypertension and chemotherapy-induced febrile neutropenia.
Let’s circle back and talk a little more about Aim 2 of your proposal, which includes the development of decision-making tools and education initiatives. What can you tell us?
Our goal is to determine what we can do clinically if we find a biomarker, which of course, requires collaboration with people who develop decision-making tools. We have an amazing team of experts in place to formalize a program in which these genetic markers can be used and to understand where these markers can be clinically implemented. A special “thank you” goes out to Mary Lou Smith and Ann Partridge, both of whom I mentioned earlier, and also to Karen Sepucha, who is Director of the Health Decision Sciences Center in the General Medicine Division at Massachusetts General Hospital. They have been instrumental in the development of this program, as well as numerous advocacy efforts.
Thank you so much for spending some time with us today, Dr. Schneider. Is there anything else that you’d like to tell our readership?
Yes, I’d like to say that it’s important for everyone at ECOG-ACRIN to know that I’m grateful for the ECOG trials and being able to work with all of the wonderful people who have made our research contributions possible.
Abbreviations/acronyms used in the Q&A: ASCO stands for American Society of Clinical Oncology; AVADO, an abbreviation for a clinical trial evaluating AVAstin and DOcetaxel; FDA, US Food and Drug Administration; GWAS, genome-wide association study; RIBBON-1, A Study Evaluating the Efficacy and Safety of Bevacizumab in Combination with Chemotherapy in Untreated Metastatic Breast Cancer; SNP, single nucleotide polymorphism; VEGF, vascular endothelial growth factor.