At the Beyond the Genome conference in Washington, DC, this week, Matthew Ellis from Washington University in St. Louis discussed his approach of "genome-forward oncology," which aims to extract maximum medical benefit from next-generation sequencing. When performing sequencing studies to discover something new about disease, it's necessary to do it within trials with excellent clinical annotation and under sample-rich conditions, Ellis said. Further when mutations are discovered, they need to be associated with clinically important phenotypes like drug resistance or disease relapse. There also has to be some kind of way to prioritize mutations as druggable targets, he added.
Some years ago, Ellis said he and his group conducted a trial of neoadjuvant endocrine therapy for breast cancer and found there was a benefit to patients if they were given drugs before surgery. That way, the tumors would regress somewhat, and the surgery would be less severe. The researchers took this as a proof of principle that they could conduct next-generation sequencing and discover interactions between mutations and outcomes in cancer, Ellis said. In a recent study of ER-positive HER2-negative breast cancer, Ellis and his group found several mutations, which they explored to see how they individually affected disease. In the case of PIK3CA mutations, Ellis said, it's possible to find approaches or drugs to trigger apoptosis, and p53 mutations can mark possible failure of treatment and poor prognosis. And in MAP3K1 mutations, the tumor tries to get rid of the kinase because it's apoptotic and tumor suppressing. Though this is not new a finding, Ellis added, it might be that loss of this kinase leads to a survival event for the tumor and resistance to treatment.
There are two benefits to applying genomics to the study of cancer, Ellis said. As researchers get better at targeting pathways, they can make a difference in or transform the treatment of breast cancer if the patient has a mutation. But this kind of research also shows how the mutations present in cancer overlap in interesting ways, and allows researchers to explore what those overlaps mean and how these compound mutations affect outcomes like survival or treatment resistance.
Ellis added that he is not a fan of the "genome reverse" approach of taking a drug that already exists and sequencing genomes to find the tiny populations of patients that have biomarkers that mean they could be treated with the drug. This approach restricts the questions researchers can ask because they are limited by the biology of the drug, he said. The genome forward approach means finding drugs for the patients and the mutations, instead of finding mutations for the drugs.