2 Clarke Drive
Suite 100
Cranbury, NJ 08512
© 2024 MJH Life Sciences™ and OncLive - Clinical Oncology News, Cancer Expert Insights. All rights reserved.
Researchers at Dana-Farber Cancer Institute and Brigham and Women’s Hospital are attempting to improve personalized medicine by analyzing DNA samples for known cancer biomarkers.
Barrett Rollins, MD, PhD
Large panel gene assays come with so much information that oncologists may have difficulty putting all of the data to practical use, Barrett Rollins, MD, PhD, chief scientific officer at Dana-Farber Cancer Institute, said. “I think the practitioner understands very well that for certain diseases, there are genetic biomarkers that are very important, and he or she understands how to use that, but when confronted with a report on 400 different mutations at different locations in the genome, it’s not clear to me that people know how to make sense of that,” Rollins said.
Dana-Farber and Brigham and Women’s Hospital have been working on a large scale DNA analysis project that could one day put much more useable information into the hands of oncologists. Known as Profile, it involves the collection and analysis of thousands of tumor samples from cancer clinics. “That’s the next big frontier,” Rollins said. “We—the American oncology community, the research centers, and the cancer centers—need to work harder on translating all of this data into useable, understandable data for the practitioner, and I don’t think we’re there yet.”
One advantage of Profile, begun in 2011, is that it uses preserved samples from pathology labs rather than requiring patients to undergo fresh biopsies. A surprising fact has emerged from the analysis: researchers have found that most of the samples indicate that patients’ cancers are candidates for treatment with FDA-approved or experimental agents. For example, Profile researchers were expecting just 15% of the useable tumor samples would provide clinically actionable information. Seventy-three percent of useable samples have provided actionable information, Rollins said.
“It has become one of the arguments we use with third-party payers that this kind of test really ought to be reimbursed,” Rollins said. As a result, Dana-Farber is working with Massachusetts General Hospital and Blue Cross Blue Shield of Massachusetts to pilot a form of coverage for the Pro le testing. “The big question we’re trying to answer, and this is going to take a couple of years, is: does this change clinical outcomes?” He expects that efforts will show that it does, “and once we do that, I think this will become standard of care reimbursed by insurers.”
OncoPanel is the Profile test, and it reads the codes of hundreds of genes. Rollins said the per-patient cost ranges from $800 to $1000. The tests are conducted using tumor tissue from patients who have been examined at Dana-Farber, Brigham and Women’s, and Boston Children’s Hospital.
Although there are many genetic assays on the market, what sets OncoPanel apart is that it can be modified to examine specific genes that researchers feel confident about, Rollins said. “We like our homegrown test because every couple of months we can swap in tests. It gives us some flexibility that dealing with a [gene assay] company we may not have.” Clinical samples are preserved in formaldehyde and embedded in paraffin wax before they arrive in the laboratory. Rollins explained that researchers wanted to find a way to test without requiring patients to undergo additional biopsies. In addition, soliciting frozen samples from clinics would have entailed a rush and a risk of decay.
Although Profile researchers wanted to spare patients from discomfort and inconvenience, they were concerned that the quality of the preserved samples would not be high enough for use in genetic testing. They were happily proved wrong on that point, Rollins explained. “It really is feasible to get high-grade sequencing information from clinically acquired specimens. Our technology is so robust that routine fixation of tissues for pathology is perfectly fine for our test to be done.”
Half of the samples they received from pathology labs were not useable, but for reasons the researchers hadn’t anticipated. Much of the time, the tissue sample was too small, such as that collected via needle biopsies; there wasn’t any cancer in the sample collected; or the samples were preserved incorrectly. However, among the samples that were acceptable, 90% could be sequenced, Rollins said.
In the search for actionable mutations and other cancer-related DNA abnormalities, the Profile project has created 15,000 genetic profiles of patients’ tumors. These data are made available to researchers through an ambitious data sharing project known as the Project Genomics Evidence Neoplasia Information Exchange, or Project GENIE, which is overseen by the American Association of Cancer Research. The registry collects sample data from multiple institutions and makes it available for outside analysis, hoping that this will contribute to the success of translational research, validate suspected biomarkers of cancer, suggest new uses for approved drugs, and support the argument that payers should broaden their coverage for genomic testing.
Rollins said all of the GENIE data temporarily remain proprietary but are later released for public use, in the interests of keeping science in the public domain. “We’ve done this with the commitment that after a year of exclusivity, this data could be released to the public so that anyone with an interest in looking at cancer genomics could get into this database and ask questions,” Rollins said. “This has been so successful that there are cancer institutions from around the world who are basically petitioning us to join. We just sent out a request for applications for the next round of institutions to join the GENIE consortium.”
So far, biomarkers have proved most useful in targeted therapy, but the potential of the Profile and GENIE projects could extend this utility to immunotherapy, Rollins said, noting that too little is known about how cancers are going to respond to immunotherapies. “We don’t yet have the kind of biomarkers for immunotherapy that we have in targeted therapy. We don’t know for sure what characteristic it is about somebody’s lung cancer that makes them respond to a PD-L1 blocker, but we need to know that, so the next big phase of Profile and these other tumor profiling projects is to incorporate immunologic measurements as well.”