Scott Gottlieb, MD
The problem with FDA involvement in the regulation of molecular diagnostics testing is not only the esoteric scientific behind them but also the rapid-fire developments in the field that can make slow moving FDA determinations irrelevant by the time they become official, a group of expert panelists stated in the opening salvos of the 4th Annual Patient Centered Oncology Care conference in Baltimore, hosted by AJMC, last Thursday.
“The FDA does not have a good way to solve for these challenges right now,” said Scott Gottlieb, MD, former FDA deputy commissioner for medical and scientific affairs. Gottlieb recalled that during one period of HIV resistance testing the FDA was looking at genomic tests and eventually would give an approval, but they’d be obsolete before anybody could come in for testing. “Doctors realized that these tests changed all the time. Genomics changed as the viruses mutated,” Gottlieb recalled.
The panel discussion was extremely timely, as last week the FDA released a report on why it feels that stiffer regulatory oversight of laboratory testing is necessary, basing its conclusions on 20 case studies of problematic outcomes. “Laboratory developed tests (LDTs) serve an increasingly important role in health care today. They also have become significantly more complex and higher risk, with several notable examples of inaccurate tests placing patients at otherwise avoidable risk,” the report stated in its executive summary.
With a plethora of testing companies and academics involved in the development of these tests, physicians and patients need clear guidance about what they are getting when they order these tests, but it still needs to be determined what type of a general role the FDA will play as a review body—whether to weigh in on clinical validity or clinical utility, the panelists said.
“The customers need to be confident that the product they’re getting and buying is something they can use to make an intelligent decision,” said Michael A. Kolodziej, MD, Aetna’s national medical director for oncology strategy. “A lot of people do not believe that is occurring. When an oncologist orders the test, they have no idea where it’s going and whether it’s being done well. Patients would be horrified by that statement. Payers do not have the expertise to look at all of these very complicated molecular tests. We need some kind of proficiency testing. Is the test being done right? That’s clinical validity.”
Nevertheless, genomic testing is well established by now. There are 155 biomarkers listed on FDA drug labels, according to session moderator Dennis P. Scanlon, PhD, a professor of health policy and administration at Penn State University. “Fifty percent of all cancer treatments in early clinical development rely on biomarker data,” he noted, and despite the problems associated with genomic testing, the value is indisputable. Women with breast cancer who receive a genetic test of tumor prior to treatment enjoy a 34% reduction in chemotherapy, he said.
Numerous other issues are associated with this form of diagnostics, among them problems that complicate the job of establishing utility for patients, the panelists said. Numerous tissue samples are found to be inadequate and a large percentage, upward of 30%, are returned as unacceptable by laboratories; but also, matching tests with patients is a tricky business, and large numbers of tests are performed without true merit, panelists said.
“A key question comes back to what is the test designed to do, and are we selecting the right test?” said panelist Joy Larsen Haidle, MS, CGC, president of the National Society of Genetic Counselors. “We may be getting an accurate result, but test may not have been designed for that particular patient,” she said. “From the patient’s perspective, they have to be able to count on that result. It impacts not just them but their families, too.”
Another side of this problem is that biomarker testing is very prejudicial to large segments of the patient population, meaning that especially when recruiting for clinical trials, it’s hard to assemble a patient cohort that meets the specifications on a particular molecular level. “A drumbeat of trials that are failing is predicted because they haven’t been able to accumulate enough patients,” Scanlon said. “This incidence rate is really the elephant in the room. Various trials have already run into this issue.”
The US genomics testing market was estimated at $5.9 billion in 2011 by research firm Booz Allen Hamilton, with nearly 2900 different tests available in that year. Subsequent research released in June by Grand View Research predicted that the genomics testing market just in the United States would hit $27.87 billion by 2022.