With limited peer-reviewed literature, health plans take an evaluative approach to tumor profile testing
Tests that search for individual genetic biomarkers to guide drug therapy in cancer are giving way to a game-changing technology. Instead of looking for one specific genetic alteration, tumor profiling tests are using both first and next generation sequencing systems to race through millions of DNA base pairs in an entire tumor genome to identify all of the alterations that exist in a cancer patient’s tissue sample.
A tumor profile test may encompass a panel of more than 200 genes. It looks for all alterations that have been associated with any solid tumor cancer. For example a tumor-profiling test may identify the presence of HER2, KRAS, or EGFR alterations in cancers outside of the one where they are well established. Profiling tests also attempt to identify a slew of lesser-known alterations.
“Cancer has long been defined by the tissue of origin, in part because that is all the diagnostic tools allowed us to do,” says Mike Pellini, MD, CEO of Foundation Medicine, a molecular diagnostics company. “Now, we understand there are underlying molecular pathways that are not specific to any one tissue and new technologies that allow us to look for all pathways that are turned off and on.”
Tumor profiling is a leap forward, providing new insight to the genomic drivers and even the heterogeneity of tumors, and its use is skyrocketing.
It has captured the attention of academic researchers, drug companies that use it to identify new targeted therapies, upstart clinical labs that market direct to consumers, and physicians who are looking for information to guide their treatment decisions.
“About 30% to 40 percent of the tests are ordered for late stage disease. This may suggest the initial treatment is no longer working,” says Pellini. “Another bucket is comprised of uncommon or rare cancers, like small-bowel cancer or uncommon diseases that lack clear treatment options — where the oncologist may be looking for a rational approach to therapy. A third example is where only a small tissue sample is available and the physician is concerned that there might not be enough tissue to run a series of individual molecular tests.”
There is a risk that tumor profile reports may overload physicians with a complex list of possible causes of a cancer. The frequency of many of these alterations is generally low but it has become a common practice for profiling labs to suggest that the results may guide treatment. Test reports often say some of the alterations may be “potentially actionable targets.” Some reports also list the drugs that could be used.
Therein lies the problem for clinicians and health plans. Not all molecular profiling is backed by solid evidence from gold standard clinical trials. Yet clinicians face the real-world problem of finding treatment options for tough cancers like those with acquired drug resistance. Health plans face the challenge of meeting their coverage benchmarks of cost effectiveness and clinical utility. “The insurance industry is somewhat cautious; we like evidence from scientific peer-reviewed literature, and these tests are still in the evaluation stage,” says Aetna’s Ira Klein, MD.
A recent meeting of the American Society of Clinical Oncology included several abstracts of molecular profiling studies that showed the ability of these assays to identify new alterations, but most studies so far have shown insight only to the genetic basis of different cancers.
“We are concerned about who will be able to give the most reliable test,” says Klein. Many companies are developing assays with unique gene panels, and there is no specific regulatory oversight of genetic testing by the Clinical Laboratory Improvement Amendments program, better known as CLIA.
Klein says there is another big concern. “The practical application of profiling results and the results of application are not yet known. It is still unclear how and when to use test results to guide therapy.
“At this point physicians do not understand the significance of all of this data,” Klein continues. “Data overload could prompt doctors to order more tests to narrow results, or it may lead to inappropriate treatment decisions.
“One really important aspect of this is that providers do not have the data infrastructure to translate all of the information in a report into outcomes. We need to build data libraries and maintain the data to determine where there are measurable results.”
In May, the U.S. Oncology Network announced a personalized medicine initiative aimed at incorporating molecular profiling and other genetic tests into clinical practice. Its new approach addresses many of the issues Klein identified.
“It’s an attempt to have an organized approach to how we handle precision medicine,” says Marcus Neubauer, MD, medical director of oncology services for McKesson Specialty Health, the parent organization of U.S. Oncology.
The chosen labs are Med Fusion and Foundation Medicine. Med Fusion already has close business ties with U.S. Oncology as a clinical reference lab. Foundation Medicine has developed a tumor assay that is used in several academic medical centers. It encompasses 236 genes covering all solid tumor cancers.
A key component of the initiative is a data system for ordering all types of tests and for receiving lab results including genetic profiles in a consistent electronic format.
“NSCLC ... has become one of the most interesting [cancers] because a number of mutations have been identified that may lead to the use of targeted drugs,” says Marcus Neubauer, MD, medical director of oncology services for McKesson Specialty Health.
“We will beef up the results section in the IKnowMed EHR [specifically for cancer patients] to create structured and searchable information from the profiling reports,” says Neubauer. “When you consider the possible combinations of results from a large gene panel, the only way to gain meaningful insight to the possible value of molecular profiling is to have structured information across a large number of patients. We’re going to start with non-small-cell lung cancer [NSCLC] and breast and colon cancer. These are common diseases with known DNA alterations.
“It used to be that from a pure scientific perspective the treatment of NSCLC was one of the least interesting cancers. Now it has become one of the most interesting because a number of mutations have been identified that may lead to the use of targeted drugs,” says Neubauer. “Lung cancer has become an example of how molecular diagnostics may affect real-world patient care.”
The U.S. Oncology Network is known for its work on oncology pathways. Neubauer says that when sufficient profiling evidence is developed, it will work its way into those pathways.
The U.S. Oncology Network has laid out a clear plan for how it will take advantage of molecular profiling. With a few rare exceptions, the official stance of most health plans on molecular profiling is that it is experimental and not medically necessary. That strategy may work for now, but Foundation Medicine’s Pellini says, “It is not a question of ‘if’ molecular profiling will be covered — it is a matter of when.”