NEW YORK (GenomeWeb) – In the current uncertain regulatory and reimbursement landscape for molecular diagnostics, many genetic counselors are questioning whether they should recommend that ovarian cancer patients who already know they have a BRCA gene mutation be retested by a US Food and Drug Administration-approved companion diagnostic in order to receive a new drug.
The FDA approved AstraZeneca's Lynparza (olaparib) in December for advanced ovarian cancer patients who have received three or more prior lines of chemotherapy and who harbor germline BRCA mutations "as detected by an FDA-approved test." The agency simultaneously approved Myriad Genetics' BRACAnalysis CDx, a blood test that doctors can use to determine whether patients have inherited mutations in the BRCA1 and BRCA2 genes and therefore have a good shot of responding well to the drug.
However, these products have entered the market during an unsettled period within the molecular diagnostics space. New genetic tests, employing advanced sequencing technologies, are entering the market claiming improvements upon older technologies. The regulatory status of tests developed by labs is currently in flux as the FDA has issued preliminary plans to oversee their safety and effectiveness. And the extent to which insurers are willing to pay for old and new testing strategies remains a pressure point for diagnsotics developers. In this environment, oncologists must figure out how best to establish the BRCA status of an ovarian cancer patient before they can decide whether to prescribe Lynparza.
Lynparza is part of a new class of drugs called PARP inhibitors, which are designed to block a protein involved in repairing breaks in DNA. In clinical trials, the drugs seem to work particularly well in patients with BRCA-mutated tumors who already have hobbled DNA-fixing capabilities. By giving a patient Lynparza, the idea is to kill the cancer cells by overwhelming them with DNA damage.
Although the indication section on Lynparza's label doesn't mention Myriad's BRACAnalysis CDx by brand name, it is mentioned in the clinical studies section of the label as the test that was studied with the drug and is currently the only FDA-approved companion test available for identifying best responders to the treatment. Conversely, the label for BRACAnalysis CDx does mention that the test is intended to identify patients eligible to receive Lynparza. That's because the test was used in clinical trials for Lynparza and analyzed samples from patients treated with the therapy. The labeling language reflects the FDA's regulatory stance that companion diagnostics – tests used to determine whether or not a patient receives a specific drug – are devices that need to be reviewed by the agency before they enter the market. The agency has issued a guidance on this topic.
But genetic counselors aren't convinced that in the case of Lynparza the FDA approved test is necessarily the best option for every ovarian cancer patient. When a woman inherits a BRCA mutation it increases her risk of breast and ovarian cancer. Based on her family and medical history, she could receive testing for the mutations well before she develops ovarian cancer or is eligible to receive Lynparza. Those women could have been tested using a number of different diagnostics – Myriad's non-FDA approved, non-CDx version of BRACAnalysis, its next-generation sequencing based 25-gene panel, called MyRisk Hereditary Cancer, or a test performed by another lab.
In order to receive Lynparza as a covered benefit, however, many genetic counselors are wondering whether these women have to be retested on Myriad's FDA-approved companion diagnostic. Some of these counselors, advising patients about their testing options, feel FDA's policy on companion diagnostics doesn't quite fit the real world scenarios in which cancer patients receive care. Others have pointed out that for payors, focused on reining in unnecessary spending in the healthcare system, retesting patients may not be cost-effective.
"It's a weird, interim time right now, and we're all struggling to figure out how this is going to affect patient care," Andrea Forman, a genetic counselor at Fox Chase Cancer Center, told GenomeWeb. "Is this the test we have to order when we see an ovarian cancer patient because she might need this medication in the future? But it's not necessarily the best test we have to offer patients. So it's frustrating, especially if insurance is going to be charged twice."
Finally, there is the concern as to whether a patient previously tested for BRCA mutation by another lab or even by Myriad's non-FDA approved version of BRACAnalysis received the right result. If doctors prescribe Lynparza based on a previously conducted test and the result wrongly determined she a deleterious BRCA mutation, then the patient may receive a drug from which she is unlikely to benefit. This is one of the concerns that the FDA has repeatedly cited about laboratory-developed tests (LDTs) that currently aren't overseen by the agency.
It's unclear whether AstraZeneca is at all concerned that patients not tested on the FDA-approved CDx might be prescribed its drug. AstraZeneca didn't respond to questions for this article.
Retesting rates for BRACAnalysis CDx could certainly impact Myriad's revenues. In a statement, Myriad spokesperson Ron Rogers highlighted the language on the FDA-approved labels for Lynparza and BRACAnalysis CDx."Our marketing efforts on behalf of BRACAnalysis CDx must be consistent with the FDA-approved product label for Lynparza and the 'Intended Use' statement about BRACAnalysis CDx," Rogers said.
Myriad began selling BRACAnalysis as a lab-developed test in 1996, and until the summer of 2013, that was the only commercially available testing option for women who wanted to learn if they had a heightened risk of hereditary breast and ovarian cancer. So, conceivably, some of the ovarian cancer patients to whom oncologists are currently thinking of prescribing Lynparza would have been tested on the non-CDx, non-FDA-approved version of BRACAnalysis.
"We're talking about patients who have had their third recurrence of ovarian cancer. So, they're being treated for the fourth time. Some of these patients underwent genetic testing eight to 10 years ago, when it was all being done through Myriad," Elizabeth Swisher, professor of obstetrics and gynecology and an adjunct professor of medical genetics at the University of Washington, told GenomeWeb.
Myriad estimates there are more than 20,000 newly diagnosed ovarian cancer patients in the US annually and 25 percent of these women have received BRCA testing. Due to the low testing prevalence among newly diagnosed patients, the company figures that as many as 44,000 women diagnosed with ovarian cancer in the past five years have not been tested for their BRCA status. "Of those who were tested with our legacy BRCAnalysis test prior to FDA approval, we estimate there are 1,500 patients who were found to have a deleterious BRCA mutation," Rogers said. "We recommend these patients speak with a licensed physician who has discretion to discuss all available treatment options."
According to Joy Larsen Haidle, president of the National Society of Genetic Counselors, Myriad's genetic counselors have told her that they are running the FDA-cleared CDx in a different portion of their lab set up to meet the agency's regulations. However, Haidle told GenomeWeb that Myriad's counselors also said that BRACAnalysis and BRACAnalysis CDx are "analytically the same."
God bless the FDA. They're doing their best, but they didn't get it.
"Both BRACAnalysis and BRACAnalysis CDx sequence the BRCA1 and 2 genes to determine if there are mutations in either of these genes," Myriad states on a web page about the companion diagnostic. "BRACAnalysis CDx is the only FDA approved laboratory developed test that indicates whether or not a patient with ovarian cancer may be eligible for treatment with Lynparza."
But the retesting confusion is not just about women who have received testing from a non-FDA approved Myriad test. The BRCA genetic testing market suddenly became very crowded in 2013, when the US Supreme Court invalidated several of Myriad's patent claims on isolated BRCA sequences. Myriad had asserted many of these patents over the years to keep other labs from offering commercial BRCA testing. Although the company has maintained that it didn’t use its IP position to restrict research, this didn't endear the company to researchers and genetic counselors, many of whom banded together with support from the American Civil Liberties Union to try to invalidate Myriad's patents in court.
After the Supreme Court's ruling, a number of labs jumped at the chance to offer competing tests, some of which employ next-generation sequencing platforms and gauge a number of genes, in addition to BRCA1/2, associated with breast and ovarian cancer. Having tested more than a million patient samples during BRACAnalysis' reign as the only commercial option, Myriad has maintained that its test is the most accurate in determining whether a BRCA mutation is “deleterious,” or associated with cancer risk. As of 2013, Myriad's lab couldn't definitively classify as deleterious or benign 0.6 percent of BRCA1 mutations and 1.6 percent of BRCA2 mutations detected by its test.
Competing labs acknowledge that Myriad's test has the most market experience, but they also claim that their tests are accurate, competitively priced, yield relatively low rates of variants of unclear classification and the rate is getting lower with the help of data-sharing efforts. For example, relying on data from pubic databases, researchers from Invitae have shown 99.8 percent concordance between its BRCA test results and Myriad's.
Still, genetic counselors say that the message they have gotten from Myriad is that patients who have been tested for BRCA mutations by other labs need to be retested by its FDA-approved test. On a "frequently asked questions" web page for the BRACAnalysis CDx, under the question, "Do I need to retest patients who have BRCA results from another laboratory?" Myriad states that patient samples tested by another lab "will receive the BRACAnalysis CDx test."
According to NextGxDx, a health IT company that curates information about available genetic tests, there are 30 labs conducting BRCA testing in the US as single-gene tests or as part of a multi-gene panel. Of these, 26 labs (including Myriad) are gauging germline mutations, and three labs are doing tissue-based testing.
Given the numerous options on the market, oncologists have to consider a number of factors in choosing the right BRCA test for their patient. With an eye toward future treatment with Lynparza, should doctors only test a woman on the FDA-approved BRACAnalysis CDx early in her disease cycle? Or should they go with a broader panel test, like Myriad's 25-gene panel myRisk test (list priced at around $4,000) or Invitae's 29-gene panel (list priced at $1,500)? Should they test just for germline mutations for hereditary cancer risk or also for mutations in the patients' tumor?
It's worth noting that BRACAnalysis CDx has FDA's approval but it gauges germline mutations in the blood – inherited markers that occur in every cell in the body. As such, it would miss somatic mutations that only show up in a patient's tumor tissue. Myriad CEO Pete Meldrum said during a recent earnings call that the so-called BRACAnalysis Tumor CDx the firm is launching in Europe will detect 30 percent more responders to Lynparza than the conventional germline BRCA test. Myriad and AstraZeneca have not publicized plans to launch the tumor CDx in the US for Lynparza.
Ultimately, in the context of treatment with Lynparza, what matters, according to Maurie Markman, president of medicine and science at the Cancer Treatment Centers of America, is whether a patient has BRCA abnormalities in the cancer cells. In his view, the present confusion is directly attributable to the word "germline" in Lynparza's label, where the FDA indicates the drug for patients with inherited mutations. Most women with BRCA mutations inherit it in their germline, which means they also have the mutation in the tumor; but some have somatic mutations, where the genetic abnormality is restricted to the tumor. "The concern and the confusion we have right now is the direct result, I believe, of an error," he said. "Is it in the germline? Quite frankly, it may not even be. It might actually be in the tumor and we know that actually happens."
"But it's a complicated world," Markman reflected, explaining that when a woman has a BRCA mutation it raises a lot of questions about her risk of breast, ovarian, and other cancers, whether she should get surgery, and the type of treatment she receives. A somatic mutation has immediate implications for the carrier's individual prognosis and treatment. But depending on whether that mutation is in the germline, that also has implications for her family's risk of inheriting the same mutation and the disease. So, Markman isn't surprised that this issue has caused confusion among genetic counselors. "These conversations are related but they are really different," he said.
"God bless the FDA," Markman said. "They're doing their best, but they didn't get it."
A fake market?
So far, Swisher and her colleagues at UW have successfully prescribed Lynparza to two ovarian cancer patients who already knew their BRCA mutation status. On the Lynparza ordering form where it asks whether the "patient has a deleterious or suspected deleterious germline BRCA mutation (as detected by an FDA-approved test)," the oncologist just wrote in the mutation the patient had. The insurer in those instances covered Lynparza. "It would be completely redundant to test somebody with a known mutation, which has been identified in a reliable laboratory," Swisher said.
At Cleveland Clinic, doctors have also prescribed Lynparza to ovarian cancer patients "with success," Caris Eng, director of the Center for Personalized Genetic Healthcare, told GenomeWeb. But Eng and her colleagues have only given the drug to patients who haven't had BRCA testing, and so they tested these patients on the FDA-approved CDx.
However, in Eng's view, typically a test performed by an appropriately certified lab provides sufficient guarantee of its accuracy so that the results can be used in clinical trials, as well as for prescribing a targeted drug. Historically, LDTs have been regulated by the Centers for Medicare & Medicaid Services according to the Clinical Laboratory Improvement Amendments (CLIA). Many complex labs doing clinical testing are also accredited by the College of American Pathologists (CAP). These LDTs have long existed alongside and competed with tests with the same intended use approved by the FDA.
So, if one of her patients has already been analyzed for a biomarker by a CLIA/CAP-certified lab, would Eng retest that patient for the same marker using an FDA-approved companion diagnostic? Eng said she would not. "I try to be careful with healthcare resources," she said. "For other trials and [in other] instances … CLIA/CAP-approved test results have been perfectly acceptable."
Eng is not alone in her view. When the agency in 2011 approved Genentech's melanoma drug Zelboraf alongside Roche's BRAF mutation companion diagnostic, some hospitals refused to adopt the FDA-green-lighted kit. Labs at these hospitals had already been performing their own internally developed tests to gauge BRAF mutations in patients. As these lab professionals saw it, they were more comfortable running their own LDTs, and incorporating the new FDA-approved test was an unnecessary expense.
But this is problematic for companies like Roche and Myriad, which have dedicated significant time and resources toward taking tests through the FDA regulatory review process. FDA-approved companion tests are validated in trials where the tests analyze samples from patients who received a specific drug. While unapproved LDTs may have been analytically validated by labs according to CLIA and CAP requirements, the FDA seems particularly concerned that the tests may not have been validated using patient samples treated with the drug of interest.
On the BRACAnalysis CDx "frequently asked questions" page, Myriad lists all the work that went into garnering FDA approval: 17 non-clinical analytical verification studies; six comparator studies; one clinical bridging study; two extraction studies; a process validation study; and validation of equipment and software. The firm also underwent two onsite laboratory inspections and a variant classification process.
All it does is it creates a fake market.
FDA's policies on companion diagnostics are promulgated through guidances, one specifically on companion tests and a draft plan for regulating LDTs more broadly. In both of these documents, the FDA makes clear that companion diagnostics are devices that require premarket review.
The agency's LDT oversight plan, in particular, hasn't had a warm reception among lab industry players, who would point to the retesting confusion with Lynparza as a prime example of why that framework doesn't fit how medicine is actually practiced. FDA's plan to regulate companion diagnostics "should be blown out of the water" because it doesn't work, Swisher said. "All it does is it creates a fake market." Companies may be spending a lot of money to take their tests through FDA approval, she acknowledged, but by the time the FDA approves a test, the state of the art has moved on, other labs have developed better tests, treatment guidelines are different, and the FDA-approved test is outdated.
For example, the National Comprehensive Cancer Network last year updated guidelines to recommend that healthcare providers test colorectal cancer patients considering treatment with the drugs Erbitux (cetuximab) and Vectibix (panitumumab) for KRAS and NRAS mutations. The FDA first updated the labels of these drugs in 2009 to note that patients with mutations in the KRAS gene were unlikely to respond to the therapies, and subsequently approved companion diagnostics that identified those with KRAS mutations. Noting that guidelines now recommend testing also for NRAS mutations, Swisher wondered whether healthcare providers were still expected to follow FDA's labeling and order tests that didn't pick up biomarkers, which according to treatment standards are now considered important for informing patient treatment.
Where do payors stand?
Ultimately, it all depends on what insurers are willing to pay for. "When I talked with a counselor at Myriad, he seemed to suggest that it may come down to the insurance companies to get clarity on if they are willing to accept the results of another lab in order to get authorized access to Lynparza," NSGC's Haidle said. "It feels like a nebulous space right now because it's all so new."
The list price for BRACAnalysis is $4,040, but when the CDx version is prescribed according to insurers' guidelines in the context of administering Lynparza, testing is covered. AstraZeneca has projected Lynparza could bring in as much as $2 billion per year. Noting the high price of most personalized cancer treatments, UW's Swisher said it doesn't make a lot of financial sense for insurers to pay for re-assessing patients if they've already had BRCA testing. "They are already going to be spending a ridiculous amount of money on the drug, why would they add an unnecessary diagnostic to it?" she posited. "I don't know that the insurance companies would require that."
Swisher explained that patients' insurance coverage often dictates the type of BRCA testing she and her colleagues will perform. Sometimes that means they perform their own next-generation sequencing panel, called BROCA, Myriad's test, or another lab's test. Although UW is not currently retesting patients with established BRCA status on the FDA approved CDx in order to prescribe Lynparza, Swisher said that in the unlikely scenario an insurer asks her lab to repeat testing with the FDA-approved test, they would do it.
Aetna's policy leaves open the possibility of coverage for retesting patients for BRCA status to determine if they should receive Lynparza. According to the policy, Aetna will pay for BRCA testing, when it is used for assessing the risk for hereditary breast and ovarian cancer, once in the person's lifetime. But in a footnote to this policy, the insurer adds the following exception: "Repeat BRCA testing with BRACAnalysis CDx (Myriad Genetics) is considered medically necessary for women with ovarian cancer who had another brand of BRCA test and who are being considered for treatment with olaparib (Lynparaza) after three or more previous lines of chemotherapy."
Adam Myrick, a spokesperson for Medicare contractor Palmetto GBA, didn't address the situation with Lynparza and BRCA testing specifically, but provided a more general perspective. He acknowledged that currently "there are many drugs that are premised on specific CDx tests that are covered based on LDT versions of tests for the core biomarker."
But in the future, payors might take a more favorable view toward companion tests with FDA approval over non-approved LDTs. Palmetto runs a program called MolDx, which reviews the evidence on molecular diagnostics in order to determine their coverage. In a recently published document explaining its test evaluation process, Palmetto noted that FDA-approved companion diagnostics have fast-track status where the test will not have to go through further evidence review, but will be reimbursed as long as the test and the drug are deemed medically necessary for Medicare patients. In the past, Palmetto has also granted premium pricing for certain FDA-approved companion diagnostics over LDTs.
It would be a "terrible mistake," according to CTCA's Markman, if an insurance company were to turn down a claim where a woman had her BRCA status determined by a CLIA-certified lab or a tissue-based cancer diagnostic that characterizes markers in the tumor. "I suspect that most insurance companies will understand that what we're looking at is the tumor," Markman said, adding that if a patient received a tissue-based test, the insurer might ask that eventually she receive a germline test to understand the disease risk for her family.
Insurers are already considering the use of next-generation sequencing-based panels to ascertain information on hereditary cancer syndromes. In fact, one of the largest private insurance companies in the US, UnitedHealthCare, inked a three-year contract with Myriad around the use of its NGS-based myRisk panel that gauges patients' risk of eight hereditary cancers.
Myriad's Meldrum said during an earnings call last year that the contract allows UHC enrollees who meet certain eligibility criteria to gain access to myRisk, and enables those who have previously been tested by Myriad’s first-generation risk diagnostics, such as BRACAnalysis, to be retested by the more advanced NGS test. It's unclear how this policy fits with the approval and availability of the BRACAnalysis CDx. UnitedHealthCare didn't agree to an interview for this article.
While "it's absolutely true that the BRACAnalysis CDx has been approved [by the FDA] for the specific purpose of identifying patients for ovarian cancer therapy," said Lakshman Ramamurthy, who heads up FDA regulation and policy matters at the consulting group Avalere Health, he noted that that payors would determine whether or not to pay for the CDx version based on whether it provides information that the previous LDT version doesn't. To retest patients on the CDx would "be a waste of resources only if the new test does not offer new information," said Ramamurthy. "The payors are fairly savvy and they will look to see if there is new information or not."
This article has been updated to specify that Lynparza's indication doesn't mention BRACAnalysis by name. However, the test is mentioned by name in the "clinical studies" section of the drug label.