By Turna Ray
This article has been updated from a previous version to clarify a comment by an AstraZeneca official. This article has also been corrected to note that DxS was acquired by Qiagen.
With Medco's entrance into the personalized medicine space, a new strategy is emerging for diagnostic developers looking to commercialize pharmacogenomic tests without a deep-pocketed pharmaceutical partner. And it might even prove to be a better return on investment for the small diagnostic company than the typical Rx/Dx partnership can offer.
At a conference in Baltimore last month, Bryan Dechairo, head of personalized medicine at Medco, invited an audience full of representatives from diagnostic firms to partner with the pharmacy benefit manager.
Dechairo said that Medco is open to working with small diagnostics firms with laboratory-developed phamacogenomic tests, even if the clinical utility of the marker hasn't been proven in randomized-controlled trials. Tests that are approved by the US Food and Drug Administration are nice when available, but Medco will not turn down validated laboratory-developed tests, he added. And while tests that gauge response to drugs that are off-patent or about to go off patent may not be of interest to pharmas, they may be a perfect fit for Medco's Genetics for Generics program, which the pharmacy-benefit manager launched last year with the aim of using genetic testing to incur healthcare savings in the dispensing of generic drugs for its customers (PGx Reporter 10-07-09).
What Medco does concern itself with first and foremost is that the gene markers at the center of its partnerships with Dx firms have been clinically validated in independent sample sets and analyzed in CLIA-certified labs.
In comparison, according to the heads of R&D at several major drug companies, diagnostic developers wishing to partner with them must ensure not only that their tests are clinically valid, clinically useful, and approved by the FDA, but they should also have the marketing and distribution capabilities to accommodate big pharma's multinational operations. And Dx firms can forget about getting a slice of the drug revenue.
Ultimately, in the still emerging and rapidly evolving filed of personalized medicine, Medco is offering independent diagnostic developers the opportunity to obtain clinical utility and cost effectiveness data within its vast network of pharmacists, doctors, insurers, and patients. Having garnered the answers to the utility and cost questions, which often pose the greatest barrier to the adoption of a genetic test, diagnostic developers will then be able to introduce the test to Medco's customers, with a much greater chance of reimbursement and adoption.
In their conception of Rx/Dx partnerships, pharmaceutical companies simply aren't offering this kind of pledge to Dx shops. "This is just a new avenue for diagnostic companies to consider when they are trying to find a market that brings value to patients," Dechairo told Pharmacogenomics Reporter. "We're just a new partner in that mix.
And with the annual revenues from a companion diagnostic estimated to be between $10 million and $50 million in a typical Rx/Dx partnership, Dx firms would welcome new strategies that would distribute a return on investment more equitably. In this regard, Dechairo, who comes to Medco from Pfizer, believes the PBM "can improve the value proposition for diagnostic companies" who are trying to bring pharmacogenomic tests to market independently.
The Big Pharma Bump
Conventional wisdom holds that firms looking to get into the companion diagnostics business need the help of a drug company to validate the gene response markers in large Phase II and Phase III trials, and to help subsidize the development and commercialization of the test.
One example of this classic structure is Pfizer's partnership with Monogram Biosciences to develop a companion test for the HIV drug Selzentry. As part of the deal, Pfizer took over global distribution for the test and invested $25 million in the diagnostics company (PGx Reporter 05-10-06). The drug giant also took up the responsibility of distributing the diagnostic alongside the drug in multinational markets.
A slightly different example is the experience of DxS. The UK-based personalized medicine firm independently developed a test to gauge KRAS mutations, which indicate patients unlikely to respond to EGFR-inhibiting colorectal cancer drugs Vectibix and Erbitux. However, drug makers didn't come around to inking deals with DxS for a companion diagnostic until Amgen, the developer of Vectibix, was denied a marketing license by European regulatory authorities (PGx Reporter 02-24-10).
The examples of Monogram and DxS suggest that partnering with large pharmaceutical and biotech firms makes small diagnostic shops attractive acquisition targets for large laboratories and instrument providers who also want a slice of big pharma's business.
[ pagebreak ]
After inking Rx/Dx deals with pharma companies, Monogram was acquired by Laboratory Corporation of America and DxS became a subsidiary of Qiagen. An acquisition by a larger shop typically increases the smaller shop's capabilities, allowing it to launch new tests and maybe even attract other pharma partners.
LabCorp has said that with Monogram under its aegis, it plans to advance into molecular diagnostics development in areas such as oncology and infectious diseases. And after partnering with Amgen and Bristol-Myers Squibb to develop companion tests for Vectibix and Erbitux, DxS was approached by AstraZeneca to commercialize a test for its non-small cell lung cancer drug Iressa.
However, there is a downside for Dx shops in these traditional Rx/Dx partnerships, and it's financial. Since drug makers view the role of the therapeutic as central to any Rx/Dx collaboration, the value proposition in most deals tends to tilt in favor of the pharma partner.
Speaking at a February conference in Philadelphia on drug/diagnostic partnerships in the field of personalized medicine, Stafford O'Kelly, head of molecular diagnostics at Abbott Molecular, said that "the therapeutic is the overwhelming driver of value in an Rx/Dx collaboration." For the diagnostic, "it's all about supporting the Phase III trial."
In O'Kelly's view, small diagnostics firms cannot afford to wait around for a pharmaceutical company to make their technology profitable — particularly in the face of evolving regulatory policies.
The FDA has indicated that in order for a genetic test to be mentioned by brand name in a drug label, the test must be approved by the agency. In addition, the FDA has expressed a desire to regulate higher-risk LDTs that were previously allowed to come to market with a laboratory certification from the Centers of Medicare & Medicaid Services.
"The risk of a diagnostic is the same as the risk of the drug," O'Kelly said. "If you're going to give a drug based on a test, a laboratory-developed test is not acceptable for companion diagnostic development."
The companion test "would require [pre-market approval] and 510(k) clearance," O'Kelly said, adding that without a nod from the FDA and without a patent-protected biomarker, "no one will pay for the test."
Of course, Abbott Molecular, with over $3 billion in diagnostic sales in 2009, is a big fish in the diagnostics space, and well positioned to be the kind of partner towards which big pharma seems to gravitate.
In the early part of the decade, Abbott navigated FDA's regulatory system to garner approval for one of the first companion diagnostics on the market, the Vysis PathVysion HER2 DNA Probe Kit that detects the HER2 gene and determines which patients should receive Herceptin. The label for Herceptin mentions the availability of the Vysis test, along with several other FDA-approved test kits for the indication. In addition to regulatory know-how, Abbott Molecular has its own sales force, and has access to global distribution channels.
These are exactly the qualities GlaxoSmithKline was looking for last year in seeking a diagnostics partner to develop the commercial version of a companion test for its investigational MAGE-A3 antigen in non-small cell lung cancer. Although GSK worked with a smaller Dx firm, Response Genetics, to screen patients in early clinical trials for the MAGE-A3 antigen, when it came to developing a commercial test, the drug maker went with Abbott, reasoning that the larger shop had more scale-up capabilities and regulatory experience (PGx Reporter 07-15-09).
After GSK inked an Rx/Dx partnership with Abbott Molecular, other drug firms recognized the advantages of teaming with a larger outfit. Pfizer, last September, announced it would work with Abbott to commercialize a genetic test that will determine which non-small cell lung cancer patients should receive Pfizer's investigational drug known as PF-02341066 (PGx Reporter 09-02-09).
Additionally, GSK this year extended its existing Rx/Dx partnership with Abbott to develop and commercialize a diagnostic test to select patients for an investigational skin cancer immunotherapy.
Value for Agnostics and Loners
At the same meeting on Rx/Dx partnerships in Philadelphia, representatives from Merck and AstraZeneca — drug firms that have inked several Rx/Dx partnerships and have publicly expressed a company-wide focus on the development of personalized medicines — itemized a few other must-haves for their diagnostic partners.
For example, a diagnostic partner must have an IP portfolio around the biomarkers of interest, a willingness to partner early, and the ability to meet pharma companies' timelines for conducting prospective and retrospective studies. As far as payment, most Dx developers receive a fee for service from their pharma partners, without any reach-through to drug sales.
With all these conditions for Dx partners, however, drug makers are keeping their own options open. Pharmas aren't ready to tie themselves to a favorite technology when it comes to companion diagnostic development, and at this point in the game, most aren't looking to buy a large instrument developer. And most drug developers aren't ready to develop the test themselves, as very few pharmas and biotechs have invested in launching diagnostics arms, among them Roche, Abbott and Novartis.
[ pagebreak ]
Pharmaceutical firms would not be interested in purchasing a company like Qiagen because they "want to remain flexible to gain access to different platforms," Cecilia Schott, business development director of personalized healthcare at AstraZeneca, said at the Philadelphia conference. As the personalized medicine field migrates from genotyping scans to sequencing, drug companies aren't looking to make any exclusive commitments to any one technology or Dx firm. It's better for pharmas to remain platform agnostic, Schott said.
AstraZeneca has an Rx/Dx partnership with Dako for companion test development in various types of cancer. The drug company also is working with DxS and several other labs in the UK to provide genetic testing for free to NSCLC patients on Iressa. And Myriad Genetics has said it has an Rx/Dx partnership with the drug developer (PGx Reporter 04-07-10; 03-10-10)
While big pharma shops around for the best platform, the prospects for profitability are pretty bleak for the small diagnostics developer on the other side of the Rx/Dx partnership, as there is often little guarantee that a biomarker discovery partnership with a drug firm will turn into a deal for commercial diagnostics development.
Even if it does, for an Rx/Dx combination product, the sales projections for the companion test don't fall anywhere in the same stratosphere as the sales figures of the drug. Iain Miller, senior director of oncology strategy and theranostics at BioMerieux, noted at the Philadelphia meeting that although DxS' KRAS test is projected to save the healthcare system $400 million annually, the typical market share for a companion test such as the KRAS test is expected to be around $10 million. This paltry amount will then have to be shared between multiple collaboration partners, Miller said.
DxS' KRAS test predicts which patients will respond to two colorectal cancer drugs, BMS/Merck Seronon/ImClone's Erbitux and Amgen's Vectibix. In 2009, Vectibix netted $233 million and Erbitux garnered €697 million ($948 million) in global annual sales.
While the numbers for companion diagnostic development are hardly enticing, the alternative is to embark on test development and commercialization alone. That far riskier scenario has only one real success story: Genomic Health's Oncotype DX. After launching Oncotype DX as a laboratory-developed test in 2004 — a test that predicts whether breast cancer patients' disease will recur and whether they will respond to tamoxifen — the company invested millions in conducting validating studies, educating physicians, and convincing payors.
As of December 2009, the company reported that 9,000 physicians had ordered more than 135,000 Oncotype DX tests in over 55 countries. The majority of payors in the US reimburse for the test, and the company reported $150 million in revenues for Oncotype DX last year.
In the end, Abbott's O'Kelly discouraged companies from attempting to develop a companion test without a pharma partner, since such an Rx/Dx collaboration will be instrumental in garnering a genetic testing recommendation from the FDA in the drug's labeling.
"The pharma company owns the label for the drug," O'Kelly said. "So, independent development of a companion diagnostic is unlikely to be successful unless it is associated with a significant safety issue."
A Third Option
The current value proposition for test developers in traditional Rx/Dx partnerships has Dx firms shopping for other options.
"The so-called success stories of Rx/Dx partnerships also have diminished returns," BioMerieux's Miller said. "Twelve years since [the launch of] Herceptin," often considered the first drug to be successfully marketed with a companion genetic test, "I don't think we are where we should be," he added. "I hope in [the next] 10 years we're not talking about the same thing."
This is where Medco hopes to play a role. By inviting diagnostic companies to partner with it, the PBM is offering a third option in the companion test development game.
Particularly in a scenario like that of Oncotype DX, where the diagnostic firm has decided to go for it without the help of a pharma, the test developer will have pay for clinical studies required by payors, patient recruitment, and related operational costs associated with validation of the test.
In such instances, Medco can "offer some value by partnering to garner the cost effectiveness and clinical utility data," Dechairo said. "We can run these studies in a real-world setting, where we see which drugs patients are on. We can contact those patients. We can get them enrolled without having to have academic sites, and monitoring, and contract research organizations involved.
"So, the costs to the diagnostic company to get the clinical utility and cost effectiveness data, by working with Medco, is often cheaper than them doing it on their own," Dechairo explained.
Medco recently presented the data from such a real-world observational study, which investigated the clinical utility of pharmacogenetic testing to predict the right dose for patients receiving the anticoagulant warfarin. In order to recruit study participants, Medco mined its pharmacy database to determine which patients were on warfarin throughout the country and contacted their physicians to gain permission to test their patients.
The PBM would not say how much it cost to conduct the nearly four-year study, which genotyped 900 patients and compared their outcomes to 2,688 patients dosed with warfarin using clinical data. However, Medco emphasized that its study methods are much cheaper than the traditional process of conducting clinical trials.
As a point of reference, last week researchers at the Washington University School of Medicine in St. Louis announced they had received $3.7 million from the National Heart, Lung, and Blood Institute for a five-year, randomized, double-blinded study, called the Genetics Informatics Trial of Warfarin, which will investigate the utility of genetic testing to dose the anticoagulant in 1,600 knee or hip replacement surgery patients at Barnes-Jewish Hospital and other institutions (PGx Reporter 04-12-10).
Last year, NHLBI launched the Clarification of Optimal Anticoagulation through Genetics trial, a prospective, randomized-controlled trial comparing clinical outcomes of one group of patients given warfarin based on just clinical factors and another cohort administered the anticoagulant based on clinical factors and genotype. The two-year, 1200-patient COAG trial received just under $10 million in fundng from NHLBI and the National Human Genome Research Institute (PGx Reporter 02-18-2009).
Of course, conducting the trial cheaply isn't the only hurdle for test developers looking to launch a new genetic test. As large as Medco's customer base may be, ultimately, doctors and experts in the field have to trust that the clinical data is robust in order to adopt the test more broadly.
[ pagebreak ]
Although Medco's warfarin trial found that genetic testing to dose the anticoagulant in heart patients reduced hospitalizations by 30 percent, the PBM has received some pushback from practicing physicians and academics, who have taken issue with the fact that the warfarin PGx study wasn't a randomized-controlled trial. Critics of the study have asserted that the reduction in hospitalization may be caused by the so-called Hawthorne effect — in which study subjects improve behavior simply in response to being studied and not as a result of the intervention (PGx Reporter 03-17-10).
However, Medco, along with others in industry and some at the FDA, are of the view that advances in pharmacogenomics require new clinical trial designs that accommodate a changing drug development paradigm in which traditional randomized-controlled trials are often unfeasible or unethical.
The observational warfarin PGx trial was Medco's attempt at presenting such an alternative that investigated the utility of an intervention outside of a clinical setting, in the real world. Ultimately, its collaborations with diagnostics firms around PGx tests also offer Medco a way of testing out new clinical trial models and challenging old paradigms.
Projects undertaken with diagnostic firms for pharmacogenomic tests will fall into Medco's personalized medicine program, which has already rolled out PGx testing services for widely prescribed drugs, such as warfarin and tamoxifen, to 200 customers serving 7 million people. However, as one of the largest PBMs in the US serving 60 million lives, Medco expects to eventually expand its personalized medicine program to more customers.
However, Dechairo emphasized that in order to form partnerships with a Dx shop, Medco would first need to see that the biomarkers in question are clinically valid.
"Our hurdles are around the clinical validity, not clinical utility, of the marker," he explained. Diagnostic developers "need to show us that if they're running the marker in their lab, that that lab is CLIA certified, that they've tested their results against some other methodologies, and that they've done discovery and validation studies to build a receiver-operator curve that's valid."
Medco doesn't want to talk to test developers who have a set of a hundred gene markers still in the discovery phase. Firms interested in approaching Medco with a partnership strategy better have their biomarkers "locked down," the platform for their test picked out, and the sensitivity and specificity of the technology validated in an independent population, Dechairo said.
But "once we feel that the biomarker is valid, we partner with the diagnostic company to do the clinical utility and cost effectiveness studies," he noted.
Celera is an example of the kind of diagnostic company Medco would be willing to partner with. Medco and Celera are working together to conduct a prospective, randomized study to evaluate whether the knowledge that patients harbor the KIF6 gene variant — which confers a 55 percent increased risk of coronary heart disease — makes them more likely to adhere to statin therapy (PGx Reporter 09-23-09).
Celera developed the KIF6 test without the help of a drug company and had validated it before approaching Medco. The company is offering it as an LDT through its subsidiary Berkeley HeartLab's CLIA-certified laboratory and several other labs, but the test has not been widely adopted. In collaborating with Medco, Celera is hoping to drive greater utilization of the test and give payors the clinical utility data necessary to begin paying for the test.
Another important requirement for Dx developers wishing to partner with Medco is biomarker IP. Industry observers agree that the only way for diagnostic company to truly improve the value proposition in an Rx/Dx partnership is to own the patent around the biomarker and the diagnostic methodologies.
"The IP definitely brings value to the diagnostic company, because if the data and the markers are in the public domain and the algorithm for using that is also in the public domain, and there is no IP around it, then any lab can run that test," Dechairo said.
"Medco has its own CLIA lab, so we can run that test without requiring a diagnostic partner," he pointed out. "But if a diagnostic partner has IP on a [proprietary] algorithm for how the test should be used, like in a methods patent, then they are the ones who own that … and therefore we would partner with them … because we can't do that on our own."
However, Dechairo disagrees with the view of Abbott's O'Kelly that FDA approval is critical to garnering investment for a companion diagnostic. In response to O'Kelly's comments at the Philadelphia conference, Dechairo countered that Medco would invest in advancing a companion test that was performed at a certified laboratory.
"Because we work by ourselves or with partners to generate that cost-effectiveness and clinical utility package before we bring this commercially to our clients, we are confident in the value of that test with our internal data," Dechairo told PGx Reporter. "That's why we don't require the extra hurdle of FDA approval. It's nice to have and we'd be happy to have that. But we're happy to work under the CLIA guidelines."
[ pagebreak ]
Finally, Medco claims it can offer the most value when it comes time for garnering reimbursement for the test. Since the clinical utility and cost effectiveness data on the test will be gathered within Medco's existing customer base, it won't be too far a leap to present this data to other insurers within its system and expand utilization of the test.
Also, by working within Medco's system, doctors performing these tests on patients won't need to depend on CPT codes for reimbursement. "Because Medco has a relationship with the client [who are payors], if we're facilitating the test or running the test with our diagnostic partner, we can charge our clients for the program or the number of tests they've done," Dechairo said. "We don't require a CPT code, because we're not going through the health plan."
By departing from traditional Rx/Dx partnerships and offering Dx shops a different value proposition than big pharma, Medco is hoping to hasten the incorporation of personalized medicine into healthcare. A survey Medco conducted with the American Medical Association revealed that only 12 percent of physicians were using diagnostic tests that were part of personalized medicine.
"When we implemented our programs for warfarin and tamoxifen we saw around 65 percent of physicians putting in a prescription for that diagnostic," Dechairo said. "We believe [that when] diagnostic companies partner with us, we can show the value to the patient, show the value to the physician, show the value to the payor, that might help get utilization for a valuable diagnostic that brings benefit to patients.
"Compared to the standard way to promoting a test by going to each physician, our way might be a better way to bring that benefit to patients."