NEW YORK (GenomeWeb) – Despite all the consternation about the reimbursement and regulatory challenges facing the field of personalized medicine, getting the science right is still the primary challenge for drug and diagnostics companies interviewed by Tufts University.
The Tufts Center for the Study of Drug Development (CSDD) issued a report (see here) in which leading drugmakers developing personalized treatments said that currently 73 percent of cancer compounds are being studied in the context of biomarker data; while 42 percent of compounds across all indications employ biomarkers.
Moreover, the drug and testing firms Tufts interviewed said they expected to increase investment in personalized medicine by 33 percent in the next five years, and projected that drug development in the space will grow by 69 percent over that time.
"Since 2010, groups have bought into this idea of personalized medicine and really gotten behind that, and it's demonstrated in the industry in terms of the investment in the space, with dollars and resources," said Glen Ferguson, VP of corporate development at HealthTell, a diagnostic company marketing ImmunoSignature, a technology that gauges an individual's immune system response to specific diseases.
"The contract research organizations are jumping into the space and there are a lot of service and diagnostics companies coming up to support all that," added Ferguson, who was not involved in the Tufts survey.
However, in his view the report also captures the fact that industry is increasingly challenged in its development efforts, having to contend with more data about diseases and biomarkers. "The more we're learning, the better our tools are becoming, [and] the more we're understanding the complexity of the science related to disease and the human body," said Ferguson.
Five years ago, when Tufts interviewed the 21 top-ranking pharma companies about their work in the space, the firms said that between 12 percent and 50 percent of their pipelines involved personalized medicines, and said they had grown their investment in developing molecularly targeted drugs by more than 70 percent between 2006 and 2010.
That investment is bearing fruit as there are now many more drug/test combinations on the market. The US Food and Drug Administration has approved seven treatments co-developed with a companion test since the agency approved the personalized medicine poster child Herceptin (trastuzumab) along with Dako's HercepTest in 1998. Additionally, 14 FDA-approved drugs have companion diagnostics with which doctors can identify best responders, and Tufts estimates there are several dozen more therapies that may be individualized with the help of non-FDA approved tests.
Additionally, the Tufts report shows a steady uptick in the proportion of marketed drugs with pharmacogenomic information in their labels, from 10 percent in 2010 to 13 percent today. Joshua Cohen, research associate professor at Tufts' CSDD, pointed out that in 2001, less than 1 percent of drugs (most of which had been developed before the completion of the Human Genome Project) had PGx data in the label. "So, we see an increase of almost 1 percent every year," Cohen told GenomeWeb in an email. "That is significant."
For the latest report, Tufts surveyed 15 leading drug companies advancing personalized therapeutics, and interviewed 12 diagnostics shops and 10 pharmas involved in the field. "The inclusion of diagnostics firms represented the biggest change from the 2010 study," Cohen said.
While personalized therapies still comprise a small slice of drug developers' overall revenues, the Tufts report reflects that industry players are more willing to recognize and acknowledge they need diagnostics to identify the patients that will respond to their drugs. "Diagnostics firms are crucial for personalized medicine," Cohen said. "Theyre sort of like the forgotten sibling who's been quiet for quite some time, yet integral to the functioning of the family (in this case, personalized medicine)."
In the oncology setting, where the Tufts report shows the most personalized medicine activity, drugmakers have become particularly comfortable incorporating diagnostics as part of the drug development process. For example, Roche subsidiary Genentech is using an internally developed immunohistochemistry test to identify patients with PD-L1 expression in all its lung cancer trials for the immunotherapy MPDL3280A. Moreover, given the number of pharma companies advancing immunotherapies with companion diagnostics, drug firms for the first time have agreed to work together to explore the differences between the tests they're using in therapeutic development and publish a report to educate the public.
Back in 2010, surveyed pharma companies projected their investment in personalized treatments would increase by 50 percent in the coming five years. According to the Tufts report, they've exceeded that mark, having increased their average investment in the field by 87 percent in the last five years.
"This said, firms do not expect to continue to almost double their investments every five years," Cohen noted. The drug companies Tufts interviewed expected a 33 percent increase in investment from 2016 to 2020. In the early to mid-2000s, pharma commitment to personalized medicine was relatively low, but as interest grew, there were big jumps in investments in the latter part of the decade. "That baseline is much higher now," Cohen reflected.
Pharma willingness to invest in the field has remained relatively steady despite certain unremitting barriers. In Tufts 2010 report, drugmakers seemed focused on challenges around aligning co-development of the drug and the test. However, the surveyed companies in the latest report ranked "partnering with diagnostics firms" as less of a worry after scientific challenges, regulation, and reimbursement issues.
The availability of lab networks has made it easier for diagnostic shops to standardize companion tests, and greater availability of next-generation sequencing panels has opened up the possibility of developing universal companion tests that could be used alongside multiple drugs. Moreover, the FDA introduced a draft guidance of CDx development in 2011 and finalized it last year. The document lays out the agency's thinking on when drugmakers should pursue a companion test in drug development efforts.
Additionally, since drugmakers seem willing to eventually move single-analyte companion tests to multi-marker NGS platforms, industry players are also encouraged to see the FDA ironing out the types of data it will accept when it comes to complex diagnostics. For example, in clearing Illumina's MiSeqDx cystic fibrosis tests, FDA looked at the clinical performance of the markers in those tests against a Johns Hopkins University database, as well as considered the NGS platform's analytical performance using a human reference genome developed by NIST.
Although drug and test makers are becoming more comfortable working together on drug development programs, they still ranked regulation and reimbursement as equally challenging to their personalized medicine efforts. The FDA last year introduced its controversial draft guidance for regulating lab-developed tests (LDTs), and the lab community appears steadfast in its resolve to fight the FDA from implementing this oversight plan. Since LDTs are readily used to identify best responders to drugs (even when they have an approved CDx), any change in this regard stands to impact drug companies as well.
Moreover, the "Protecting Access to Medicare Act of 2014" (PAMA), signed into law last year, will bring big changes on the reimbursement front once implemented. PAMA will move payment for molecular tests to a market-based payment system starting in 2017. The Centers for Medicare & Medicaid Services will pay labs for tests according to the weighted median of private payor rates, and labs will have to begin reporting rates from payors to CMS in 2016.
However, CMS is working through the MolDx program to help test developers garner the clinical utility for their tests using coverage with evidence development mechanisms and registries. According to Tufts' survey, more than 70 percent of the interviewed companies perform specially designed clinical trials to prove the clinical utility of their personalized medicine products to payors and 50 percent operate evidence-gathering programs.
"I believe we'll continue to see more work done in the holy grail of personalized medicine: co-development of drug/diagnostic combinations," according to Cohen, who sees payors and regulators favoring co-development given its promise to lower costs and improve success rates within pipelines.
Although industry players are more vocal about the regulatory and reimbursement challenges, their biggest hurdle remains the science. This is not a surprise, since as the field's understanding of disease mechanisms grows, so does the amount and diversity of the data drugmakers must consider when crafting treatments.
"The science has always been complex, both in terms of biomarker identification and also test and drug development," Cohen said. "Part of the complexity is due to multiple factors impacting a patient's susceptibility to adverse events or proneness to being a better responder to a drug than another patient. There may be multiple biomarkers and other factors involved. Finding out the weight attached to each factor is very difficult."
Alzheimer's is a key example, where the complexity of the disease has continued to baffle researchers and defeat drug development programs focused on amyloid as the key biomarker. Some drugmakers, such as Biogen, are now hoping to do better with the help of advanced imaging tools that may enable researchers to identify patients at the point in their disease when they'll benefit most from an amyloid-targeting agent.
Amy Miller, executive VP of the Personalized Medicine Coalition, which commissioned the latest report conducted by Tufts, recalled a conversation she once had with a bench scientist who had worked on identifying a predictive biomarker for the AstraZeneca's lung cancer drug Iressa (gefitinib). AstraZeneca had spent a lot of money searching for a predictive biomarker for the drug and researchers had many false starts before the field landed on EGFR mutations.
This scientist described the whole experience like a "who done it," Miller recalls. "I think that is why the companies are so confounded when regulatory and payment policies do not support targeted therapies," she said. "The science is so hard and the benefit to patients so striking that the rest should naturally follow, in the eyes of the developers."