Diagnostics Needed for Personalized Medicine

To fully utilize the great advances in understanding the genomic basis of disease, clinicians need medical diagnostics able to identify those individuals most likely to benefit from pharmaceutical therapy, and those at risk for adverse events. Unfortunately, only a handful of therapeutics—primarily cancer treatments—have been approved with validated companion diagnostics (CDx's), and only 10% of marketed drugs have pharmacogenomics information on their labeling, according to the Tufts Center for the Study of Drug Development. Murky regulatory and legal pathways for linking "next generation" diagnostics and break-through drugs appear the culprits in limiting progress in 'omics-based medicine.

A sign that this may be changing is the recent Food and Drug Administration approval of a test for the KRAS mutation linked to colorectal cancer to inform the use of the cancer drug Erbitux. FDA also approved in January 2012 Vertex Pharmaceutical's cystic fibrosis therapy Kalydeco (ivacaftor), after a speedy three-month assessment. The drug is effective for a few thousand patients tested for the G551D mutation, and carries a controversial $294,000 price tag.

Still, FDA's long-promised guidance on drug-diagnostic co-development remains on hold. Instead, the agency issued draft guidance in July 2011 on procedures for developing and approving CDx's linked to specific therapeutics. FDA officials are reviewing comments on that proposal, while contemplating new strategies for using genomics in drug development, as discussed at an Institute of Medicine (IOM) workshop last March. Experts from academia, industry, and FDA explored the incentives and challenges in collecting and validating complex genomics data needed to approve new diagnostics, with an eye to applying these techniques to serious conditions besides cancer.

Labs seek limits

One obstacle to clarifying FDA oversight of genomics diagnostics is the ongoing controversy over regulation of LDTs or lab-developed tests vs. commercial medical devices. An IOM report issued in March 2012 called for stronger FDA regulation of 'omics tests to ensure that the data and methods used in research studies are reliable, but that move is strenuously opposed by independent commercial test labs that conduct a growing volume of gene-based analyses. Under current law, test laboratories are certified by the Centers for Medicare and Medicaid Services under the Clinical Laboratory Improvement Act. FDA claims authority to regulate LDTs as medical devices, but currently does not exercise that power and primarily approves the test equipment used in labs.

The IOM addressed this issue following a scandal in 2010 involving genomics-based tests developed at Duke University that failed to accurately predict which chemotherapy would be most effective for cancer patients in oncology trials. Consequently, many individuals failed to receive appropriate treatment, and research reports had to be thrown out amidst charges of fraud and data manipulation. Still, commercial labs have gained support from Congress to block more active FDA oversight and other regulatory initiatives.

Testing companions

While the debate continues over LDTs, FDA is moving to clarify its process for approving companion diagnostics, as outlined in its July 2011 guidance. CDx's can assist clinicians in deciding what medication to prescribe, but are not intended to diagnose disease or to check status of patients under treatment, explained Alberto Gutierrez, Director of the Office of In-vitro Diagnostics (OIVD) in FDA's Center for Devices and Radiological Health, at the "Next Generation Dx Summit" in August. By predicting patients likely or unlikely to respond to treatment or to experience adverse reactions, CDx's are particularly useful in selecting appropriate patients for clinical trials. Elizabeth Mansfield, Director of OIVD's Personalized Medicine Staff, advised pharma sponsors at a CBI Forum in March 2012 to notify drug review divisions early that a diagnostic will be used in a drug trial so her office can provide advice to ensure that pivotal trials determine that the diagnostic is informative. If the therapeutic trial is successful, there usually is no need for further clinical validation of the diagnostic. The CDx will be labeled for use with that therapy, and the drug will be labeled as safe and effective only in patients identified by a certain diagnostic test.

Trial sponsors can run into trouble, though, if they use a diagnostic that has not been fully validated prior to use or if the study uses multiple diagnostics to enroll subjects in a trial, Gutierrez warned. If a test is part of inclusion/exclusion criteria in a trial, sponsors should use a central site to do all testing for trial enrollment, or use the same validated test at all sites. ­—Jill Wechsler