The on-going war over the proper role of regulation of molecular diagnostics got off to an early start in 2014 with the publication in the year’s first issue of Nature Biotechnology of a quixotic commentary: “A cancer trial scandal and its regulatory backlash” by Kurzrock et al.
These authors took aim at what they perceived as unnecessary fall out to the 2011 scandal at Duke involving use of faulty genomic markers to select therapy for cancer patients. In addition they took exception to a March 2012 Report published by the Institute of Medicine that followed this scandal recommending that (1) assay validation be accomplished by a defined multistep process and that (2) tests used for patient selection in clinical trials undergo FDA review.
Arguments against FDA regulation of laboratory tests (most commonly laboratory developed tests or LDTs) are not particularly new or newsworthy. But usually these are posited with the theory that CLIA controls are sufficient to meet regulatory needs. To my knowledge this is the first published piece that simultaneously argues for elimination of both CLIA and FDA controls over tests used for patient decision making.
To launch such an argument about CLIA in the ashes of the Duke debacle is simply mind numbing. Such an argument ignores the realities of the differences in control of the operating environment and the quality systems that typically exist between research and clinical laboratories. While there is no reason why a research laboratory cannot function as well as a clinical laboratory, without proper regulation that is not assured and hardly the norm. For research laboratories operating at a quality level equivalent to a CLIA certified laboratory, the additional time/expense to become properly credentialed are trivial in the scope of the cost of the entire developmental life cycle of a new diagnostic device. For research laboratories operating more loosely, laboratory studies can be performed and analyzed but not used for final patient decision making providing a safe harbor for exploration of new diagnostic information.
The concern with oversight of FDA in the investigational phase of diagnostic device development represents a fundamental misunderstanding of the objectives and processes involved in the investigation device exemption (IDE) program. Central to the IDE oversight process is determination that risks to the subjects are not out weighted by anticipated benefits and that the study is scientifically sound. For the sponsor that has done this right FDA has one of the shortest turnaround times for any review program: 30 days. For the sponsor that has not, there is ample opportunity to get it right by virtue of several types of interactive submissions including the pre-sub program, the pre-IDE program, and the IDE program.
Kurzrock et al. frame their argument against regulation in the context of excessive costs and lost time for new products to reach patients as a result of the need to comply with regulatory requirements. They argue that therapy directed by molecular diagnostics is likely to be more effective than therapy provided as a shotgun approach to unselected populations. One certainly hopes this is true since the success of personalized medicine rests on this assertion.
But they seem to reach the implicit conclusion that getting the test exactly right doesn’t really matter very much. I disagree and suggest the obverse; this argument speaks instead to the importance of having reliable tests not to winging it? For example, if a drug with significant benefits and/or risks is to be given to a patient on the basis of a companion laboratory test, it seems to me the test becomes an equal partner to the drug in contributing to the safety and effectiveness profile of the treatment. Kurzrock get it partially right when they offer the observation “a faulty set of lenses for patient allocation … is undesirable.” But in my view they missed the mark. It is more than undesirable, it is unacceptable and regulation can help.
Perhaps these authors read the wrong IOM report. They might do well to review the publication “Genome-Based Diagnostics: Clarifying Pathways to Clinical Use: Workshop Summary”.
Kurzrock et al. might be surprised to learn that in discussion with stakeholders it was noted “regulation of genomic diagnostic tests can be a critical factor in the extent of use of those tests and in the competitiveness of companies.” They might also be interested in knowing that some believe “standards, quality control, regulatory guidelines, and technology assessment all can facilitate the movement of a test from bench to bedside.” They might pause to consider the merits of the argument that “the lack of a coherent oversight system could create a chasm between the use of genomic diagnostic tests and improved health.”
Finally these authors may do well to look at the comments by Hayes in the IOM report and in his subsequent published paper on this subject: “Tumor-Biomarker Diagnostics – Breaking a Vicious Cycle”. As those following Myraqa blogs will know, Hayes makes the modest proposal that if diagnostic devices are ever to be properly appreciated and valued they will require more up-front attention not less. This is perhaps not the message Kurzrock et al. want to hear. But it is one too important to be missed.