Search the Site
join our mailing listsubscriber login
|
News Analysis: Trials Went On Despite FDA Statements
publication date: Feb 2, 2012
From the Jan. 27 issue of The Cancer Letter:
By Keith Baggerly, MD Anderson Cancer Center
In 2007 and 2008, Duke initiated three clinical trials in which genomic “signatures” of sensitivity were used to determine patient allocation to treatment arms. Duke has acknowledged that the three now-terminated clinical trials were conducted without FDA approvals in the form of investigational device exemptions (IDEs), despite the fact that FDA views such signatures as medical devices for which approvals must be obtained before they are used in “significant risk” situations. Working with Kevin Coombes, a fellow biostatistician at MD Anderson, we raised questions about the science underlying how these signatures were derived—and concerns regarding patient safety—and published objections in September of 2009. After our analysis was published, Duke suspended the trials pending an independent review board-sponsored investigation of the science in October 2009; announced the trials would be reopening for enrollment in January 2010; and then re-suspended the trials following new concerns about the CV of one of the PIs, Anil Potti, in July 2010; before finally terminating the trials in November 2010. It is now acknowledged that the underlying science was flawed, and that the trials should not have been run. Related to the scientific question, however, is a regulatory one linked to patient safety: what types of approval are required to run a trial in which signatures are used to direct therapy? In the context of oncology trials, where therapies can be dangerous, both FDA and IRB approval may be required. As noted above, the FDA views genomic signatures used to guide therapy as medical devices, which would thus require IDEs to be used in an experimental setting. When the clinical trial is also subject to the need for approval of an investigational new drug (IND) application, then significant risk device issues can be addressed either through the IND application or through a coordinated IDE application. In either case, however, qualitatively similar questions need to be addressed. Which approval is viewed as primary also determines which center within the FDA has direct oversight: the Center for Devices and Radiological Health (CDRH) for devices, or the Center for Drug Evaluation and Research (CDER) for drugs. In the Duke trials, the institution’s IRB initially determined that FDA approvals would not be required because all treatment arms were seen as offering “standard of care,” and therefore didn’t pose “significant risk,” which is the threshold for such determination. However, two treatments can work equally well in the general population, and thus qualify as “standard of care”, while having different odds of working in an individual patient. Indeed, the Duke signatures were introduced in hopes of exploiting this distinction. By intelligently choosing the “better” therapy, they hoped to skew the odds from the 50:50 that might apply to each drug overall to something like 75:25. |
