Drug/Diagnostic Co-Development: When to Begin the Tandem Ride
Executive Summary
Targeted drug development raises a new conflict of interest possibility when a company hoping to market a drug is also trying to win approval for the diagnostic that proves the drug's efficacy. That problem of "self-validation" is one aspect of the even broader drug/diagnostic question: when and how a company should go about validating a biomarker it wants to incorporate onto a drug label.
Targeted drug development raises a new conflict of interest possibility: the company hoping to market a drug may also be trying to win approval for the diagnostic that proves the drug's efficacy.
That's what Pfizer Inc. realized when it decided to participate in the Alzheimer's Disease Neuroimaging Initiative (ADNI), a consortium of industry and academic participants whose goal is to validate imaging tools and markers for tracking Alzheimer's disease (AD) progression. "We now realize the importance of having other companies take some of the burden of getting the biomarkers peer-reviewed," to avoid any issue of self-validation, explains Peter Snyder, PhD, of Pfizer Global Research & Development, who chaired the ADNI study design group. "There's been a culture shift at Pfizer," in this regard, he says. Indeed, that motivation was sufficient to offset the usual industry concerns about a consortium approach, including having to share data and methods openly, and moving at a slower pace than what a company is accustomed to.
Pfizer's concern is one aspect of a broader drug/diagnostic question: when and how a company should go about validating a biomarker it wants to incorporate onto a drug label. In developing AD drugs, for example, the consistency between surrogate endpoints and other measures like neuropsychological or memory tests isn't well established. "Many biomarkers turn out not to be valid" during clinical trials, cautions Susan Molchan, MD, project officer at the National Institute on Aging (part of the National Institutes of Health ) for ADNI. For this reason, in advising ADNI, "the FDA has been very careful" to point out the need to go slowly to ensure validation, she points out. Put in the context of a pivotal AD clinical trial, it may take 7-10 years of follow-up to establish clinical utility of a new biomarker used in combination with a therapeutic to set dosage.
Aside from regulatory reasons, a pharma company should be concerned about investing in a Phase III trial without confidence in the clinical relevance of the target and the performance of the assay, adds Angus Hastie of SynergysDx, a division of Roche 's Roche Diagnostics unit. SynergysDx was formed to collaborate with pharma companies to take prototypical biomarkers identified outside Roche that are intended to be used in combination with the pharmas' compounds and turn them into optimized research-grade assays for clinical trials or ultimately, in vitro diagnostic products.
But the FDA is not clear on the subject, Hastie says. In an April 2005 Preliminary Concept Paper on Drug-Diagnostic Co-Development, it recommends that a diagnostic test be fully clinically validated at the end of Phase II. Yet even as it advocates this, the FDA acknowledges that the position does not reflect current practice. Indeed, since the publication of the Concept Paper, the agency has heard this message repeatedly from industry.
"What's depicted in the Concept Paper is an idealistic, optimistic scenario for how to do development better," in cases where a drug company has an idea about the target and is developing a biomarker to use in early-stage trials as a research tool to move trials forward, explains Felix Frueh, PhD, associate director for genomics at the FDA's Center for Drug Evaluation and Research. Unfortunately, he says, the diagnostic often comes into development as an afterthought.
The relevant question is when to assess clinical validity: whether a target can be measured and whether modulating it leads to a clinical effect. "It's reasonable to take that into consideration in Phase II" for drugs such as Genentech Inc. 's trastuzumab (Herceptin) and the EGF receptor (EGFr) drugs cetuximab (Erbitux) and erlotinib (Tarceva, also from Genentech), says Frueh, which from the beginning were intended as targeted therapies.
The commercialization of those drugs, while successful, nonetheless points up problematic co-development issues. "With Herceptin, the test used in clinical trials turned out not to be commercially feasible. It was only very late in the process that the firm Dako developed a test that could be co-approved." Indeed, he believes, "you could say [Genentech] got lucky" with Herceptin. On the other hand, he acknowledges, a development plan could also call for testing a drug against all comers, and to identify subpopulations in Phase III only if it fails the all-comers trial—the counter-strategy to having a molecular target and developing a test for it.
Developers of AD drugs already have some information to help guide their co-development programs. The main thrust of ADNI is to create a gold standard reference library with which to confirm or validate the sensitivity of potential markers that have already been identified as likely candidates, and so are already on the road to clinical validation. "We are not going in blind," says Pfizer's Snyder. "If we are pressed to use imaging markers for compound development programs in the next 1-3 years, we already know of some good ways to go."