Personalized medicine should be adopted as a prospective strategy in drug development rather than primarily serving as a tool to salvage failing drug candidates, participants at FDA's Science Forum suggested.

Personalized medicine is being positioned as the future of drug development - a holy grail that will pair the right drug at the right dose with an individual patient.

HHS Secretary Michael Leavitt has lauded personalized medicine's potential to provide preventive, predictive and preemptive treatments. He recently announced that making personalized medicine a reality will be his focus for the remainder of the Bush Administration.

Personalized Medicine: So Far, So Little

To date, however, the tools of personalized medicine have been primarily used to resuscitate troubled drug candidates.

The most cited example of how a personalized medicine approach has been applied as an afterthought to rescue a drug is AstraZeneca's Iressa (gefitinib).

Research indicated that the non-small cell lung cancer treatment was effective in a subset of patients. Based on a Phase II study aimed at identifying responders, Iressa was allowed to remain on the market with a narrowed indication despite its failure to demonstrate a survival benefit in a confirmatory trial (¹ Also see "Iressa Gets Narrowed Indication Instead Of Withdrawal: New Model For FDA?" - Pink Sheet, 27 June, 2005., p. 18).

Another example is NitroMed's heart failure treatment Bidil (isosorbide/hydralazine), which was approved for use in African-Americans after previous trials failed to show benefit in the overall population. Using race as a surrogate for genomic markers, FDA approved Bidil in June 2005 (² Also see "BiDil Approval In Blacks Is Step Toward Personalized Medicine, FDA Says" - Pink Sheet, 27 June, 2005., p. 5).

One notable success of a personalized medicine strategy is Genentech's cancer therapy Herceptin (trastuzumab), which is targeted for over-expression of HER-2. Herceptin exemplifies the way that "individualized treatment" can be attained not just by targeting the patient's genomic profile, but also by targeting the profile of the disease.

Although it has not been widely embraced, prospective use of the personalized medicine approach is now within reach, CDER Deputy Director Douglas Throckmorton suggested April 19 at the Science Forum.

"Personalized medicine is enabled by the new science and the technology. We just simply didn't have them years ago. Those things give us the opportunity to take the methods we had and apply them to opportunities currently as well as developing ones."

"We are not in a place where personalized medicine is hype," he maintained. "This is the future."

AstraZeneca Learning From Past Experience?

Some companies are taking steps towards actualizing personalized medicine. AstraZeneca has adopted such an approach for its clinical development programs, Global Product Director-Oncology Catherine Wheeler said. However, she noted that the tools may not yet be able to support a personalized medicine strategy for all products.

"We now have much more interest within the company in developing an R&D infrastructure that really investigates the ability of personalized medicine for all projects in the portfolio."

"We are looking to find personalized medicine approaches for just about everything that we do," Wheeler said.

"Within the company we now have much more of an emphasis on the linkage of the target and the disease so that rather than seeing ourselves as a chemistry company that is trying to make molecules and get them out there, we really are beginning to understand the biology."

Personalized medicine may not be a realistic goal for every development program, however, Wheeler noted. "What is really challenging is that it may not work, or it may not be evident in the case of every drug that you are trying to develop. Trying to get that linkage between the target and the physiology and the outcome is extremely challenging."

Wheeler added that AstraZeneca has begun engaging with FDA to address issues surrounding translational science, biomarker development and trial support for biologic samples to facilitate its personalized medicine approach.

"I think we are right up against what are the limits of what science can actually do, and I think we are all trying very hard to make it work and to be able to realize this vision, but to expect it to be working for every single product may be asking more than what is reasonable in the short term," she said.

Throckmorton agreed that a personalized medicine approach may not be appropriate for every drug candidate.

"I think you look at personalized medicine as the right tool under many circumstances, and you use it under those circumstances. There are other times, simple drug development products, places where there is not a lot of inter-individual variability in things, where those sorts of tools are not necessarily required in efficient drug development," he said.

"What you need to do is make sure that it is a part of your discussion, try to decide whether or not those sorts of techniques...need to be applied or not, and then try to facilitate their use when appropriate."

Where Opportunity Lies

Throckmorton suggested that there are four areas of opportunity in the field of personalized medicine "that are either being done today and we can do better, or that I believe the tools exist at the present that could make it more usable today."

The four areas "we should be talking about trying to implement today" are: drug metabolism, better definitions of disease subsets, prevention of adverse events and response monitoring.

Developing drugs with a greater understanding of their metabolic variability will provide a better understanding of variability of response in trials, enable better dosing based on knowledge of an individual's metabolism and enhance the commercial feasibility of drug candidates, Throckmorton said.

For metabolically complex products, a pharmacogenetic understanding of individuals' metabolic responses would help tailor the dose to individual patients.

For example, metabolism of warfarin is very heterogeneous, Throckmorton said. "If we could do a better job of predicting that for individuals, we could likely do a better job of dosing that product with obviously a big safety payback."

"We're planning a clinical trial to try to assess the impact of such testing to get better labeling, better pharmacogenetic information into the labeling."

FDA recently announced a Critical Path Institute initiative to validate genomic biomarkers to optimize warfarin dosing (see ³ Also see "Warfarin Dosing Biomarkers Could Be Model For Genetic Tests In Labeling" - Pink Sheet, 8 May, 2006.).

While FDA has increasingly included pharmacogenetic dosing information in labeling, Throckmorton suggested that the agency could be doing more. "This is an opportunity we've been working at - one that we could be doing better at with available tools."

An understanding of drug metabolism and dosing based on personal characteristics creates development opportunities for metabolically complex drug candidates, Throckmorton asserted.

"It could potentially allow complex products, products that are now avoided by sponsors, complex metabolic pathways, to actually come to market because they would be able to define the...metabolic profile [and] the proper dosing."

Another aspect of personalized medicine that Throckmorton maintains can be realized today is identification of disease subsets.

"New tools can provide that discrimination by providing additional information for us on pathologic process and disease mechanism to give us insight that we simply haven't had available to us in the past," he said. Having that information could help to improve clinical trial design.

Developing biomarkers to identify likely responders to therapy or patients with high likelihood of an event has been a high priority for FDA under its Critical Path initiative.

"Cancer is the place where a lot of the initial progress has been made. Other disorders I think are waiting to have this applied to them," Throckmorton said.

He noted heart failure as another area where work is underway to identify disease subsets. "Finding a population that is at high risk for a new myocardial infarction or high risk for having to be hospitalized for heart failure or something like that is another way of making trial design more efficient, another way of making us better at developing novel therapeutics."

Determining the genetic basis for adverse events is also currently attainable by tying adverse events to genetic data. Throckmorton stated that "we have the tools at hand at present to study...the prevention of adverse events."

"Rare and serious adverse events occur in postmarketing, and if there is a way that we could find to marry their genetic information in a bank and somehow look back and make an association with the occurrence of that adverse event and changes in their genetic makeup, understanding of their genetics, we would be able to prevent that adverse event from occurring or at least describe it."

"Racial variance in adverse events may be a good place to start," he added.

FDA is currently requesting genetic material from patients who have experienced adverse events related to torsade de pointes to "predict the risk of torsade in products that are known to carry that risk in large populations" and try "to understand better the genetic basis of that risk."

Drug developers also have the capacity today to "do a better job of predicting non- or poor responders," Throckmorton said. "We have the tools...if we can find the population whose efficacy is waning for a given product and look to see what's happening with their genetics, to see if there was way to predict the loss of efficacy."

Where FDA Can Help

The onus lies not just with industry, Throckmorton acknowledged. "We need to fundamentally reengineer the way we're thinking about product development. FDA needs to take the regulatory steps to make that all possible; sponsors need to be open and willing to make changes where possible to make this happen as well."

Throckmorton highlighted three regulatory challenges that FDA needs to address to facilitate the development and clinical use of personalized medicines.

Focus On Diagnostics

First, he said, the agency needs to support the commercial availability of diagnostics that provide the genetic data necessary for clinical decisions. "It's critical for us to be able to move to personalized medicine to have tests available so that an individual patient can make an individual decision," Throckmorton said, and "co-development of the diagnostic and therapeutic are possibly as critical."

However, co-development of products and companion diagnostics poses both regulatory and business hurdles.

FDA is working on a guidance on drug-diagnostic co-development (⁴ Also see "Biomarkers To Get Individual Guidances Under FDA Plan To Promote Development" - Pink Sheet, 17 October, 2005., p. 20). Partner diagnostics are also getting significant attention through the agency's Critical Path initiative.

The C-Path Institute is also taking action to facilitate diagnostic development. The Tucson, Ariz.-based group plans to host a meeting with diagnostic companies to discuss ways to improve pathways for this type of research tool.

Moving Forward With Adaptive Trial Design

In addition to promoting the availability of diagnostics, FDA also needs to become more comfortable with adaptive trial designs, Throckmorton said.

"We need to overcome an inertia around the discussion of the adoption of novel trial designs wherever possible. That inertia...exists both within the sponsors and within the regulators, and we just simply need to be able to confront that," he said.

"FDA's response up to now has been to try to encourage earlier conversation around product development with sponsors; earlier ideally being at the end-of-Phase IIa meeting or pre-IND phase where both sides are still open to novel trial designs, able to make different choices about the way a product might be developed."

FDA has indicated that it will take a case-by-case approach to accepting development programs with innovative trial designs (⁵ Also see "FDA To Consider Critical Path Trial Design Proposals On Case-By-Case Basis" - Pink Sheet, 10 April, 2006., p. 12). The agency's Critical Path opportunities list highlighted streamlining clinical trials - including adaptive, enrichment and non-inferiority designs - as a key to improving the drug development process.

The agency also needs to continue to promote biomarker development to gain a "better understanding of the mechanistic basis of individual response to therapy to improve patient information and physician information," Throckmorton said.

"We know that biomarker development is unfocused and somewhat inefficient. In the future, I think we hope to link development and biomarker development very closely so that we can make the most efficient use of resources and time."

FDA is currently involved in several biomarker projects, such as the multi-party Oncology Biomarker Qualification Initiative (⁶ Also see "Tri-Agency Oncology Biomarker Initiative Commences With PET Scan Study" - Pink Sheet, 20 February, 2006., p. 3).

Personalized Medicine Is Closer Than You Think, FDA’s Throckmorton Says

More from Archive

More from Pink Sheet