BioWorld Today Correspondent

SAN DIEGO - It now is possible to design clinical trials so that accumulating data can be used to inform modification on an ongoing trial, leading to ineffective drugs being dropped at an earlier stage and to a greater chance of success in Phase III.

With the FDA due to issue guidance on the use of adaptive clinical trial design later this year, the concept is being embraced by all the large pharma companies.

Wyeth in particular is re-engineering its clinical development function around that principle. But it also is highly relevant to biotechs, which can use adaptive design to get better value out of their clinical development budgets, Judith Quinlan, vice president of Adaptive Design trials at Cytel Inc., told delegates in a session at the 2008 Biotechnology Industry Organization (BIO) meeting, titled "Adaptive Clinical Trials - an Evolutionary Concept."

A survey conducted in February 2008 found 59 examples of ongoing trials with adaptive designs. Of those, 20 are being conducted by biotechs.

"The aim is to maximize the amount of information you get for every $1 spent," Quinlan said. Rather than getting to the end of a traditional linear trial and discovering that wrong assumptions were made along the way about formulation, patient numbers or dosing, accumulating data are used to decide how to modify aspects of a trial as it goes along.

"The secret is to do that without undermining the statistical validity of the trial, and this means lots of planning up front," Quinlan said. "Adaptive design is a new methodology. . . . It is not an excuse to make any change you want in the course of a study."

At the start of a trial, there is a huge amount of uncertainty about many different factors, but even at Phase III, some uncertainties remain.

"In other words, adaptive design can be applied across development," Quinlan said.

In the early stages of development, biotechs likely will have a small number of compounds and will be trying to work out the best way to use limited funds to make the most of the portfolio. "At the stage they are living with a high level of uncertainty about the predictability of animal models, there will be the possibility of multiple indications," Quinlan said. "The goal of adaptive design in Phase I is to pick up the first signals of efficacy to transition to full development."

Until recently, Quinlan was director of biostatistics at London-based GlaxoSmithKline plc.

She told delegates that, from her perspective of buying biotech assets, adaptive design has the potential to increase the asking price for products that are being out-licensed. "If you have holes in your development plan which require a pharma partner to do re-work, it will weaken your position," she said.

At the same time, if companies plan to continue development in house, it will decrease the development risk being carried forward.

Adaptive design offers a "new approach to fundamentally better answers about safety and effectiveness of new products," said Michael Krams, assistant vice president of clinical development at Wyeth Inc. "We are enthusiastic pioneers in trying to implement this new paradigm in developing drugs."

Krams told delegates that in the past year Wyeth has saved $13 million on one trial budgeted at $20 million, saved $1.5 million on another and saved $3.6 million on a $12 million trial.

Adaptive trial design is a fast way of telling the difference between the good, the bad and the ugly, but it won't make drugs that don't work, work.

"We want to fail efficiently, and at the same time, take great care not throw out the baby with the bathwater. In other words, this is not just about getting rid of poor drugs sooner," Krams said.

However, the gains from adaptive design do not come for free, because it is necessary to spend more time up front modeling possible trial designs.

"You can't just change things at will. Adaptive designs are intensive on the preparation side looking at many examples of 'What do we do if this happens,'" said Don Berry, head of the quantitative sciences division at the MD Anderson Cancer Center at the University of Texas. "Calculating the operating characteristics of a trial is enormously complicated. You have to do it 10,000 times for each assumption."

MD Anderson Cancer Center has conducted 200 trials in the past five years. "Many biotechs come to us to ask about effective trial design," Berry said. As a sign of the way the wind is blowing, one CEO said the reason for seeking advice was that the lead investor was not prepared to back the trial unless it used an adaptive design.