Translating innovation into new approaches for applied and clinical research

Guest Columnist

Donald A. Berry,
Professor, The University of Texas MD Anderson Cancer Center

See additional guest columns »

Alexandria Summit: Oncology 2011 Leads to Neuroscience 2012

Clinical research differs in neurologic diseases as compared with cancer. Perhaps most notably, tumors can be seen and felt. They can be sliced, poisoned, and burned. And they announce themselves when and if they return. Neurologic diseases can be as devastating as cancer, but they don't have the same obvious physical presence. We don't say, "She had a large Parkinson's disease." Affecting the disease means affecting symptoms, and it's difficult to tell whether a drug effect is on the disease or only on the symptoms. Still, the commonalities between oncologic and neurologic research are greater than the differences. For example, heterogeneity within the diseases is almost as great as it is across them. In the extreme, different patients with the same diagnosis may have unique diseases that respond only to therapies individually tailored to them. Given the rapid progress in cancer biology and discoveries about cancer pathways, in a few years every cancer patient will have an orphan disease. The same may be true in neurology, but the time frame will be longer.

It is important for researchers to learn from successes and failures in diseases other than their specialty. Unfortunately, we continue to follow failed strategies too long. Success rates of phase III drug programs indicate that neurology fares better overall than oncology (55% compared to 34%), which is the worst among all therapeutic areas. However, neurology's rate is driven largely by successes in developing pain medications and therefore deceptively high. For example, the last 14 phase III drug trials in Alzheimer's disease have failed to show a treatment benefit. This dismal and durable record suggests that the foundations for launching phase III in Alzheimer's are flawed and that we haven't learned from our mistakes.

Such evidence of poor planning in phase II highlights the importance of the FDA's Critical Path Initiative with its focus on "biomarker development and streamlining clinical trials." The goal is to improve the foundation of drug development, which has not been the FDA's role historically. Similarly, Alexandria Summit 2012 will bring greater awareness to the problems in developing therapies for neurological diseases and potential solutions.

Characteristics of failures in late-stage drug development are similar in neurology and oncology. One culprit is sample size. Foundations are the strongest parts of buildings – and the weakest parts of drug development. The drug pyramid is unstable because it stands on its apex. Some sponsors roll the dice and enter phase III on the basis of inadequate evidence. Others base phase III on a modest-sized phase II that showed a benefit, but they aren't able to distinguish between a random fluctuation and a real drug effect. Regression to the mean sets in during phase III, which then evinces less benefit than expected – and perhaps no benefit at all.

A complication of this culprit is subset analysis. Naïve subset analysis is an epidemic in drug development. A patient subset that responds best in phase II becomes the focus in phase III. Regression to the mean almost always leads to disappointment. The rub is that personalized medicine and appropriately identifying responding populations is the future of drug research. Restricting to identification is what makes a subset analysis naïve. Confirmation is essential during phase II and perhaps continuing into the early stages of phase III.

A related culprit is the tradition in drug development that learning occurs only in discrete phases. Clinical trial designs that pay heed to results accumulating in real time make trials more informative and lead to more efficient drug development programs. Seamless phase II/III trials are examples, but learning can occur throughout a trial and not just at the juncture between phases. For example, the initial stages of phase III could continue a focus on identifying the appropriate patient subpopulation, drug dose, schedule, combination therapies, etc.

Another major reason for failures in drug development involves endpoints. Phase II sometimes focuses on early endpoints while phase III considers long-term outcomes. The relationship between them is based on hope rather than data. Affecting the manifestation of disease (e.g., tumor burden) may not translate into longer-term clinical benefits (e.g., prolonged survival). Having a primary endpoint is important, but we should consider the whole patient, modeling various measures of disease course longitudinally so as to learn whether and how the drug affects that course. In Alzheimer's disease, for example, ADAS-cog at an early post-treatment visit is not perfectly predictive of ADAS-cog at one year, but it has some predictive ability. And FDG-PET in the early post-treatment period has some ability to predict ADAS-cog in the longer term. These early "auxiliary endpoints" do not meet any formal definition of surrogacy, but they are critical in designing efficient clinical trials and in tailoring therapy to patients.

An aspect of traditional clinical trials that greatly slows progress is addressing one question per trial. Trialists and sponsors should collaborate on building trials that examine many possible therapies, including combinations of experimental drugs, aiming to identify which patients respond to which therapies. An ongoing prototype trial called I-SPY 2 in breast cancer incorporates adaptive randomization among many therapies, imaging assessments over time, baseline biomarkers and longitudinal biomarkers. All these aspects can be incorporated into designs of trials for Parkinson's and Alzheimer's as well as for other neurological diseases. The more therapies involved in such a trial, the greater the efficiency and the more rapid the progress.

We have an obligation to patients to make clinical research more efficient. This will be a double win because it is also essential for the economic well-being of our healthcare system.

See additional guest columns »

For more information on Alexandria Real Estate Equities, Inc., visit

Copyright © 2011, Alexandria Real Estate Equities, Inc. All rights reserved. Confidential and Proprietary - Do Not Copy or Distribute.
Alexandria and the Alexandria Logo are registered trademarks of Alexandria Real Estate Equities, Inc.
Terms of Use