Use of biomarkers in early stage clinical trials has grown considerably in the past few years.

• Lilly's Roberts, who has responsibility for trials through Phase II, says 80 to 90 percent of early stage clinical studies have some form of biomarker measurement to support the elucidation of drug activity, safety, and efficacy parameters.
• Pfizer's Williams also puts the figure for his company at 80 to 90 percent. He says the exceptions are situations when efficacy trials are simple and inexpensive to conduct, or when traditional tests are so good that biomarkers are not expected to substantially improve decision making.
• At Bayer, Carney reports that biomarkers are included in all programs at some level. Since oncology is a top priority, 90 to 100 percent of these clinical trials include biomarkers.

Areas of Impact

Various clinical areas are expected to move toward adoption of biomarkers at different rates.

In addition, oncology has a strong pipeline of drugs in development and has a number of smaller companies engaged in innovative research. Already, physicians have access to tests that help identify breast cancer patients at low risk for recurrence, who may not benefit from systemic adjuvant therapy (see "Breast Cancer: The Test Case" ). In another area of cancer, Novartis' successor to Gleevec (imatinib) is in Phase II/III and appears to be less susceptible to the development of drug resistance. This drug is being developed with a battery of tests to define which patients should receive it.

In Roberts' opinion, more and more new drugs in oncology are being pursued with parallel development of a diagnostic test. She also expects to see more changes in drug labels associated with genotyping of drug-metabolizing enzymes in the future.

Cardiovascular At first glance, cardiovascular may not look like a good area for biomarker development. The cardiovascular field is complex, and includes a number of overlapping diseases. Also, the area has been well served by simple, low-cost tests, such as blood pressure and cholesterol measurements. Nonetheless, some gains in individualization of treatment for cardiovascular patients are expected.

To Vonderscher, the "one size fits all" paradigm of treatment encourages physicians to practice trial-and-error medicine to find the best prescription for an individual patient. But this sort of physician-level segmentation means that a company often loses more patients than it gains if its new drug is effective only in a subset of a patient population; it will be more practical to actively segment using a biomarker.

And though blood pressure and cholesterol measurements have been effective, Pfizer's Williams points out that they don't address issues such as plaque stability and size. He predicts that the clinical biomarkers in cardiovascular will include intravascular and carotid ultrasound, and in vivo tests for plaque composition and stability using imaging. Carney predicts that biomarker tests to guide new treatments for cardiovascular disease will be available over the next five years.

Roberts envisions the cardiovascular area as progressing, since key academic leaders are pushing for more and better tests, including those encompassing metabolic syndrome, a constellation of health-oriented changes affecting cardiovascular, diabetic, and obese patients.

Neuroscience In Lester's view, and that of Vonderscher, Dracopoli, and Edmonds, biomarker improvements in neuroscience will come first in several difficult-to-treat diseases where the potential for gain is enormous, including Alzheimer's, schizophrenia, and pain.

The biology of these diseases is not well understood, and the phenotype is complex. This hampers the biomarker discovery and development process, but there is promise in Alzheimer's to assess the benefit of treatment using imaging and other biomarker techniques. Roberts expects complex diseases, such as schizophrenia and depression, to be most difficult, with biomarker applications developing at a later date.