NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US) Forum on Neuroscience and Nervous System Disorders. Neuroscience Biomarkers and Biosignatures: Converging Technologies, Emerging Partnerships, Workshop Summary. Washington (DC): National Academies Press (US); 2008.

Cover of Neuroscience Biomarkers and Biosignatures

Neuroscience Biomarkers and Biosignatures: Converging Technologies, Emerging Partnerships, Workshop Summary.

Show details

1Biomarker and Biosignature Principles

Workshop participants discussed the opportunities, challenges, principles, and best practices associated with identifying the necessary research tools, regulatory considerations, and partnerships for a biomarker that would provide a near-term impact. The Foundation for the National Institutes of Health (FNIH) Biomarkers Consortium, which catalyzes new partnerships for the development of biomarkers that lie within the precompetitive space, was identified by participants as one potential mechanism that may facilitate additional collaboration and investment. Further, participants highlighted other models of public-private partnership that seek to accelerate development of new therapeutics, spanning effort across multiple areas of biomarker development, including the Alzheimer’s Disease Neuroimaging Initiative (ADNI).


Organized by an independent planning committee, the Forum hosted a public workshop on biomarkers for nervous system diseases, inviting experts from industry, academia, government, and advocacy groups. The goal of the workshop was to discuss strategies to identify a high-impact biomarker, including a proof of concept, and provide a framework for how the Forum may facilitate its future dialogue and interactions among academia, government, and the private sector. Each speaker was asked to present data and stimulate discussion on the following questions:

  • What processes can be used to accelerate scientific advances relevant to biomarker development?
  • What models of public/private/academic partnerships have been successful in this and other arenas?
  • What disciplines should be brought to bear? How can interdisciplinary perspectives be promoted?
  • What tools are available? What tools are needed?

Besides stimulating discussion on these important topics, it was the Forum’s ambition to contribute specifically to accelerating the availability of at least one important nervous system biomarker, both for its intrinsic value and for its value in exploring modes of Forum engagement. Thus, a stated goal of the workshop was to identify at least one high-impact biomarker, suitable for public-private partnership and potentially accomplishable in the near term, whose development might be accelerated by the Forum by facilitating interactions among stakeholders. It was recognized that to be suitable for public-private partnership, a biomarker would need to be useful to therapeutic development in industry context.

The chairman of the workshop’s planning committee, Dr. Dennis Choi, described the workshop’s goals in greater detail. He underscored the importance of setting realistic expectations for biomarker development, considering that few biomarkers (of varying qualification levels) besides risk genes have been developed for nervous system diseases to date. He and other members of the planning committee noted that bio-markers that reflected disease activity, drug safety,1 or effectiveness were most likely to be of value in aiding clinical trials and, hence, to be of interest to industry. Yet a disease risk biomarker could also have quick impact on clinical trials—for example, if it helped to identify a subgroup of patients who had a higher probability of responding to a given candidate therapy. A 5-year time frame for development of a single biomarker is probably achievable, said Choi, but a longer time frame is needed for the biomarker to meet the Food and Drug Administration’s (FDA’s) rigorous regulatory requirements for qualifying as a surrogate marker (a topic later discussed by Dr. Janet Woodcock, deputy commissioner and chief medical officer at the FDA). While it is important to understand the regulatory requirements for a biomarker capable of serving as a primary end point in a clinical trial, Choi noted that exclusive focus on regulatory requirements at outset may deter innovation.

Another workshop goal was to help catalyze public-private partnerships. The partnerships could be cultivated through a variety of ways, including the FNIH Biomarkers Consortium, a new mechanism discussed by speaker Dr. Thomas Insel, the model provided by the ADNI, or, more narrowly, via a direct relationship between government and one private sponsor. Regardless, to be viable these partnerships must have sufficient commercial potential to engage the private sector and sufficient public health or research potential to engage the public sector.


To facilitate public-private partnerships for biomarker development, Dr. Insel, director of the National Institute of Mental Health, outlined one major new mechanism, the FNIH Biomarkers Consortium. The mandate of the consortium is to accelerate biomarker discovery, development, and qualification. The creation of this consortium was among the prime motivations behind this workshop. This workshop will also serve to inform the Biomarkers Consortium, according to Insel, who is both a member of the Institute of Medicine’s Forum and sits on governing bodies for the FNIH.

The Biomarkers Consortium was launched in October 2006 as a new initiative of the FNIH. The latter is a nonprofit organization associated with, yet independent of, the National Institutes of Health (NIH). FNIH is authorized by Congress to broker relationships between NIH and industry, academia, and philanthropies. Responding to scientifically worthy proposals, FNIH seeks funding from NIH institutes and pools their resources with those of private partners. One of the arrangements already created by this unusual pooling of public and private resources is an imaging project for lymphoma and lung cancer. NIH’s partners under the Foundation’s auspices are usually a group of pharmaceutical companies, rather than a single company. Projects vary in size, depending on their purpose and the Foundation’s success at fund-raising.

Unlike an NIH institute, FNIH does not operate from a fixed budget; rather, it solicits funds from its public or private partners depending on the proposals it selects. FNIH is the administrative headquarters for each of the projects and is responsible for the entire process of proposal solicitation, proposal review and selection, and post-award management. Proposals can be submitted by any researcher and need not be restricted to those affiliated with organizations holding membership on the Foundation. Above all, FNIH’s selection process is independent of the typical NIH peer-review system, and its criteria for selection are in keeping with its mission to expedite and expand the development of medically useful biomarker technologies and products.

The Biomarker Consortium’s policy, like that of its parent Foundation, is only to solicit projects within the so-called precompetitive space. The concept is based on the premise that precompetitive projects are unattractive to academic, government, and industry partners alike, although for different reasons. For academicians and government researchers, developing biomarkers is too expensive and process oriented. For a single drug or device maker, biomarker development is too risky and removed from commercial payoff to justify the investment. The Consortium fills the gap by funding precompetitive projects that none of these entities would undertake on their own. Emphasis on precompetitive projects ensures that results are widely useful to the field as a whole. As elaborated upon by Choi, precompetitive projects sit somewhere between being not too hard to accomplish in a reasonable time frame and not so easy that individual companies can accomplish them and thereby gain competitive advantage.

One prominent example already funded through FNIH is the Alzheimer’s initiative to identify biomarkers. It is precompetitive because it does not test any particular drug; rather, it is a prospective observational study that tracks the course of Alzheimer’s disease. By contrast, a competitive project is one in which an individual company stands to gain financially, such as by testing a particular medication or diagnostic test for Alzheimer’s disease. There are other mechanisms that enable those types of partnerships to occur outside the purview of the Consortium. Nevertheless, the Consortium is still so new that it has not defined the exact boundaries between precompetitive and competitive space, according to Insel.

The Biomarkers Consortium faces many of the same policy issues as its parent Foundation. However, one key issue, conflict of interest, is less relevant to the Biomarkers Consortium because all members have agreed that all work will be public and, as described above, in the precompetitive space. In addition, device manufacturers and diagnostic companies are not represented in the Consortium, eliminating those who may see the biomarkers discovered as within their intellectual property space and of considerable value. Finally, all members sign extensive disclosure documents and agree to transparency in their interactions. Another issue concerns project solicitation and selection. Here, the obvious criteria apply: scientific merit, responsiveness, feasibility, and quality. But the most difficult issue is to find what types of scientifically meritorious projects are best suited to the Consortium’s mission. The ongoing debate is whether scientific merit should be defined as having clinical impact as its foremost objective. A project can be of tremendous scientific value, for example, without having immediate clinical impact. On the other hand, a project can have immediate clinical impact without being at the cutting edge of science. Consortium members who sit on the committees that approve proposals often wrestle with these tensions.

The Consortium gained from the policy already developed by its parent Foundation regarding antitrust laws. Those laws normally preclude leaders of the pharmaceutical industry from meeting together and working on joint projects. Once the Foundation defined its role in facilitating projects in the precompetitive space, industry representatives were willing to participate without fear of violating antitrust laws. Two other thorny issues—intellectual property and data sharing—still are formidable because of the trade-off between encouraging commercialization, on the one hand, and meeting the public health need for transparency and openness on the other. For that reason, these issues are worked out on a project-by-project basis and are subject to the approval of the Foundation’s oversight bodies.


FDA is deeply concerned about the limited innovation of bio-markers, stated Dr. Janet Woodcock, deputy commissioner and chief medical officer of the FDA. That concern prompted several policy initiatives, such as publication in March 2006 of FDA’s Critical Path Opportunities Report. The document overtly encourages development of biomarkers and other tools to shorten the time necessary for new drug and device development and their clinical use (U.S. Department of Health and Human Services, 2006). In keeping with that landmark publication, Woodcock said that her presentation was designed to clear up misunderstandings about FDA’s definitions of biomarkers and to explain the agency’s regulatory requirements for different types of biomarkers. The misunderstandings, in her view, have set back biomarker development, because FDA’s requirements for most types of biomarkers are erroneously perceived as too onerous. With the exception of surrogate biomarkers, most other types of biomarkers do not require a high bar for regulatory use, she stated.

FDA became concerned about the lag in biomarker development, relative to a surge in therapeutic development, when it realized that many biomarkers are discovered but never submitted for regulatory review. They are developed in academic laboratories and published in the biomedical literature as case series. They may even become commercially available as a lab service. But few are integrated into widespread clinical care because the evidence base is too slim or controversial. The main hurdles, according to Woodcock, are that academicians do not understand FDA’s requirements and that the business model for diagnostic development is not as robust as that for therapeutics.

To clear up some of the misunderstanding, Woodcock began by giving FDA’s current regulatory definition of a biomarker. The following definition was developed by an NIH-convened working group of which FDA was a part (Biomarkers Definitions Working Group, 2001):

A biomarker is a characteristic that is objectively measured and evaluated as an indicator of normal biologic processes, pathogenic processes, or pharmacological responses to a therapeutic intervention.

As it would for any other evolving field of medicine, FDA further modified the definition in its pharmacogenomics Guidance into possible, probable, and known valid categories of biomarker, depending on scientific evidence available to support the biomarker (Food and Drug Administration, March 2005).

Meanwhile, FDA has backed away from using the regulatory term “validation” of biomarkers because the term acted as a deterrent to bio-marker development. To signal the lower threshold of evidence needed for most biomarkers (except surrogate markers), FDA began to refer to its regulatory process of evaluating biomarkers as a qualification process rather than a validation process. The purpose of a qualification process is “to evaluate the utility of a biomarker” (U.S. Department of Health and Human Services, 2006). Biomarker qualification is essentially an evaluation of a marker’s fitness for use, that is, whether the evidence supports a biomarker’s use for a given purpose. The level of evidence needed to qualify for fitness for use is highly variable. Many types of diagnostic biomarkers, for example, do not have a high threshold of evidence for approval by FDA. Biomarkers used in drug development that do not have an extremely high threshold of evidence, said Woodcock, include those used for safety assessment (e.g., markers that predict early signs of toxicity and/or signal potential for severe toxicity) (U.S. Department of Health and Human Services, 2006). Additionally, genetic tests for drug metabolizing enzymes or other determinants of starting dose may be utilized without undergoing the rigorous “validation” required for surrogate endpoints. Similarly, biomarkers used to stratify patients in order to enrich a trial with those who should receive therapy (e.g., as is the case with Herceptin for breast cancer) can usually be studied within a particular drug development program and do not need extensive separate trials.

A surrogate endpoint requires the most rigorous level of evidence. It is defined as a biomarker intended to substitute for a clinical endpoint. The surrogate is expected to predict clinical benefit (or harm, or lack of benefit) based on epidemiologic, therapeutic, pathophysiologic, or other scientific evidence. The clinical endpoint for which the surrogate is being developed is a characteristic or variable that reflects how a patient feels, functions, or survives (U.S. Department of Health and Human Services, 2006).

Woodcock proceeded to describe the multistep process that FDA requires for qualification of biomarkers. Qualification of any biomarker first requires analytic validation, a process that includes evaluation of test parameters such as stability of reagents, standardization of assays, assessment of sensitivity, specificity and predictive value of assays, and the biomarker’s robustness in various sites. Analytic validation is the area where academia falls short, in part because academic scientists are not well compensated or rewarded in the academic sector for applied science. The second step is clinical validation, which includes evaluating the performance of the biomarker in clinical samples or in people with varying characteristics. The sponsor must establish that the assay continues to measure the same thing with reasonable accuracy under varying conditions or in different populations. The third step is to establish clinical utility, that is, to show that the biomarker has some clinical significance. Establishing clinical utility is not very onerous because many stand-alone diagnostic biomarkers can meet this criterion. Biomarkers that show ways to stratify patients based on prognosis, that show natural history of the disease process, or that predict pharmacokinetics based on variation in drug-metabolizing enzymes are the most common ways to establish clinical utility. A somewhat higher bar is reserved for bio-markers used to diagnose, or contribute to diagnosis of, pathology. In some cases, the purpose of biomarker qualification is to establish a linkage between the biomarker and a therapeutic intervention. These types of biomarkers are used to select patients to receive therapy (or not) or used for dose selection. These types of biomarkers do not have extensive regulatory requirements for clinical utility.

Woodcock stressed the point that biomarkers falling under any of the aforementioned categories may shorten the duration of clinical trials. A surrogate marker, in other words, is not the only type of biomarker that can hasten the process of drug development. Examples of nonsurrogate markers that can shorten trial duration are ones used to enrich trials with patients whose prognosis is worse or patients who are likely to exhibit a more rapid time to an event. For example, enrollment criteria often restrict entry to patients who meet certain prognostic criteria. New bio-markers such as gene expression arrays in cancer or markers of inflammation in heart disease may be used to identify individuals at high risk for recurrence or myocardial infarction, respectively.

Qualification of surrogate biomarkers requires great rigor, including evidence showing biological plausibility, statistical correlation with a clinical outcome, and success in clinical trials (Box 1-1). Although these criteria seem formidable, FDA has accumulated more than a decade of experience with surrogate markers. Over that time, its position on what constitutes a surrogate marker has evolved. Among FDA’s key modifications, said Woodcock, is to understand that there is no gold standard for clinical outcome measurement of a particular disease. Patient outcomes are too multidimensional in that a single outcome measure can miss domains of interest. It is very difficult to capture with a single measure both the benefit and the harm predicted by a surrogate endpoint. Woodcock said she foresees the future of surrogate endpoint development as featuring composite outcome measurements (i.e., biosignatures). She also envisions responder rather than population mean analyses and individualized therapy based on biomarker-derived strata.

BOX 1-1

Qualification of Biomarkers for Use as Surrogate

Biological Plausibility

  • Epidemiological evidence that marker is a risk factor
  • Marker must be consistent with pathophysiology
  • Marker must be on causal pathway
  • Changes in marker reflect changes in prognosis

Statistical Criteria

  • Changes in marker must be correlated with clinical outcome (but correlation does not equal causation)

Additional Support for Biomarkers as Surrogate

Success in Clinical Trials

  • Effect on surrogate has predicted outcome with other drugs of same pharmacologic class
  • Effect on surrogate has predicted outcome for drugs in several pharmacologic classes

Other Benefit/Risk Considerations

  • Serious or life-threatening illness with no alternative therapy
  • Large safety database
  • Short-term use
  • Difficulty in studying clinical end points

SOURCE: Temple, 1999.

Regarding applications to nervous system disorders, Woodcock expressed optimism. Few nervous system biomarkers are available today because the disorders are marked by subjective diagnostic criteria, highly variable rates of responses, a high need for preventive interventions, and current therapeutic interventions with safety or adherence problems. All of these features create an opening for the development of new biomarkers. Woodcock views neuroscience as a leading candidate for new biomarker development.

To facilitate biomarker development, FDA has carved out several roles for itself. Through its Critical Path Opportunities Report initiative and its “qualification” process, it hopes to encourage adoption of new biomarkers for preclinical and clinical product development. It also hopes to encourage partnerships and consortia to share the burden among all stakeholders who benefit from the new biomarkers. FDA participates in at least five other consortia dealing with other fields of medicine. Finally, FDA plans to develop regulatory guidance on pathways to market, and it plans to promulgate further advice on the design of qualification trials.


ADNI is an approximately $60 million public-private partnership sponsored by FNIH in collaboration with other federal agencies, the National Institute on Aging (NIA), and private companies and organizations. Its overall goal, over a 5-year period, is to develop a validated biomarker for Alzheimer’s disease clinical trials. The emphasis of the initiative is to find biomarkers through neuroimaging, as the name of the initiative implies. Serial magnetic resonance imaging and positron emission tomography scans are being used to image several parameters of the brain, including the volume and boundaries of the hippocampus and the entorhinal cortex (two sites most affected by Alzheimer’s disease), whole-brain atrophy, and cortical thickness. But other biomarkers from cerebrospinal fluid and urine are also being collected that as described later, offer great interest and potential. The study’s costs are being borne by NIH ($40 million) and the Alzheimer’s Association and several drug companies ($27 million) according to Dr. William Potter, vice president, Franchise Integrator Neuroscience, at Merck Research Laboratories, the presenter who described the initiative and the lessons drawn thus far.

The study seeks to identify biomarkers for the progression of mild cognitive impairment and early Alzheimer’s disease. It is not a clinical trial testing a particular drug; rather, it is a prospective, naturalistic study tracking several groups of patients over the course of 2 to 3 years. One group (n = 400) has pre-Alzheimer’s (i.e., a mild cognitive impairment found in the earliest stages of Alzheimer’s disease). These patients constitute a critical population for prevention or for slowing further progression. Another group has early onset Alzheimer’s disease (n = 200), and the control group consists of cognitively normal older adults (n = 200).

Although ADNI commenced before creation of the Consortium, it illustrates the kind of public-private partnership envisioned by the Consortium for other nervous system diseases. ADNI will furnish one large dataset for analyzing a host of potential biomarkers or biosignatures over the course of disease, with the goal of determining the most useful ones. As such, it is a precompetitive project, with broad applicability for eventual use in wide-ranging clinical trials. In the past, clinical trials of the same agent had been plagued by inconsistent or conflicting findings partially attributable to the use of different biomarkers or different methods of analysis. ADNI’s sponsors have agreed to full data sharing on a real-time basis; ongoing results are publicly accessible on the Internet.

What are the motivations behind the public and private partners? For the primary public partner, the NIA, the trial serves a crucial public health need to find a biomarker to stimulate development of new therapeutics. And that public health need is growing due to the demographic bulge of aging baby boomers. For industry partners, explained Potter, there are several reasons:

  • the greater commercial demand due to the demographic growth in older persons;
  • the longer period of patient usage based on the expectation that once a biomarker is identified, it will be possible to treat patients earlier in the course of disease;
  • the desire to expedite NIA’s research findings so that they can be applied to clinical trials more quickly;
  • the ability to participate in study design; and
  • the inexperience of industry partners in conducting prospective observational studies in which no drugs are being tested (as opposed to their wealth of experience in clinical trials).

On the downside, investment in Alzheimer’s drugs is still considered highly risky because of the exorbitant costs of drug development, including clinical trials, and the poor return on investment. Several drugs have been unsuccessful in reaching the market, whereas other drugs that have reached the market have not met with high demand because of modest clinical gain.

The major lessons learned, observed Potter, come from the active role that industry members have carved out through a special advisory committee. Their contributions have been essential in two prominent areas of study design: the collection and standardization of cerebrospinal fluid (CSF) and the standardization of imaging (Box 1-2). Industry members have collaborated so well, in the opinion of Potter and other workshop attendees, that they set a model for industry contributions to future Biomarkers Consortium projects.

Box Icon

BOX 1-2

Active Role of Industry’s Participation in Study Design. Raised the percentage of patients for CSF collection from 20 percent to close to 60 percent Standardized collection, handling, and storage of CSF through development of a best practices (more...)

While many advances are coming from the analysis of the image data, Potter expressed that collection of CSF may turn out to be more important than imaging biomarkers. This contribution was made possible through the advisory committee’s success in working with its public partner to modify the study design to collect CSF at much higher rates than initially called for in the protocol and by encouraging more patients, via an educational video, to willingly undergo lumbar puncture. The protocol originally called for 20 percent of each group to undergo lumbar puncture; however, through the aforementioned strategies, investigators have been successful in collecting CSF from close to 60 percent of participants. With this success also comes the challenge of analysis of the CSF. ADNI was not originally designed to perform detailed analysis on this quantity of CSF; therefore, according to Potter, new partnerships are necessary to take advantage of this opportunity.



A safety biomarker can be used to identify patients at high risk for serious side effects, to monitor early signs of toxicity, or to predict the likelihood for severe toxicity.

Copyright © 2008, National Academy of Sciences.
Bookshelf ID: NBK53116
PubReader format: click here to try


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.5M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...