3
FDA Perspectives

The Food and Drug Administration (FDA) must balance competing needs in its approach to evaluating biomarkers, and it must do so with limited resources (IOM, 2010). On one hand, biomarkers are viewed as a route to reducing the cost and time required to develop effective new drugs, devices, and biologics to address chronic diseases; on the other hand, patients and consumers must be protected from risks associated with biomarker use. The apparent efficiency/safety dichotomy also applies to the use of biomarkers in health claims for foods, which the FDA regulates under a different framework than is applied to drugs, devices, and biologics.

These complexities were apparent in workshop presentations and discussion reflecting the FDA’s perspective on the report. Speakers representing three FDA centers involved in using biomarkers in regulatory decision making—Center for Food Safety and Applied Nutrition (CFSAN), Center for Drug Evaluation and Research (CDER), and Center for Devices and Radiological Health (CDRH)—were asked by the committee to describe the process their centers currently employ to evaluate biomarkers, to compare that process with the committee’s recommended evaluation framework, and to consider how the FDA might make use of their report. To the last question, Marc Walton replied that at the time of the workshop, the report had only recently been published, and would require detailed reading by several people at each FDA center involved with biomarker evaluation before the agency could respond to it. “We’re clearly very interested in the report and are going to be thinking about it



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 29
3 FDA Perspectives The Food and Drug Administration (FDA) must balance compet- ing needs in its approach to evaluating biomarkers, and it must do so with limited resources (IOM, 2010). On one hand, biomarkers are viewed as a route to reducing the cost and time required to develop effective new drugs, devices, and biologics to address chronic diseases; on the other hand, patients and consumers must be protected from risks associ- ated with biomarker use. The apparent efficiency/safety dichotomy also applies to the use of biomarkers in health claims for foods, which the FDA regulates under a different framework than is applied to drugs, devices, and biologics. These complexities were apparent in workshop presentations and discussion reflecting the FDA’s perspective on the report. Speakers rep- resenting three FDA centers involved in using biomarkers in regula- tory decision making—Center for Food Safety and Applied Nutrition (CFSAN), Center for Drug Evaluation and Research (CDER), and Center for Devices and Radiological Health (CDRH)—were asked by the com- mittee to describe the process their centers currently employ to evaluate biomarkers, to compare that process with the committee’s recommended evaluation framework, and to consider how the FDA might make use of their report. To the last question, Marc Walton replied that at the time of the workshop, the report had only recently been published, and would require detailed reading by several people at each FDA center involved with biomarker evaluation before the agency could respond to it. “We’re clearly very interested in the report and are going to be thinking about it 2

OCR for page 29
0 BIOMARKER AND SURROGATE ENDPOINT EVALUATION very carefully,” he said. “It’s clearly a very extensive, detailed, effortful report … so there will be a lot that we want to think about.” Discussion following the presentations focused on two main top- ics. The first was the definition of the term risk biomarker, as used in the two presentations from representatives of CFSAN and employed in its charge to the committee. As previously noted, the committee did not use this term, due to its potential for confusion with the precisely defined term risk factor. Elizabeth Yetley, a consultant to the National Institutes of Health (NIH) Office of Dietary Supplements and to the committee, asked whether CFSAN considered a risk biomarker to be a type of sur- rogate disease biomarker. Speaker Paula Trumbo, supervisory biologist at CFSAN, replied, “risk biomarkers can be surrogate endpoints, but not all risk biomarkers are surrogate endpoints.” This exchange prompted Thomas Fleming to ask whether CFSAN’s goal was to be able to determine whether a given biomarker is a reliable way to assess the level of risk a patient has for a certain event, or whether they seek biomarkers that can represent reliably whether a treatment will alter a patient’s risk for such an event, and can therefore serve as a sur- rogate endpoint. “The former [case] simply requires a biomarker to be a correlate,” he said, but the latter requires the effect on the biomarker to reliably predict the full effect on the true clinical endpoint. Dr. Fleming further explored this distinction in his presentation, which is summarized in Chapter 7 in this volume. A second topic of discussion focused on whether a surrogate endpoint must have biological plausibility: that is, that the biological connection between the biomarker and the process it represents is known, if not fully elucidated. Stephen Williams noted that, in an article he and John A. Wagner had written recently reviewing the history of surrogate end - points, they had argued against requiring biological plausibility for sur- rogate endpoints (Lathia et al., 2009). They reasoned that highly plausible biomarkers had failed as surrogate endpoints (for example, ventricular arrhythmia, as previously discussed by Ronald Krauss), and also that this criterion would exclude some types of biomarkers (for example, microRNA) for which plausibility might be difficult or impossible to determine. However, Dr. Williams added, an editorial in the same issue of the journal in which their article was published disputed their conclusion (Gobburu, 2009). Therefore, he asked the FDA representatives how they viewed the role of biological plausibility as a criterion for selecting sur- rogate endpoints. The committee’s report did not specifically recommend that surrogate endpoints have biological plausibility, Dr. Williams noted. “From my point of view, plausibility is always helpful,” Dr. Walton stated. “For one, it provides enough enthusiasm to expend the significant

OCR for page 29
1 FDA PERSPECTIVES amount of energy and effort it takes to prove [a surrogate] endpoint,” he said. “If it seems to be just magic, it can be hard to sustain the effort over time.” He also said that thinking about the apparent mechanism underly - ing a biological process helps researchers plan experiments or studies to test the endpoint in question. From a theoretical perspective, Dr. Walton agreed that a biomarker with an unknown link to a biological process, but with robust supporting data, could have utility as a surrogate endpoint; however, he noted that it might not be clear whether this endpoint reflected a late, critical stage in the disease process, or an early step in the response to an intervention. Moreover, he said, if a biomarker’s biological significance is known, that information might suggest further applications for the biomarker, leading to the testing of additional hypotheses and the possible expansion of its use. Without a mechanistic understanding of the biomarker’s biological role, its use would be restricted to a specific, data-constrained context. “It really is important to do the best we can to objectively understand the data and formulate hypotheses,” Dr. Fleming said. According to Dr. Fleming, biological plausibility is particularly critical to formulating a hypothesis regarding a putative surrogate endpoint; thereafter, prospec - tive trials are needed to validate the effect of an intervention on the bio - marker as compared to a clinical endpoint. “Now, in that validation trial, I will also have had prespecified biological mechanisms that will be key supportive endpoints, and those can, in fact, be more readily interpreted as confirmatory,” he explained, “yet I’ll explore the data even in those trials … but those additional exploratory analyses of biological mecha - nism have to be viewed with great caution.” Committee member Victor De Gruttola, professor and chair of the Department of Biostatistics at Harvard School of Public Health, noted that the association between a given surrogate and a clinical endpoint may be confounded. Given that possibility, unless the underlying biological mechanism is understood, “there isn’t really a statistical way to be sure that confounding has been appropriately dealt with,” he said. “Surrogacy analyses, especially the ones that attempt to ascertain whether the Pren - tice condition holds or not, will always be subject to confounding.” There- fore, he said that clinical trials must be conducted on surrogate endpoints in order to assess causality and determine the impact of a randomized treatment on an endpoint. PRESENTATION By PAuLA TRuMBO, CFSAN Dr. Trumbo reviewed the current process for biomarker evaluation at CFSAN, which occurs in the course of reviewing petitions for autho - rized and qualified health claims for foods. Scientific evidence offered in

OCR for page 29
2 BIOMARKER AND SURROGATE ENDPOINT EVALUATION support of such claims often includes academic studies employing “risk biomarkers [that] are not validated surrogate endpoints for chronic dis - ease risk,” she reported. A health claim is a causal relationship between a substance, which can be a food or a food component, and a disease or a health-related condition, Dr. Trumbo said. Such claims apply to healthy people in the U.S. population who are free of (but possibly at high risk for) the disease that is subject to the health claim. For example, she said, a person who has elevated low-density lipoprotein cholesterol (LDL-C) levels could be included in an evaluation of a heart disease health claim. Health claims can also be crafted to address a particular subpopulation such as women or the elderly, she added. Health claims are not about treating, preventing, curing, or mitigating symptoms of a disease, Dr. Trumbo said, which differentiates them from drug claims. “Certainly, there are drugs that are available for reducing modifiable risk factors, i.e., surrogate endpoints,” she said. “But when we evaluate the evidence for a health claim, we’re either looking at a sur- rogate endpoint or a clinical disease endpoint … we won’t be looking at evidence that is about mitigation of a disease … [such as] progression of an existing cancer or joint pain from osteoarthritis.” The statutory laws governing the regulation of foods and drugs are very different, said Dr. Trumbo. Provided they are free of pathogens, foods are generally recog- nized as safe (GRAS), and CFSAN focuses its premarket safety evalua - tions on determining whether ingredients that are added to foods or to dietary supplements are safe. CFSAN evaluates health claims according to an evidence-based review system that has been in place for about 5 years, according to Dr. Trumbo. This system establishes scientifically rigorous criteria for two categories of health claims—authorized health claims and qualified health claims. Dr. Trumbo noted that what differentiates these categories is the strength and quantity of the evidence supporting them. A qualified health claim must be presented to consumers using qualifying language because the evidence behind the claim is not as strong as that supporting an autho- rized health claim, which does not require qualifying language. Evidence for health claims includes intervention studies, clinical stud- ies, and observational studies. Dr. Trumbo noted that CFSAN considers intervention studies to be their gold standard, particularly those that mea- sure a clinical endpoint. More often, however, CFSAN must evaluate evi- dence for health claims based on observational studies involving surrogate endpoints. “For instance, one of our qualified health claims is on monoun- saturated fats and olive oil, and another one … [is for] walnuts and coronary heart disease,” both of which utilized LDL-C data to support the claims. In evaluating the strength of evidence presented to support health

OCR for page 29
 FDA PERSPECTIVES claims, CFSAN considers the type, size, and quality of studies conducted and their relevance to the general or target population of the claim, said Dr. Trumbo. In addition to surrogate endpoint biomarkers, CFSAN also encounters biomarkers of intake (for example, beta-carotene and other carotenoids) as part of health claims. In this case, she said, “if needed, we will do side reviews to evaluate whether there is good evidence to suggest that an intake biomarker is a good reflection of intake.” Dr. Trumbo presented the following list of disease endpoints CFSAN has been petitioned to review, coupled with associated biomarkers: • Coronary heart disease (CHD)—Total/LDL cholesterol, blood pressure • Colon/rectal cancer—polyps • Diabetes—Blood sugar levels, insulin resistance • Osteoporosis—Bone mineral density • Dementia—Mild cognitive impairment None of these endpoints has been determined to be a surrogate endpoint by CFSAN, she said. “We don’t have the expertise to make that decision, so … we rely on the various institutes within the National Institutes of Health, as well as our colleagues at CDER” to evaluate surrogate endpoints and keep CFSAN informed of new developments in this area. However, CFSAN came to wonder whether they could apply similar criteria to the qualification of biomarkers as they do to the health claim review process. This question prompted CFSAN to approach the Institute of Medicine (IOM), and is the basis for their request for a biomarker evaluation frame- work, said Dr. Trumbo. PRESENTATION By kATHLEEN ELLWOOD, CFSAN Kathleen Ellwood, director of nutrition programs staff in the Office of Nutrition, Labeling, and Dietary Supplements in CFSAN, further outlined the challenges that led to CFSAN’s request for the IOM study. She noted that there are few known “validated modifiable risk biomarkers for chronic disease risk,” and “not all risk biomarkers are surrogate endpoints.” In particular, she said, many chronic diseases lack surrogate endpoints. Filling this gap will not be easy if it requires reliance on costly long- term clinical trials using clinical outcomes, said Dr. Ellwood. Therefore, the use of “biomarkers and biomarkers of disease risk” is important to CFSAN, especially in the chronic disease setting. However, there is “an absence of an agreed-upon, systematic, and transparent process for quali- fying surrogate endpoints,” and a need for research to identify potential surrogate endpoints and the scientific evidence needed to support their

OCR for page 29
 BIOMARKER AND SURROGATE ENDPOINT EVALUATION use. “We need to identify these information gaps, including the type of additional studies, [and] the level of evidence that’s needed to qualify a risk biomarker,” she said. Dr. Ellwood noted that a 2007 IOM report funded by CDER rec- ommended that government agencies and other stakeholders develop “a transparent process to create well-defined consensus standards and guidelines for biomarker development, validation, qualification, and use” (IOM, 2007a). As a result, Dr. Ellwood and her colleagues at CFSAN worked extensively with their counterparts at CDER to develop the request for this study, which was funded through the FDA’s Critical Path Initiative.1 To achieve their goal of developing “a framework on eviden- tiary standards for chronic disease risk biomarkers,” CFSAN asked the IOM to recommend “a process, focusing on risk biomarkers and surrogate endpoints in chronic disease, and … asked [the IOM] to look at existing prototypes.” CFSAN requested that the IOM committee demonstrate application of the biomarker evaluation framework through case studies, as reflected in the committee charge. Dr. Ellwood noted that they wanted to focus the scope of the case studies to biomarkers and surrogate endpoints for CHD, a relatively well-researched area. For example, she noted that “LDL cholesterol is considered a surrogate endpoint, but HDL-C [high-density lipoprotein cholesterol] has not quite gotten to that level. So what would it take for HDL, then, to be considered a surrogate endpoint for a chronic disease?” CFSAN’s objective in requesting that the IOM recommend a framework and demonstrate its use through case studies was “to assist FDA in developing a framework for the evidentiary standards for the qualification of biomarkers as potential surrogate endpoints of chronic disease risk.” PRESENTATION By MARC WALTON, CDER CDER’s interest in biomarkers rests primarily in their ability to improve drug development and to improve CDER review decision mak- ing. “If all works well, then we’ll wind up with more drugs that get successfully developed, and those that do get developed may be better optimized than without the biomarkers,” said Dr. Walton. Within this general purpose, he identified the following ways in which biomarkers could help optimize drug development: 1 The Critical Path Initiative is a strategy, launched by the FDA in 2004, to drive innovation in the scientific processes through which FDA-regulated products are developed, evalu - ated, and manufactured (http://www.fda.gov/ScienceResearch/SpecialTopics/CriticalPath Initiative/ucm076689.htm).

OCR for page 29
 FDA PERSPECTIVES • Selection of candidates for development • Study design during development • Marketing application review • Labeling information • Enabling otherwise infeasible studies to be conducted (for example, surrogate endpoints) As a result of the importance of biomarkers within CDER, Dr. Walton noted that the center is working actively with various groups to aid them as they try to develop the biomarkers and the data to support their use. To date, biomarkers have become accepted for use within CDER through a two-stage process, according to Dr. Walton. Initially, a bio- marker is proposed for use within a specific drug application, when a sponsor asks to use a particular biomarker in the drug approval process. The sponsor will discuss with CDER the data supporting their request, and then CDER makes a case-specific decision as to whether the biomark- er’s use is appropriate. “As these cases occur time and again in slightly different situations, over an extended period of time” CDER gains experi- ence with the biomarkers, including a broader concept of the breadth of their use that is reliable and informative,” Dr. Walton said. Based on that deeper experience, he said, CDER “will come to accept biomarkers in a broader sense.” CDER continues to refine and expand biomarker qualification along two main tracks, Dr. Walton continued. In some cases, biomarker tests will need to be codeveloped to evaluate a biomarker along with the drug whose use will be tied to that biomarker test result, he said; guidance for that process is being developed in collaboration with CDRH. In addition, Dr. Walton noted that CDER has established the Biomarker Qualification Process, which has been in development and slowly maturing over the past couple of years. CDER is currently developing a guidance for this program. “Within CDER, when we talk about biomarker qualification, we are talking about a conclusion that within some very carefully and specifically and clearly stated context of use, this biomarker has been demonstrated to be able to support appropriate decision making,” Dr. Walton stated. Once a biomarker is qualified, CDER is careful that it is being used within this specific context, and that new evidence has not since come to light that contradicts its qualification. As long as that is the case, drug developers can employ the biomarker without extensive or detailed submission of primary data to support its qualification, said Dr. Walton. “Drug develop- ers will be able to rely upon that [biomarker] qualification, and the CDER reviewers will also be able to rely upon it.” When CDER qualifies a biomarker, “what we qualify is the substance,

OCR for page 29
 BIOMARKER AND SURROGATE ENDPOINT EVALUATION or the analyte, that is … being measured,” Dr. Walton said. “It is not the method that measures it, nor is it the device that measures the biomarker.” Rather, CDER expects that a robust biomarker would have several differ- ent assay methods capable of measuring it appropriately, along with infor- mation on the performance characteristics of each of these assay methods. In its capacity to review devices for clinical testing, CDRH would clear or approve some of these assay methods, he said, but such clearance is not equivalent to CDER qualification of the biomarker being tested. Dr. Walton characterized “context of use” as a “shorthand term for a comprehensive statement of the manner and purpose of use of the bio - marker.” He noted that context of use might, for example, be restricted to certain disorders or drug classes, or for nonclinical biomarkers to certain animal species. Context of use is also likely to specify how clinical sam- ples must be obtained, he continued, although sample handling methods are considered to be part of the assay (and as such, guided by CDRH). “Very importantly, how the results get interpreted and applied must be clearly stated within the context of use,” Dr. Walton stated. For CDER, “a context of use defines the boundaries of the known reliability of the biomarker,” he said. “It’s not going to at all define the boundaries of non- reliability, because there are going to be a great many potential uses that have not [yet] been evaluated.” Those uses are candidates for case-by-case considerations, he added, and if successful, for expanding the qualified context of use for that biomarker. After reviewing the multistep biomarker qualification process cur- rently practiced at CDER (see Box 3-1), Dr. Walton noted that the guidance being developed for this process will be incorporated into a more general guidance for “drug development tools (DDTs).” These include patient- reported outcome tools as well as biomarkers, he said, because their development and evaluation processes are quite similar. The guidance will describe the process of working with CDER to develop DDTs, explain CDER’s managed process for consistent scientific evaluation of these tools, and establish the legal framework for CDER to post its qualification decisions themselves as guidance, he said. “It is important to us that this [guidance] becomes available for public access,” he said. In addition, as with all guidances that FDA publishes, this allows for public comment regarding specific biomarkers and CDER’s evaluation process. Regarding the committee’s recommended biomarker evaluation framework, Dr. Walton noted that, in some elements, the conceptual orga- nization of the process is somewhat different from that used at CDER. He emphasized the importance of context of use, which he said must be clearly identified and specified in order for CDER to determine whether or not a biomarker is determined to be appropriate for use (qualified in CDER’s terminology). He added that the analytical method should be

OCR for page 29
 FDA PERSPECTIVES BOX 3-1 CDER Process of Biomarker Qualification • ubmission of a short letter requesting initiation of formal interactions with S Biomarker Qualification Program - ubject to CDER acceptance of request S • Advice and consultation stage begins • Submission of a briefing document - dentification of biomarker I - ntended context of use I - xisting state of knowledge E - dentified knowledge gaps I - roposed approach to fill the gaps P • Meeting with Biomarker Qualification Review Team - iscussion, and advice on plans D • epeated interactions as needed by sponsor to complete work of develop- R ing comprehensive evidence on biomarker for specific intended use • Meeting with Biomarker Qualification Review team when data complete - Comprehensive evidence overview assessment - greement that all reasonably apparent knowledge gaps eliminated A - vidence supportive of qualification for the intended context of use E • Qualification review stage begins • ubmission of full, comprehensive, detailed evidence package with primary S data • Biomarker Qualification Review team evaluates data, writes review(s) - akes recommendations on qualification decision M - epeated interactions as needed by biomarker qualification reviewers R - ublic advisory committee or other public consultation if warranted P • Letter of qualification sent to sponsor when CDER-wide concurrence • Qualification letter and reviews to be posted on CDER website - fter CDER Process Guidance published A SOURCE: Walton presentation (June 21, 2010). reliable, reproducible, and adequately sensitive, and fit for the intended purpose. Dr. Walton noted that “for some purposes, a method [of measur- ing a biomarker] may have adequate sensitivity, and for other purposes, it may not be adequately sensitive.” Therefore, “the validity of the method and its performance characteristics have to be thought about with the context of use in mind.” PRESENTATION By FEDERICO gOODSAID, CDER Federico Goodsaid, chemist at CDER, began by echoing Dr. Walton’s emphasis on context, as well as remarks by Dr. Williams, who argued

OCR for page 29
 BIOMARKER AND SURROGATE ENDPOINT EVALUATION that absent a context of use, it is difficult (if not impossible) to evaluate a biomarker. “We have had a process for qualifying biomarkers over the past 100 years,” Dr. Goodsaid said. “You wait long enough until you think you have all the references in the world needed to convince someone that a biomarker is valid. And when people become a little tired of arguing about it, they say, ‘okay, it’s qualified,’” Dr. Goodsaid said. “I think we can do better.” The real problem with qualification is defining the evidence required to qualify a biomarker for a particular use, Dr. Goodsaid said. For example, if one examined some of the biomarkers presented as case studies in the report without prior knowledge of their shortcomings, “what would we ask of a qualification process to conclude that they are qualified?” He said this is the overall question the FDA would like to be able to address, and it led him to further inquire, “how do you convince people that a biomarker is qualified for a specific purpose? What level of evidence is needed? If it is going to be for safety, what are the levels of evidence? If it is going to be for efficacy, what are the levels of evidence? Does it matter?” There have not been many discussions regarding biomarker qualifica- tion for nonclinical uses, Dr. Goodsaid noted. He mentioned one excep - tion: an account of a multiyear effort that resulted in the qualification of seven renal safety biomarkers by the FDA and the European Medicines Agency (EMEA) (Dieterle et al., 2010), based on the submission of drug toxicity studies and analyses of biomarker performance from the Predic - tive Safety Testing Consortium (PSTC). He described this achievement as a first step toward the goal of getting new biomarkers as efficiently as possible with comprehensive data supporting their claims. In closing, Dr. Goodsaid made the point that, unless biomarker qualifica- tion offers immediate financial gain for companies, motivation will be lacking for the development and qualification of new biomarkers. “We need to think about the fact that many qualification efforts will be closely linked to con- sortia that will be pulling their resources together to be able to try to qualify biomarkers,” he argued, and offered as examples the PSTC and an effort by the Chronic Obstructive Pulmonary Disease (COPD) Foundation. According to Dr. Goodsaid, the ultimate goal is to qualify biomarkers that are needed for a broad range of applications and to do so as quickly as possible, provided sufficient data are available to support specific contexts of use. Dr. Goodsaid noted that a sensible approach to making better and wider use of biomarkers would begin by qualifying various biomarkers for limited contexts of use, followed by attempts to expand their uses based on additional evidence. However, he added, the context of use must be known in order to know what data would be needed to support that application of the biomarker.

OCR for page 29
 FDA PERSPECTIVES PRESENTATION By ROBERT BECkER, CDRH Robert Becker, medical officer at CDRH, began his presentation by commenting on the IOM report: “It is notable, from my perspective, to see the extensive and nuanced assessment of biomarkers that has been provided in the report. We recognize that this is a challenging scientific area, and there is very obviously admirable product with respect to the effort and the report itself that has been generated.” He then provided an overview of device regulation at CDRH, which include some biomarker tests. CDRH regulates devices for treatment, diagnosis, or any other application with a medical context. Like drugs, regulation focuses on safety, which he defined as the reasonable assurance that the probable benefits of a device will outweigh its risks, and effective- ness, which he defined as the reasonable assurance that use of the device will prove clinically significant. Medical devices are classified by CDRH according to the level of risk associated with their use, said Dr. Becker. In the case of biomarkers, the risk is the potential harm to patients if the biomarker does not accurately reflect the clinical endpoint. Uses deemed high risk are often associated with cancer diagnosis, he explained, or with large population exposure, as occurs for some screening procedures (for example, mammography). In these cases, CDRH requires a premarket application, which establishes the safety and effectiveness of the individual device for its intended use, he said. Devices classified as moderate risk include products used to determine cancer prognosis and monitor diagnosed cancer in patients; these require a premarket notification by the sponsor. By contrast, low- risk devices such as tongue depressors and toothbrushes are usually exempt from premarket submissions. In response to an audience question, Dr. Becker noted that biomarker tests can also be approved through another pathway defined by the Clini- cal Laboratory Improvement Amendments (CLIA).2 “There are, of course, two paths by which tests used for medical care have made their way to market over the past 3 decades or so,” Dr. Becker said. While the FDA approval process focuses on the test itself, the CLIA regulations focus on the laboratory in which the tests are conducted, and are aimed at ensuring the quality of its procedures, staffing, qualifications, and management. 2 The Centers for Medicare & Medicaid Services (CMS) regulates all laboratory testing (except research) performed on humans in the United States through the Clinical Laboratory Improvement Amendments (CLIA). In total, CLIA covers approximately 200,000 laboratory entities. The Division of Laboratory Services, within the Survey and Certification Group, under the Center for Medicaid and State Operations (CMSO) has the responsibility for implementing the CLIA Program. The objective of the CLIA program is to ensure quality laboratory testing. See http://www.cms.gov/clia/.

OCR for page 29
0 BIOMARKER AND SURROGATE ENDPOINT EVALUATION Over the past several years, the FDA has expressed an increased interest in regulating aspects of laboratory development tests, he noted, adding that this is a topic of active discussion. A growing area of involvement by CDRH concerns the regulation of biomarkers intended to inform the course of medical treatment, according to Dr. Becker. A well-known example of such “companion diagnostics” is a test used to determine whether a patient with breast cancer could benefit from receiving the drug trastuzumab, an antagonist against a specific receptor (human epidermal growth factor receptor [HER2/neu]) that is overexpressed by certain breast cancers. “This is, I think, a prime illustration of the kind of circumstances that will tend to underlie per- sonalized medicine as it evolves,” said Dr. Becker. Moreover, he added, when personalized therapeutic decisions are driven by single test results, as is becoming increasingly common, the risks associated with that test are increased due to the heightened consequences of its failure. “You are not talking about a circumstance where the test result is simply integrated with a lot of other information,” he said. “You are now talking about a circumstance where the test result itself can drive a clinical decision.” For this reason, Dr. Becker emphasized that companion diagnostics should be reviewed and approved by the FDA before the test goes to market, because “the risks for such tests equal the risks of the therapeutic products.” He added that under such circumstances, CDRH aims to opti - mize the benefit-to-risk performance associated with using the test and treatment together. Dr. Becker commended the report’s treatment of analytical validation, and said that it reflected many of the considerations CDRH has in evaluat- ing biomarker tests. He provided a list of these considerations, which he called “not complete by any means, but … typical of the kinds of features or performance characteristics” that enable CDRH to assess how well a test measures a biomarker of interest: • Precision (repeatability, reproducibility); • Accuracy; • Sensitivity, limit of detection; • Specificity (interference, cross-reactivity); • Sample type/matrix; • Sample preparation/conditions; • Performance around the cutoff; and • Potential for carryover, cross-hybridization. The steps CDRH takes to conduct analytical validation of a biomarker- based device vary according to the technologies involved, said Dr. Becker. For example, different statistical paradigms may be used to evaluate

OCR for page 29
1 FDA PERSPECTIVES qualitative and quantitative tests, or the evaluation paradigm may reflect the setting in which the biomarker test is used. Typically, these tests are kits or devices intended for laboratory use, but CDRH has also regulated tests conducted by a single laboratory. A test’s intended use strongly influences its evaluation by CDRH, according to Dr. Becker. Context of use determines which features are critical to device design; these features may include the analyte itself, how it is measured, the population in which the test will be used, and the rationale for using the test. Ideally, this information is explicitly stated in the course of defining its intended use in the approval or clearance pro- cess, but he noted that applicants to CDRH have “a long history of … var- ied adherence to clarity” in the information they provide to support such devices. This shortcoming, he added, may reflect the fact that the science of device regulation is relatively new, compared with drug regulation, and also that some devices have more uses than most individual drugs. Thus, the center finds that their regulatory paradigms are continually evolving as their experience with the uses and risks of devices grows, he added. Dr. Becker presented what he called an “unofficial” list of elements that inform device regulation by CDRH: • The description of the product, • Its intended use (clinical context, purpose, target population, means of usage), • Instructions for its use, • Performance claims and limitations (for example, is it prognostic versus predictive? How well does it fulfill its intended purpose?), • How it will be manufactured and marketed, and • How potential problems with the device may be detected and resolved. Dr. Becker pointed out that these devices may operate in a range of medi- cal contexts that include risk estimation, screening, diagnosis, residual or refractory disease, recurrent disease, monitoring, prognosis, and pre - diction of therapeutic response. He emphasized that each application is unique in the studies or trials required to demonstrate the analytical and clinical validity of the test. It is often desirable to have more than one device on the market to meet a single intended use, said Dr. Becker. This is particularly the case for biomarker-based tests, he added, because they can employ a variety of measurement methods. When a “follow-on” test is developed for a bio- marker for which a qualified test already exists, it raises several issues, Dr. Becker said. “Even in something as simple as an immunohistochemistry

OCR for page 29
2 BIOMARKER AND SURROGATE ENDPOINT EVALUATION test for HER2, you can have antibodies that might be associated perhaps with different epitopes of the antigen that might convey with that some - what different performance characteristics associated with the test,” he said. “You can find, for example, that the very same analyte … might show up in a different clinical context” (for example, two different tumor types that vary in their expression of the analyte). Using HER2 as an example, Dr. Becker offered further explanation of considerations regarding “follow-on” tests based on biomarkers. HER2 is expressed at a variety of levels that can be assessed from the perspective of either biochemistry or cellular physiology, with one method measur- ing the HER2 protein itself while the other method relies on reverse- transcriptase polymerase chain reaction (RT-PCR) to monitor expression of the HER2 gene. “These different approaches to measuring the bio- marker bring with them different aspects of analytical performance,” he said, to the extent that “you wonder at what point they actually diverge in terms of the[ir] biological implications.” In addition, two tests might measure the same biomarker in different matrices (for example, within a tumor versus serum levels). The science associated with biomarkers and their applications is itera- tive and continually evolving, and it is important to recognize that scien - tific understanding of a biomarker can be “outrun by technical progress,” said Dr. Becker. For example, gene signatures represent a descriptive or empirical compilation of biomarkers that might be shown to correlate with clinical outcomes or features, but for which the underlying biologi- cal significance is unknown. He noted that an ongoing, resource-intensive process of biomarker qualification and utilization, as recommended by the committee, could lead to increased understanding of the biological role of each biomarker and the expansion and refinement of its uses.