July 2008: Volume 34, Number 7
The Prospect of a National Lab Service
Will Creating Standard Methods and Samples Speed Approval of New Markers?
By Debrorah Levenson
For many diagnostics companies, the road to FDA approval of a new cancer biomarker assay is long, twisted, and costly. Compiling the necessary data can take years and significant capital. First companies must make a case for a test’s robustness by showing similar results from multiple labs, and then they must prospectively validate the marker to prove its clinical relevance. Company officials have long complained to the FDA that the lack of standard methods and samples complicates and extends this process. Furthermore, they fear that the actual clinical use of an approved assay may not conform to the methods used to produce data for the FDA submission. Such inconsistencies raise questions about how physicians interpret the test’s results and could ultimately undermine the clinical utility of new cancer biomarkers.
Now the Critical Path Institute (C-Path, Tucson, Ariz.) aims to overcome these issues and to help hasten approval of new cancer biomarker assays. C-Path plans to establish a national lab that it hopes would become an important facet of the nation’s system for validating in vitro diagnostics. Known as the United States Diagnostic Standards (USDS), the lab would design testing protocols and provide standardized clinical sample procurement, data processing, and readouts. “USDS would be a way to do premarket testing of diagnostics and would generate standardized data that would improve the evaluation of diagnostic tests for approval by the FDA,” explained Maryellen de Mars, C-Path’s Director, Clinical Biomarkers. “The idea is to provide standardized samples and procedures. It’s quality assurance, in a way.” According to the organization’s Web site, C-Path is an independent nonprofit research and education institute that helps FDA’s Critical Path Initiative by spurring “collaborations among the FDA, academia, and regulated business to improve the process of developing new medical products, making them faster, safer, and smarter” (See Box, below).
C-Path Initiative Targets Preclinical Safety Biomarkers
Another C-Path project aims to establish new ways of determining experimental drugs’ human toxicity before they are tested in patients. Launched in March 2006, the Predictive Safety Testing Consortium plans to help FDA and its European equivalent, the European Medicine Evaluation Agency (EMEA), qualify new biomarkers of human safety.
The consortium’s director, William Mattes, PhD, says there is a major need for new biomarkers of toxicity. For example, some preclinical tests of toxicity are as much as 60 years old, and in certain cases, related data comes from anecdotal reports, not actual scientific scrutiny.
Through their membership in the consortium, 16 pharmaceutical companies are now sharing their internally developed preclinical safety biomarkers and testing each other’s methods to see if they are reproducible. The consortium’s ultimate goal is to present the companies’ data to the FDA and EMEA, which will deem some of the markers the consortium examines as qualified for use in drug development, according to Mattes. Last year, the consortium submitted seven urinary biomarkers of kidney injury, and in May 2008, the FDA and EMEA confirmed their joint review and acceptance of these tests.
C-Path plans to submit data on more potential biomarkers. Other consortium projects are focused on markers of harm to the liver, muscle, and vascular system. Another is examining focuses on carcinogenicity in rodents.
The consortium hopes that the companies’ shared experience with the safety biomarkers will go a long way toward establishing consensus about them. “We want to develop new, better, and more sensitive and specific tools for determining safety and get broad industry, regulatory, and scientific agreement about their safe use,” explained Elizabeth Walker, PhD, the consortium’s Assistant Director.
Laboratory directors familiar with the effort believe it has the potential to benefit manufacturers, clinical labs, and patients, but say that the concept of a national diagnostic testing service raises some serious questions. How will C-Path fund what promises to be a huge and costly project and still maintain the lab’s independence? Will the existence of the USDS increase pressure on FDA to regulate lab-developed tests? And will USDS compete with academic labs that currently partner with companies on validation studies?
Fulfilling a Need
Right now, FDA submissions often raise questions about variations in sampling methodologies, assay procedures, and how well clinical data correlates to particular samples, a process known as clinical annotation, noted Fred R. Hirsch, MD, PhD, Professor of Medicine and Pathology at University of Colorado Cancer Center in Denver. Differences in how individual labs handle these processes sometimes make FDA, labs, and physicians question that data. “When there is variation in procedures, it’s difficult to interpret the data. You want to be sure that results reflect true biological variation and not just variation in methodology,” he explained.
Standardization of cancer biomarker assays could be an aid to FDA, according to AACC Past President Gary Myers, PhD, Chief of the Clinical Laboratory Branch in CDC’s Division of Laboratory Sciences, which has helped create important standards for cholesterol and HbA1c assays. That’s because the FDA must often evaluate data presented in different formats. “Having a lab where standard sample sets are used to devaluate diagnostic products, and evaluation results all follow common, standardized formats would make it easier for FDA to review and evaluate product submissions. It could compare apples to apples and oranges to oranges,” Myers explained.
A Menu of Useful Services
An executive summary of a proposal describing USDS lists several benefits that C-Path hopes to offer diagnostics companies. These include standardizing the validation process, a ready supply of standardized samples, data that compare potential products to those of competition, and neutral third-party evaluation. The summary goes on to provide details of the proposed repository, which would be “a collection of highly qualified, annotated clinical samples.” The repository would include tissues and fluids with confirmed diagnoses, clinical outcomes, and controls. USDS would be subject to CLIA, CAP, and International Organization of Standardization requirements, the summary adds.
De Mars emphasized that C-Path is in the early stages of planning for USDS. “We’re in the process of compiling an advisory board for the project to define scope and strategy. But there’s general support and excitement throughout the industry about creating standards in this area.”
A key FDA official confirmed that discussions about the project are still preliminary and that C-Path is “working out the bugs” in its concept. But Steven Gutman, MD, Director of the Office of In Vitro Device Evaluation and Safety, also emphasized that FDA’s role is minimal at this point. While the agency supports the concept of a national diagnostic service lab, C-Path initiated the current proposal for how the lab would operate. FDA has given C-Path advice but no money so far. “FDA’s position on USDS, as with many C-Path activities, is that we offer our perspective and maybe funding for small studies, and form partnerships,” Gutman explained.
While he minimizes FDA’s role in USDS, Gutman says its planned contribution to the development of new assays would be valuable. “It will improve the quality of the science in submissions. With or without regulation, if you can improve science, that’s always a plus,” he said. Any type of assay that isn’t well-standardized might benefit, he added, noting that USDS might be particularly helpful in preparing proteomic assays for FDA submission. “We haven’t cleared one yet. There’s a paucity of methods, materials, and standards for proteomics. If the new enterprise improved standardization for these assays, it could really move the field forward.”
The Pilot: A Focus on EGFR
C-Path expects that an NCI clinical study, now in its early stages, will help establish a USDS pilot program by determining the infrastructure necessary to provide planned services. The phase III trial will screen almost 1,200 non-small cell lung cancer patients to determine EGFR copy number by FISH. The trial will retrospectively determine the presence of EGFR mutations by sequencing and protein expression by immunohistochemistry (CLN, February 2008). Afterward, researchers will randomize patients, who have already had first-line therapy, to receive either the tyrosine kinase inhibitor (TKI) erlotinib or chemotherapy with pemetrexate. The trial’s primary goal is to determine the most suitable EGFR biomarker and to identify the patients who benefit most from TKI inhibitors, according to Janet Dancey, MD, Associate Chief of the Investigational Drug Branch in NCI’s Cancer Therapy Evaluation Program. She is coordinating the trial.
De Mars identified other goals that are more directly related to USDS’s mission. The study’s first phase will not only determine the best method for measuring EGFR, but will also help establish a standardized process for FDA to validate and assess predictive biomarkers. Future phases could also help diagnostic companies access valuable clinical samples for developing new biomarkers and technologies, cross-validate new tests, and support a standardized process for evaluation of diagnostics by FDA. C-Path has proposals from a total of 14 diagnostics companies with technologies that measure EGFR or other potential targets. Future phases of the trial might use these technologies, with the goal of identifying new predictive or prognostic markers.
Colorado Cancer Center’s Hirsch is coordinating the biomarker studies for the trial on behalf of Southwest Oncology Group, which has headquarters in Ann Arbor, Mich., and researchers at about 550 institutions. He described the study as a prospective biomarker-driven trial that is unique in its scope. While various groups have studied outcomes of patients whose treatment was based on assessment of their EGFR status, this trial is the first to join a large number of oncology cooperatives, NCI, and diagnostics companies, he explained. “The trial is also a demonstration that prospective biomarkers can be validated in a collaborative way among all of these partners. They have spent a lot of time building up a complex infrastructure.”
At the end of the study’s first phase, researchers hope to have 970 patients’ blood and tumor samples, all collected according to standardized methods for analysis of mutations, gene copy number, and protein expression. Meanwhile, the data will be based on standard mutation, gene copy number, and mutation methods. “While this type of standardization exists in individual labs and centers, we will have several centers and cooperatives all using a standardized methodology,” Hirsch noted.
The samples will be both precious and well-suited for tests of other markers, according to Dancey. Specifically, she noted that each of the samples will be linked to outcomes data for a certain therapy, so in the future, researchers investigating toxicity or survival will be able to trace particular data back to specific samples.
A similar but smaller study has already dealt with many of the questions the NCI trial aims to answer, noted Marc Ladanyi, MD, Chief of the Molecular Diagnostic Service and Attending Pathologist at Memorial Sloan-Kettering Cancer Center. Led by Vince Miller, MD, also at Memorial Sloan-Kettering, the trial involved 101 patients and found that the presence of an EGFR mutation was a much stronger predictor of both responsiveness to and prolonged survival on erlotinib than EGFR amplification. Tumors with EGFR amplification without mutation showed only rare responses and minimal survival benefit (Journal of Clinical Oncology 2008; 26: 1472–1478). He also questioned whether other aspects of the NCI trial are new. “In some earlier, major studies, predictive statements were based only on the small subset of patients in which paraffin samples could be retrospectively obtained for testing. But now, most major cancer centers are running clinical trials with procurement of more complete sample sets, so there is nothing special or unique about what is being proposed by Critical Path.”
Questions from the Lab Community
While USDS could provide a valuable service to diagnostic manufacturers—access to high quality samples for validating new assays according to standard procedures—Myers predicts that lab directors at academic hospitals will have many questions about it.
The first concerns funding. “C-Path has taken on a daunting task. Who will pay for it?” he asked. According to de Mars, the Science Foundation of Arizona has granted the seed money, and USDS is still developing a business model and raising more money for the project. She did not divulge the identities of potential funders, but she did note that C-Path is eying federal legislation introduced last year by U.S. Rep. Gabrielle Giffords (D-Ariz.). The Safe and Effective Drug Development Act (H.R. 2592) would allow FDA to enter into public-private partnerships that would accelerate development of medical products. While the legislation would prohibit these partnerships from accepting manufacturers’ money, it would also establish an exception that allows partnerships to accept funds from a consortium of companies with FDA-regulated products. The HHS Secretary would have to deem the situation free from conflicts of interest and issue a waiver, according to the bill.
Current C-Path plans do not call for diagnostics companies to fund USDS directly. Rather, companies would pay fees for services that would generate validation data. “The data would be theirs, and they could use it for FDA submission,” de Mars explained, adding that USDS would base fees on the scope of services provided.
Competition for Academic Labs?
Another concern is that several academic medical centers already provide many of the services USDS hopes to offer, Myers added. That’s a sentiment shared by Ladanyi. “The C-Path sample set will be nice. But this type of sample set is already being created at many cancer centers,” he pointed out.
Gutman agrees that much of what C-Path proposes that USDS do isn’t exactly new. “The idea of collecting samples is good, and any mechanism for creating a panel of reference samples is good. But they already exist,” he said. “However, this [USDS] would be more deliberate and focused than other activities. It would be interesting if it got some formal status.”
USDS would compete with academic labs that diagnostics companies now support by paying for studies to validate experimental assays, Myers continued. Addressing this concern, de Mars noted that FDA usually requires that submissions include data from three laboratories, one of which is usually a diagnostic company’s own facility. She sees room for both USDS and academic medical center labs in the pro-cess, because USDS and an academic lab could be the second and third locations. According to deMars, “USDS will initially focus on development of standards for cancer statistics. In its neutral position, USDS would have the capacity to run competing assays, from any manufacturers, on the same set of standard samples. Also, with eventual growth and expansion of services, USDS may need to outsource to other labs.”
Honing in on Homebrews
Both the USDS proposal summary and de Mars point to another role Myers believes will be of concern to many lab directors: review of laboratory-developed assays. De Mars said USDS could provide standardized evaluation of lab-developed tests, in a process she likened to certification. “This USDS ‘certification’ would aid in marketing tests that manufacturers develop in-house,” she explained, adding that companies could have USDS test their reagents on standard sample sets and publicize the data in the hope that it spurs sales.
Myers also wondered if USDS could be a mechanism for FDA regulation of lab-developed tests. Members of the HHS Secretary’s Advisory Committee on Genetics, Health and Society are among those who support FDA oversight. At its most recent meeting in February, the committee said it would recommend that all lab-developed tests come under the FDA’s watch to ensure their effectiveness and patients’ safety. At that meeting, Gutman said that while FDA does have authority to regulate these assays, it lacks the resources to do so. “It’s not possible [to oversee all assays], but the high-risk tests belong at the FDA,” he said during the meeting. But Gutman told CLN that FDA “has not established specific goals for how USDS work might be applied to FDA testing oversight.”
Whatever USDS’s scope, de Mars hopes it will bring the benefits of pharmacogenomics to more cancer patients. “Companion diagnostics are really the dawn of a new era. Before, cancer patients were all treated with a ‘one-drug-fits-all’ strategy,” she explained. “Now drugs have been developed that benefit a subset of patients. We are moving toward personalized medicine, and USDS can help advance that further.”