In an attempt to improve the development of biomarkers, the American Association for Cancer Research, the US Food and Drug Administration, and the National Cancer Institute brought together stakeholders from academia, industry, and funding and regulatory agencies to identify the hurdles in biomarker development so that solutions can be found and the process streamlined.
The resulting report, published in the July issue of Clinical Cancer Research provides the community with a good place to start, key participants said in interviews for this article.
The lack of biomarker reproducibility “is exactly why the recommendations were put together,” said Samir Khleif, MD, Head of the Cancer Vaccine Section of the National Cancer Institute and a lead author on the report.
“This is exactly why the collaborative was put together—because biomarkers have been developed, and are being developed, for a long time, the question is the uniformity of those biomarkers, the validation of those biomarkers, and the qualification of those biomarkers. This is why it is important to have a process that defines what these things are, rather than have each one developed on their own in a different way and with different standards.”
27 Consensus Recommendations
The collaborative group made 27 consensus recommendations in eight areas of biomarker development, including biospecimens, analytic performance, standardization and harmonization, bioinformatics, collaboration and data sharing, regulatory issues, stakeholder education and communication, and science policy.
Dr. Khleif said that the recommendations are not particularly surprising, but will help move things forward.
“It is important to define the exact problems in order to be able to find solutions for them. What the collaborative groups did was to sit down and think of all the hurdles or barriers, define them properly, and put together certain elements to overcome those issues, not necessarily in a hurdle-by-hurdle way, but in an integrative way.”
The problems are not trivial, but they can be solved, says Steven Gutman, MD, MBA, Associate Director of Technology Evaluation at Blue Cross Blue Shield, who was previously with the FDA's Office of In Vitro Diagnostic Device Safety and Evaluation.
“From my perspective, biomarker development is a holy mess, and it is not because people have not been trying,” he said.
“You know those series of movies, a scary movie, and another scary movie, and another scary movie? When I was at the FDA, that is how we felt about biomarker meetings. There would be a biomarker meeting and another biomarker meeting and another biomarker meeting. And always talking great thoughts, and real leadership efforts, and yet it hasn't quite happened.
“I think the reason it hasn't happened is because it is really difficult,” he continued. “You take any subsection of the AACR report—and I was on the subcommittee looking at issues related to analytical validation—and it is just daunting. It is very complex science. It is very complex issues related to regulation, to reimbursement, to actual clinical use. It is very complex sampling issues, very complex financial issues in terms of development, and the system hasn't been able to overcome the complexity.”
Dr. Gutman said that moving forward, he is particularly interested in seeing improvements in biospecimen quality control and for the FDA to develop guidance on the development of predictive tests—both of which, he noted, are being tackled currently. The FDA expects to publish a draft guidance on predictive biomarker development later this year, and the NCI Office of Biorepositories and Biospecimen Research (OBBR) is working to improve the quality and consistency of human specimens.
3 Recommendations on Biospecimens
The consensus report includes three recommendations on biospecimens, including establishing standards and promoting routine quality assessment of biospecimens, developing publicly available oncology biospecimen reference standards for validation, and promoting research on biospecimen quality.
Although human biospecimens have been the basis of medical research forever, the quality of the samples has recently become more of an issue.
“Technologies for molecular analysis are so powerful, so cheap, and so fast, it makes it possible for investigators worldwide to use them,” said Carolyn Compton, MD, PhD, Director of OBBR and Co-Chair of the Biospecimens Committee for the AACR-FDA-NCI collaborative group.
“But because they are so powerful, they have raised the bar for the quality of analytes that are interrogated: They can now produce the wrong answer with spectacular speed if you don't use high-quality analytes on the front end.”
In the past, the bar for quality was relatively low, Dr. Compton noted. For example, immunohistochemistry tests required that only one molecular species, usually a protein, remained intact and detectable in a biospecimen. By contrast, current tests can measure thousands of species at the same time.
“If you don't know anything about the quality of the sample, that brings into question the interpretation of the data. It has proven exceedingly challenging, if not impossible, to reproduce data on biomarkers from center to center, study to study, and of course the question is: ‘Is this because the thing we are testing is so heterogeneous?’ We need to standardize what we are testing if we want to have interpretable results. And that brings it back to the biospecimens.”
To tackle this issue, Dr. Compton's NCI office is promoting research into biospecimen themselves—i.e., instead of just using the specimens to answer other questions, her group is supporting research on what variables affect biospecimen quality. For example, the group is interested in learning how long a tumor sample can sit between removal from a patient during surgery until fixation before it deteriorates and is no longer of sufficient quality for a given assay.
No Quick Fixes, but Significant Start
None of the participants interviewed for this article expect that the report will improve biomarker discovery and development immediately, but they do think it is a significant start. “These are pretty diverse recommendations looking at the biomarker development process from all different aspects, so some recommendation requirements are different than others. Some require refinement and further digging into the process, and hopefully that will be done,” Dr. Khleif said.
Everyone agrees though that a key feature of the AACR-FDA-NCI effort was that it brought together different stakeholders in biomarker development, including cancer researchers, regulators, product developers, industry, and funders. “Everyone had a little bit different perspective,” Dr. Compton said. “Everyone had a little bit different stake, different incentives, and different concerns. So it was important to bring everyone together to say “Okay what will it take?”
“There is a feverish effort and a gazillion dollars going into biomarker research and we are producing basically nothing. That's wrong. How can we fix it? It will take a little bit on everybody's side to make sure issues are addressed and to do it in a way that meets everyone's needs.”