The US Food and Drug Administration has identified bioinformatics as one of six broad topics that promise to speed drug development while lowering costs before the end of the decade, but it is still unclear how informatics vendors can best take advantage of this newfound stature in the eyes of the regulatory agency.
Last week, the FDA released its Critical Path Opportunities List a catalog of 76 initial projects that it described in a statement as "a starting point in identifying priorities to be accomplished under the Critical Path Initiative."
The list, which is based on feedback from the Critical Path Initiative unveiled in 2004, includes eight bioinformatics-specific projects that range from biomarker identification to biosimulation to database development (see box, below, for the complete list of bioinformatics projects).
The list does not carry the same influence as a formal FDA guidance or promise funding, but some informatics vendors still view it as a positive sign for their businesses.
"Obviously, it would be far more powerful if there were specific funding or stronger guidance available today from the FDA, but the general momentum for the Critical Path Initiatives is a good sign," Jill Fujisaki, co-founder and vice president of scientific alliances at Entelos, told BioInform via e-mail.
"Obviously, it would be far more powerful if there were specific funding or stronger guidance available today from the FDA, but the general momentum for the Critical Path Initiatives is a good sign."
Fujisaki said that Entelos is pleased to see that the FDA is "following up in its initial push for change and has outlined specific project areas to give their original white paper more 'meat.' We believe pharmaceutical companies will be paying attention to see what happens next, and may even pick and choose among the projects that are aligned with their own needs."
The goal of the Critical Path Initiative "is to take advantage of new scientific tools to meet existing challenges to medical product development. In no field is the opportunity greater than bioinformatics," FDA said in a report outlining the Opportunities List (available here).
The FDA defines bioinformatics rather broadly in the report as "the application of mathematics, statistics, and computational, quantitative analysis to biological data," and notes that "with recent advances in the bioinformation sciences, it should be possible to analyze and mine large sets of biological data about patients, with the goals of creating robust, quantitative computer models of normal human physiology, of the natural history of certain diseases, and of the course of a disease as affected by standard treatments."
In addition, the FDA notes in the report, "The findings from in silico testing … could reduce the risk and cost of human testing by helping [drug makers] make more informed decisions on how to proceed with product testing and when to remove a product from further development."
Biosimulation companies like Entelos and Gene Network Sciences stand to gain from the FDA's endorsement of in silico testing, but companies developing software for data-mining, statistical analysis, data-management, and gene or protein biomarker discovery could all see some benefit from the initiative either through direct involvement in the Critical Path projects, or indirectly, through biotech and pharmaceutical clients that align their internal research goals with FDA specifications.
Entelos' Fujisaki said that the FDA's support of biosimulation is likely to have "a positive, but indirect, impact on adoption" of its technology because "nothing has a bigger impact on customer adoption than showing value very early on in a project that has high visibility and high impact in terms of a customer's key decision-making processes."
Nevertheless, she said, there is some chance that the FDA's priority list could sway internal technology-adoption decisions within pharma. "The way resources are allocated within a large pharmaceutical company is highly competitive," she said. "Thus, if one scientific champion [in the company] is pushing for one particular technology and another champion is promoting a different technology, it wouldn't hurt to be able to point to the FDA to help make one's case for choosing one technology over another."
From the General to the Specific
FDA first kicked off its Critical Path Initiative in 2004 with a white paper (available here) that was a broad call for new technology development to speed drug development. Last week's priority list is based on feedback from that initial white paper, the FDA said, and fleshes out the initiative's general goals with some specific projects.
The list will also help the agency set priorities. Of the six broad topic areas, FDA said that it reached a "consensus" that the two most important areas for the initiative are biomarker development and streamlining clinical trials.
As a result, the FDA announced its first formal Critical Path undertaking on the same day that it released the Opportunities List: a private-public initiative called the Predictive Safety Testing Consortium that will share laboratory methods to predict the safety of new treatments before they are tested in humans. The consortium includes the independent non-profit Critical Path Institute and its founding partner SRI International, along with five large pharmaceutical firms: Bristol-Myers Squibb, GlaxoSmithKline, Johnson & Johnson Pharmaceutical Research & Development, Merck, Novartis, Pfizer, Roche, and Schering Plough Research Institute. The FDA is not a member of the consortium, but will "assist it in an advisory capacity," according to a statement.
Ray Woosley, president of the C-Path Institute, said that of the initial projects already underway at the institute, "all of them are going to rely heavily on some aspect of bioinformatics." Many of the 76 projects in the Critical Path Opportunities List will require informatics development, he noted, because "one of the biggest challenges in drug development is managing information."
The Predictive Safety Testing Consortium, in particular, will "require a very large database to be able to maintain those for comparison of results," Woosley said.
The consortium's goal of sharing predictive safety biomarkers could be good news for bioinformatics software firms, as well.
Yelena Shevelenko, vice president and general manager of Rosetta Biosoftware, said that the initiative "is an opportunity we have been preparing ourselves for," and that Rosetta has been "focused on supporting these kinds of biomarker discovery and validation efforts for the last two years."
While Shevelenko hesitated to speak for Rosetta's pharmaceutical clients, she said that "we're very pleased to hear that pharmaceutical companies have agreed to share results" as part of the consortium, and that the project partners would certainly require "robust data management capabilities that our products provide."
As far as internal pharmaceutical projects go, Timothy Perera, a principal scientist in the oncology research early development group at Johnson & Johnson Pharmaceutical Research and Development Europe, said that the safety biomarker project only underscores a trend that has been emerging for several years. "There have been increasing efforts from pharma companies in general, and certainly within J&JPRD, to try and identify biomarkers to go with compounds, so we are trying to do biomarker discovery earlier and earlier in the development phase," he said. "So the FDA announcement certainly hasn't changed our approach there, though it underlines the fact that it is crucially important for us to do so."
Bioinformatics Projects in the
Critical Path Opportunities List
|Identification and Qualification of Safety Biomarkers: A collaborative effort to "pool and mine existing safety and toxicology data [to] create new sources for identification and qualification of safety biomarkers."|
|Virtual Control Groups in Clinical Trials: Would provide "databases, models, and/or imaging collections" that could be used by multiple sponsors as "historical controls to reduce the necessary size of control groups in clinical trials."|
|Adverse Event Data Mining: Involves combining adverse event data related to a product, a class of products, or a disease that could "enable identification of previously undetected patterns of safety events and/or co-morbidities and could elucidate drug-drug interactions."|
|Multiple Complex Therapies: Would pool data on the effects of the combined use of complex technologies to create information that would improve both patient safety and new product development.|
|Modeling Device Performance: Would develop a "rigorous model of specific aspects of human physiology [that] could allow more predictive in-silico … testing of implanted devices, prior to human testing."|
|Clinical Trial Simulation: Calls for clinical trial simulations using in silico modeling to "predict efficient designs for development programs that reduce the number of trials and patients, improve decisions on dosing, and increase informativeness."|
|Failure Analysis: Would develop a public database of information from trials of unsuccessful products that could help identify "patterns associated with failure and help sponsors avoid repeating past mistakes."|
|Natural History Databases for Rare Diseases: Calls for the development of databases recording the natural history of patients with rare diseases, incorporating observations on clinical progression and biomarkers.|
Last week the FDA also released its first request for applications under the Critical Path initiative, which includes a bioinformatics component. Entitled "Collaborative Cardiovascular Drug Safety and Biomarker Research Program," the RFA calls for, among other things, the development of "a program to develop and clinically validate a genotype-driven warfarin dosing algorithm [to] improve the therapeutic efficacy and safety of warfarin dosing." (The RFA is available here).
The FDA said that it will identify several more priority Critical Path opportunities "over the next several weeks." Some of these projects "could be undertaken by one organization; some will require collaborations coordinated and supported by the FDA," the agency said.
The Critical Path Initiative will require "unprecedented collaboration among public and private sector partners," the FDA said, adding that it intends to play a "consultive role" in many of these projects, which are likely to involve other federal agencies such as the National Institutes of Health, the Centers for Medicare and Medicaid Services, and the Agency for Healthcare Research and Quality.
Whether the FDA intends to contribute any funding to these projects is still unclear. President Bush requested $5.94 million to help pay for Critical Path in FY 2007, but the agency has not yet indicated how it intends to distribute those funds if it obtains them.
Regarding mechanisms for fostering bioinformatics projects, the FDA noted in its report that "no one company, university, or governmental agency has the necessary information bases to create computer models sufficiently robust to accurately predict product safety and efficacy."
Therefore, "Collaborative efforts will be needed to reap the benefits of model-based product development." In addition, the FDA said, "new strategies for information-sharing and safe information housing will be needed."
The agency acknowledged the "logistical, technical, and resource issues involved in intelligent mining of large data sets," adding that "stakeholders will need to address an array of technical issues, including determining how to access, organize, and ensure data quality, and the need for new IT systems and data exchange standards."
The FDA added that "development of and consensus on database structure and content, algorithms, and data quality standards are necessary, but remain unexploited Critical Path opportunities.
Bernadette Toner ([email protected])