Data sharing is critical to the validation of new biomarkers used in drug development, experts at the seventh annual Predictive Safety Testing Consortium (PSTC) workshop at the US Food and Drug Administration's (FDA) White Oak campus on Monday said.
Biomarkers are biological characteristics that can be measured to give an indication of a biological process or the presence or likelihood of a disease. They can also predict how a patient might respond to a particular intervention.
However, for FDA to rely on a biomarker as part of a product approval, a sponsor must demonstrate the validity of the biomarker in the context in which it’s being used.
Alternatively, biomarkers can be qualified through FDA's Biomarker Qualification Program, which offers a pathway for biomarkers to be qualified for regulatory use more broadly.
Speaking on the need for well-validated biomarkers, Christopher Leptak, co-director of the Biomarker Qualification Program at FDA, said, "Ultimately we're trying to improve our scientific understanding for the benefit of patients."
"In recent meetings, this whole concept of data sharing has either been a component or at least [part of the] discussions that are occurring … Not only are we talking about the same questions and topics, but we're actually getting deeper and deeper into the logistics of how it might work," Leptak said.
Because demonstrating the validity of a biomarker requires a considerable amount of data, the Critical Path Institute (C-Path) and other groups such as TransCelerate Biopharma have worked to develop ways to pool data from multiple sources.
John Michael Sauer, director of regulatory strategy and submissions at C-Path, said his vision is "to be able to use data that’s generated during the registration of drugs for biomarker qualification."
"In order to make this happen, what we need to think about is how we create a safety biomarker data repository to house nonclinical and clinical data," he said. "Is there a way that we can use data generated during the normal conduct … of drug development to go ahead and fill this biomarker database? To be able to use those data not only to qualify biomarkers, but also to change the context of use for certain biomarkers, very much like we change the label for a drug as we get new indications or we get a new safety finding?"
To create such a repository, Sauer said, stakeholders must work through a number of issues, including who will have access to the data and how it will be blinded and masked. Sauer also added that the value proposition and incentives for industry to participate must be clear.
Another issue, according to Richard Liwski, director of C-Path's Data Collaboration Center, is that when it comes to data, “often the most difficult process … is actually getting it."
Aside from finding companies willing to share their data, Liwski said considerations must be made for how data-sharing agreements are written, as well as for cybersecurity and data governance.
However, when data is pooled properly, Liwski said, "You can improve that signal to noise ratio and see patterns emerge that you wouldn't see with a single data set."
Benefits to Drugmakers
Ed Bowen, senior director of translational and bioinformatics at Pfizer, added that data sharing has significant potential to benefit drugmakers.
However, he noted that drugmakers don't always see it this way. "[Data sharing] has historically been a very scary thing for sponsors … I think the culture within the industry as a whole is changing significantly. Companies are much more inclined to share now than they ever were," he said.
"First and foremost, in terms of value," Bowen said, "is reducing the number of patients in a control arm."
He also offered an example of an upcoming Phase IIa trial that Pfizer is planning. According to Bowen, Pfizer was able to identify another company that is willing to share data on some of its patients that would allow Pfizer to cut the number of patients in the control arm in half.
This reduction, Bowen said, will reduce the time necessary to run the trial by five months and save the company $1 million in patient costs. Furthermore, he said, Pfizer was able to save an additional $7 million by eliminating the need for an observational study as part of the development program.