There’s something about cancer. Scientists taking it on are more intense, more purposeful.

In talking with the National Cancer Institute’s Ken Buetow recently, we heard a hint of that urgency. Buetow is associate director for bioinformatics and information technology and the director of the National Cancer Institute’s (NCI) Center for Bioinformatics. He is an architect of the caBIG—a network for oncology collaboration designed for long-term scientific collaboration, not short-term, trial-by-trial success.

To a large degree, the NCI has re-imagined the clinical trial data landscape in a way that is antithetical to the approach taken by the pharmaceutical industry. Buetow and the NCI team didn’t pick standalone components of their system first, and try to connect them with scotch tape later.d9A2t49mkex

Fast Start

Instead, NCI published a framework for interoperable systems, and then started designing individual components that would have no trouble talking to each other. If there were a Lasker prize for clinical systems, caBIG deserves one, and it’s barely under way. Here are the clinical trial platform page and a list of all aspects of the caBIG project.

In a few years, spending perhaps $20 million annually, caBIG has built 40 applications or databases that are at various stages of completion. That includes an arsenal of tools for basic science. On the clinical side, there is a tissue bank, a patient study calendar, an adverse events reporting system, an electronic lab notebook, an imaging archive. caBIG, in short, is an example of the U.S. government setting out to do something laudable and audacious—and then accomplishing it.

“We have to figure out ways to do trials better and faster,” Buetow says. “The key thing we’ve learned is the importance of the focus and the definition around standards, standards at a variety of different levels. What’s interesting to me is how far we have been able to leverage the novel opportunity and efficiency that comes from the use of a common electronic infrastructure.”

FDA Role

A geneticist by training, Buetow and the NCI probably face a governance challenge that is as gnarly as any in the land. Somehow he corralled the innumerable egos and agendas of federal and university-based scientists. “People can put down their proprietary and competitive forces,” Buetow says. “The one place we can get universal agreement in our community is around the structure and definition about how we collect our information and our applications can talk to each other.”

The caBIG project uses the clinical data management system from Oracle, and has solicited bids from other suppliers for other components, though the NCI has not yet chosen a system for electronic data capture (EDC).

All along, regulators have been participating in the caBIG discussions. There is a clear path to the market for any compounds that can be commercialized. “FDA as well as the academic and commercial partners have all been at the table,” notes Buetow.

Data Standards

The whole effort, Buetow says, was built with Clinical Data Interchange Standards Consortium (CDISC) standards in mind. That includes the CDISC project on case report forms (CRFs), for which special oncology CRFs are being developed. If perhaps 80 percent of the CRFs can be found in a library, a much smaller subset will need to be developed on a custom basis.

“What we see as the opportunity in caBIG is to make it as easy or easier to use standards than create [clinical systems] de novo,” says Buetow. There is no additional burden to use standards, he notes, when all the components of caBIG conform to them by design, from the outset, without the customization or modification that delay so many trials.

Thinking Different

caBIG has also kept an eye on extracting a bit of general health care data from generic hospital and physician computer systems. caXchange is a bridge between clinical and HL7 or hospital-based electronic health record (EHR) systems.

Buetow also touts a system called c3PR, or Cancer Center Clinical Participant Registry. “C3PR provides current enrollment statistics and a repository for participant information across studies, sites, systems, and organizations,” its web page observes. Not all of that information may available to a sponsor, Buetow notes, but it should be able to help find patients and collate information about them. Explains Buetow: “It is a bridge that tells which information should be transferred to a clinical research record or clinical care.”

Investigator Database

He’s also proud of the Firebird registry, which automates the collection of the clinical investigator 1572 forms. It’s a clearinghouse that will be accessible both to government backers of clinical trials and industry, and facilitate one central place where a scientist-physician’s forms can be updated and stored. The forms can be digitally signed. Pfizer and Johnson & Johnson are participating in the testing.

“It has been a true partnership between government, academia and the private sector, which is always exciting to me, when we can have common ground,” says Buetow. Firebird has been used on a handful of live trials, he says.

In general, Buetow believes the caBIG philosophy may be relevant in industry, where scientists and organizations are also separated by geography and silos of expertise. Connecting different organizations and getting science done is causing indigestion across the research world, he suggests, simply because of the difficulty of connecting so many systems.

Better Connections

“The lack of interoperability is now a real barrier,” Buetow says. “The lack of this interoperability is now slowing them down.” He speaks of “cumbersome, disconnected processes,” not to mention the inevitable shift from paper to electronic data, and it’s clear that he believes that technology can make a difference in fighting cancer.

Industry is welcome to contribute to the effort, but parts of the caBIG code must be left in the public domain. (That has discouraged some prominent clinical trial technology firms from jumping in.) “We’ve had a lot of informal conversations with industry,” Buetow says.

Some of the dynamics about what can be shared, he adds, seem to be changing: “I’m impressed at the increasing recognition of the value of starting to look at some of this stuff as precompetitive and to work and build a better system.” It will be fascinating to watch whether researchers using the caBIG tools can not only access networked samples, images, and data but move science forward.

Editor’s note: Here’s a list of FAQ about the caBIG project.