InforSense this week said that it has entered into a three-year collaboration with Dana-Farber Cancer Institute to develop a translational research informatics infrastructure that will give researchers at the institute ready access to clinical and experimental data.
The agreement marks InforSense’s third translational research deal with a major cancer research organization and helps secure its presence in a field where there is growing demand for informatics solutions, but no off-the-shelf commercial products.
The company first entered the translational medicine market in 2005 when it began working with the Windber Research Institute to adapt its flagship workflow technology to merge clinical and research data [BioInform 05-09-05], and soon followed that with a similar agreement with the Erasmus Medical Center in the Netherlands.
Since that time, the company has been able to “productize” a number of capabilities that it developed for those institutes, Jonathan Sheldon, chief scientific officer at InforSense, told BioInform. “We’ve taken it from something that we developed that was specific to Windber and we’ve made it much more generic,” he said.
The outcome of that development process is the company’s ClinicalSense application, a web-based tool that allows users to aggregate clinical data and to identify patient cohorts with particular clinical characteristics. This week, InforSense released an updated version of ClinicalSense that includes the ability to quickly create a “data mart” from information stored in a clinical data repository.
“Once you have the analytical data mart, you can then sit ClinicalSense on top of that to do the slicing and dicing of that clinical data to identify a particular cohort,” Sheldon said.
“People have been talking about translational research for a long time,” he said. “What that typically means is very large service engagements where a team of consultants comes in and builds something very bespoke.”
InforSense has tried to move beyond this model by exploiting the knowledge it has gained from its collaborations, he said. “We’ve abstracted all the solutions that we’ve built to a level where we feel we’ve got something that is a product.”
As a result, the implementation time can be “measured in days rather than months,” said Joe Donahue, senior vice president at InforSense. Deployment of the system does require a certain amount of customization from the company’s services group, but the bulk of its functionality is available “out of the box,” he said.
Donahue added that the company is seeing “real traction” for its translational research system. “For us, it’s probably one of the most exciting parts of our product line,” he said.
He noted that the company’s initial relationship with Erasmus now spans more than 30 academic research and commercial organizations throughout the Netherlands and that the firm has recently signed another agreement for the translational research platform with an undisclosed research institute.
But while academia appears to be willing to adopt the company’s tools, pharmaceutical firms have been slow to follow. Donahue and Sheldon said that industry appears to be taking a wait-and-see approach for now.
“The big pharmas are really looking at the leading academic institutes to see how they solve this problem and what impact it could have,” Donahue said.
Sheldon said that the company is currently seeing some pharma customers doing “small-scale projects” that involve both clinical data and molecular data, “just to model the processes.”
However, he added, “pretty much without exception, all the pharma we’re talking to are in the midst of fairly detailed plans around translational research,” a trend that leads Sheldon to believe that “this year and next year are going to be sort of landmark years” for the industry.
“I think there will be some major organizational changes and major process changes with respect to the way they approach the R&D process. And I think that part of that has been learning a lot from the way biomedical institutes have exploited the clinical data,” he said.
Dana-Farber’s Approach
One institute that these pharmas will likely be keeping an eye on is Dana-Farber. The institute is working with InforSense to develop a system that integrates clinical information with gene-expression data and other experimental information, as well as data from the public domain. The goal is to enable different types of researchers at the institution — whether they be clinicians, biologists, or statisticians — to easily design experiments, access information, and analyze data that is of interest to them.
John Quackenbush, professor of biostatistics and computational biology at Dana-Farber, said that researchers at the institute have been hampered by a lack of ready access to clinical information. In one example, he said, a research group wanted to design a study that would pair clinical data from frozen patient samples with microarray data about those samples
“The big pharmas are really looking at the leading academic institutes to see how they solve this problem and what impact it could have.” |
“The problem they were facing was that all the information about the samples themselves was stored in one database, and all the clinical information was stored in another database, and they’d been trying for years to link the data and they really had no effective way of doing it,” he said.
“Essentially, the process of requesting clinical information and getting clinical information could take months, and if they … didn’t make the right request ahead of time, they’d have to repeat this process of going back and going back,” he said.
Aiming to solve this problem, Dana-Farber won a two-year, $1 million Oracle Commitment Grant to develop a data warehouse that could store all of this information securely to ensure patient confidentiality.
“But then we had another problem, and that was providing access to the information to the front-line researchers” — which is where InforSense came into the picture, Quackenbush said.
“The starting point [was] really the warehouse and what InforSense brings to the table is an approach that creates dynamic data marts that can sit on top of this data, pull in the information in a structured way, and provide very rapid access to queries that we know that users are going to want to make,” he said.
The resulting three-year agreement with InforSense covers a “pilot” study for multiple myeloma, a disease area that the institute chose “because it’s a relatively small patient population … and that actually presents some advantages because we don’t have to deal with the same quantity of data” that might be required for other types of cancers, Quackenbush said.
He said that the multiple myeloma project involves a “few hundred patients” and covers “a number of ongoing clinical trials.” The data includes clinical information, patient histories, cytogenetic analyses, and microarray gene-expression data, as well as information from public resources such as GenBank and the HapMap Project. The goal is to enable researchers to create “analytical pipelines” that will help answer key questions related to translational medicine.
For example, Quackenbush said, “imagine we look at a group of patients in this multiple myeloma trial and identify a set of genes that are associated with positive response to a certain therapeutic regimen. Can we see whether those genes map to response pathways in other diseases like breast cancer or prostate cancer or colon cancer? Do they map through PubMed to a certain class of drugs that have been discovered looking at cardiovascular disease or some other application?”
Quackenbush noted that the flexibility of the InforSense system is particularly important because “we’re not just thinking about the questions of today; we’re starting to think downstream about what kind of questions people are going to want to ask in the next few years.”
That flexibility is what sets the Dana-Farber system apart from some other informatics solutions for translational research, he said.
Indeed, there are two major IT initiatives for translational medicine right in Dana-Farber’s backyard: the Harvard Medical School-Partners HealthCare Center for Genetics and Genomics, which has been working with HP for several years to develop a comprehensive IT infrastructure for personalized medicine; and the Informatics for Integrating Biology and the Bedside program, an initiative funded under the National Institutes of Health’s National Center for Biomedical Computing program, and is led by researchers at Brigham and Women's Hospital, Children's Hospital, Harvard Medical School, and Partners HealthCare.
Quackenbush said he looked at these systems before opting for the Oracle/InforSense system, but said he “really didn’t think that what they had in place at present was going to meet all of our research needs right away.”
Another option, the suite of tools being developed under the National Cancer Institute’s Cancer Biomedical Informatics Grid project, was also inadequate for the institute’s requirements, he said. “If you didn’t have a lot of legacy systems in place, you might be able to take a lot of their tools and build a new system, but given the fact we have a lot of legacy data and legacy systems, it really didn’t appear to be all that flexible,” he said.
Nevertheless, he acknowledged that the field of translational research informatics is still in its early days. “I think what you’re going to see over the years is people trying lots of different things, and looking at different approaches to address this problem,” he said.
“I won’t claim today that ours is going to be the best solution, but I think it has a lot of things going for it and I actually believe it has the potential to have a pretty big impact in the overall conduct of clinical and translational research.”