“At every academic institution, research data is a separate universe from clinical data,” MD Anderson President Ronald DePinho, M.D., says. “Clinical data are trapped in medical records and research data in different systems throughout an institution. “We’re transforming how we do things in the future by integrating research and clinical data in one system.”
Academic medical institutions generate enormous amounts of research and clinical information. But making that data accessible to researchers has been a problem.
“We do a good job of capturing data, but it ends up being vaulted, padlocked and imprisoned in different ways,” says Andy Futreal, Ph.D., ad interim chair of Genomic Medicine. “What we’re trying to do with our big data interchange is liberate that data into a secure space where we can put it all together and learn from it.”
Data from 167,000 patients
As co-leader of the big data platform of MD Anderson’s Moon Shots Program, Futreal and a team of experts have worked to build such an interchange since 2012. Years of planning led to construction of the system architecture in 2015 and the initial loading of clinical data from 167,000 patients treated since Jan. 1, 2012.
Research data from two other moon shots platforms — next-generation tumor genetic analysis by the Cancer Genomics Laboratory and immune cell monitoring data from blood and tumor samples gathered by the immunotherapy platform — are being added to the big data interchange. This will expand to include other research information over time, Futreal says.
Most research data now are scattered in various MD Anderson databases, in the computer hard drives of individual researchers, and in their papers published in academic journals.
Detailed clinical information found in patients’ medical records — treatments, side effects, pathology, test results, notes by oncologists, surgeons, radiation oncologists, images — will continue to live in MD Anderson’s clinical electronic health record (EHR) system while also being routed into the big data interchange.
Security and compliance experts have been engaged in the project since day one to build safeguards into the system.
More questions, faster answers
Researchers will be able to query the system, Futreal notes, “to facilitate and accelerate access to data and avoid the laborious searching, hunting, gathering” now required to obtain information.
For example, a laboratory scientist studying a specific molecular pathway could search the database for a specific mutation across various cancer types.
Or an oncologist might want to look at outcomes for all patients with a specific type of breast cancer who were treated with a specific regimen during a specific time frame.
Futreal is an expert in cancer genomics and led the team that discovered the BRAF mutation found in half of all melanoma cases. Patients with this mutation are now treated with targeted therapy that attacks cancer cells without harming normal cells.
But the genomic diversity of tumors is what makes them so hard to treat in advanced stages, Futreal says. When a targeted therapy knocks down cells dependent on one mutation, surviving cells that lack that mutation — or ones that have acquired new ones — come back with a vengeance.
“Because every cancer is different, we need to build this infrastructure to learn as much as possible about every patient we care for. One treatment can produce 10 different outcomes among 10 patients,” Futreal says. “The knowledge for how to do it better next time lies in those individual patients. Until we build systems that capture data properly, integrate it, and empower the analysis of it, we’re missing a trick — an important one.”