WASHINGTON, DC—At a time of increasing data overload, cancer researchers need all the help they can get to collect, collate, integrate, analyze, and share biomedical data in order to translate important new molecular and genomic findings into better patient care. So said speakers at a workshop here sponsored by the National Cancer Policy Forum (NCPF) of the Institute of Medicine, who came together to discuss the need for a U.S. cancer informatics system of excellence to accelerate research; the existing gaps and challenges; and create a vision for transforming the cancer informatics enterprise for the 21st century.
“We are truly poised on the edge of a scientific revolution, and in the informatics field, there are billions of dollars being invested and it couldn't be more important to get it right,” said the Chair of the Workshop Planning Committee, Sharon B. Murphy, MD, IOM Scholar-in-Residence and an NCPF staff member.
While the field of cancer research informatics is highly complicated, “we need to embrace the complexity,” added Amy P. Abernethy, MD, Associate Professor of Medicine in the Division of Medical Oncology at Duke University School of Medicine and co-Vice Chair of the Planning Committee.
In 2004 NCI launched caBIG, the Cancer Biomedical Informatics Grid to serve as a national framework for data exchange and speed research findings to the clinic. Improved harmonization of data across cancer centers is regarded as one of the successes of that effort, but caBIG has not lived up to its promise; it has been hampered by reduced funding and other problems and currently has no senior staff leader.
A search is now on for an NCI senior staff leader for informatics who will develop a new NCI informatics strategy, said Daniel R. Masys, MD, Chair of the caBIG Oversight Subcommittee and Affiliate Professor in the Department of Biomedical Informatics and Medical Education at the University of Washington.
There is a growing recognition that caBIG can't do it all. Among its problems, he said, were: a “cart-before the horse grand vision;” a “one-size-fits-all approach;” and an “unsustainable business model for both NCI and users.”
And, he said, while NCI's commitment to a computational research infrastructure remains strong, “in an era of tightly constrained resources, the focus will be on the highest priority needs of science.”
To focus the IOM workshop on a new vision for informatics in cancer research, several NCPF members and their coauthors circulated a draft discussion paper at the meeting titled “Achieving Data Liquidity in the Cancer Community: Proposal for a Coalition of All Stakeholders” (see box). As stated in the wide-ranging discussion paper, “At this critical juncture, biomedicine has no systematic means for rapid and efficient exchange of data so that it can be leveraged and converted into knowledge by anyone other than the original author.” The proposed nonprofit coalition would be a facilitator in that systematic data exchange.
The vision in the discussion paper is driven by unprecedented gains in scientific knowledge. “In 10 years each individual will be surrounded by a virtual cloud of billions of data points—P4 medicine,” said keynote speaker Leroy Hood, MD, PhD, President and co-founder of the Institute for Systems Biology in Seattle.
“Predictive, Preventive, Personalized, and Participatory”
Defining P4 medicine as “predictive, preventive, personalized, and participatory,” Hood predicted that in a decade an individual human genome will be a part of every patient's record; that in five years, sequencing a human genome will cost about $100; that human genomes will be sequenced in families; and that societal and scientific objections to sequencing individual human genomes will fade, “driven by actionable gene variants” which can be treated with success.
“Rapid advances in technology are resulting in a dramatic increase in the output of genomic and molecular data related to cancer biology,” said another member of the planning committee and a coauthor of the draft discussion paper, Lawrence N. Shulman, MD, Chief Medical Officer and Senior Vice-President for Medical Affairs and Chief of the Division of General Oncology at Dana-Farber Cancer Institute.
“The sheer volume of emerging data is more than any us can keep track of…or make appropriate or interesting connections between. Clinical medicine is becoming increasingly more complex due to the number of new data points required to make clinical decisions, and the number of new drugs available.”
The most successful cancer programs will “inter-connect research and clinical activities and data, in an organized and efficient manner, with as broad a database as possible,” he continued.
For this to happen, cancer researchers will need effective new informatics tools. As an example, he cited SPARKS (Synergistic Patient and Research Knowledge Systems), a collaborative institutional informatics framework to accelerate scientific discoveries into clinical practice. The ultimate goal of SPARKS is to enable early diagnosis, personalized treatment, cure, and the prevention of cancer and related diseases.
Shulman put in a plea for robust electronic health records (EHRs) that codify clinical, genomic, and molecular data in an “inter-operable, standard-adherent” way; such EHRs will be central to a good cancer research informatics enterprise, he said. But unfortunately, he said, “We are not close to this aspiration.”
“Mind-Boggling, But Exciting”
Asked by OT how oncologists can best navigate the increasingly challenging world of cancer research informatics in clinical practice, workshop panelist Allen S. Lichter, MD, Chief Executive Officer of the American Society of Clinical Oncology, said that from the point of view of practicing oncologists, “you're going to need all kinds of support services to find your way through this. A whole set of analytic tools will be needed; ASCO will have to be part of this new world. There will be a host of support services that we can't even dream of right now. It is mind-boggling, but exciting.”
He predicted that hand-held digital computational devices of the future “will have incredible power,” and that there will be new applications available to oncologists to help them use accumulating research data in clinical practice.
Last November ASCO released a sweeping report, “Accelerating Progress Against Cancer: ASCO's Blueprint for Transforming Clinical and Translational Research,” which cites the importance of employing EHRs and rapid learning systems to use real-world scientific findings for better cancer treatment in clinical care (OT, 12/10/11 issue).
What is needed in a world of massively accumulating data is “a partnership among researchers, clinicians, regulators, policy makers, and patients to design an integrated information network system,” said another member of the planning workshop and a coauthor of the draft discussion paper, William S. Dalton, PhD, MD, President, CEO, and Center Director of the H. Lee Moffitt Cancer Center & Research Institute and an NCPF member.
Dalton, who in July will become CEO of Moffitt's new Personalized Medicine Institute, described the research and health care network model implemented at Moffitt in 2006, which uses the Total Cancer Care (TCC) Protocol to give each patient an individualized approach to care. Moffitt serves as the coordinating site for the TCC consortium, which includes 17 US community hospitals.
“Expansion of consortium sites will encourage information exchange,” he said. Patients in the TCC Protocol agree to be followed throughout their lifetimes; to have their tumor studied using molecular technology; and to allow researchers to re-contact them if needed. To date, more than 76,000 cancer patients from Moffitt and its consortium members have been enrolled in the TCC protocol.
“I see this as the hub-and-spoke model,” he said—i.e., a regional network in which a cancer center like Moffitt serves as the hub and community hospitals are the spokes. “It's got to be a grassroots effort. You can't just superimpose it [a system of cancer research informatics] from the top.”
Draft Discussion Paper
The draft discussion paper circulated at the workshop recommends launching a new nonprofit information-based biomedical membership organization made up of stakeholders with the broadest possible representation. The coalition would augment the roles of groups such as ASCO and AACR in defining critical cancer data to be collected. The proposed coalition would:
- Serve as a catalyst for and help a community to develop and make available, pre-competitively, an open digital framework for biomedicine;
- Convene the cancer community to seek consensus on strategies for pre-competitive data sharing;
- Advocate for the adoption of standards for data collection;
- Ensure that the open digital framework stays current with all technological advances;
- Ensure that all biomedical organizations have access to the open digital framework, so that they can improve patient care and enhance their research;
- Serve as an honest broker within the biomedical ecosystem, linking those who need to use digital frameworks with those who provide them;
- Help to support a flourishing ecosystem of biomedical organizations that can fuel each other's activities “through frictionless flow of data;” and
- Serve as a test bed for the digital infrastructure recommended at a previous IOM workshop.
Read this article on our iPad edition to hear the opening remarks from National Cancer Policy Forum Chair John Mendelsohn, MD, and Workshop Planning Committee Chair Sharon Murphy, MD, as well as to see a copy of the full two-day agenda.
To receive our iPad issues, download the free app todayhttp://bit.ly/OT-iPadApp!