Skip to main content
An official website of the United States government

Winners Announced for the Inaugural DCEG Informatics Tool Challenge

, by DCEG Staff

Six winners have been announced for the inaugural DCEG Informatics Tool Challenge. The competitive funding challenge was conceived by DCEG Director Stephen J. Chanock, M.D. “Innovation has always been the hallmark of DCEG’s research approach,” said Dr. Chanock. “The challenge was a new approach for us, and one that I hope will help bring to bear the endless possibilities of our modern technological and informatics environment to further enhance DCEG’s research approach.”

Proposals were evaluated for their innovative use of current technologies to address a specific research need, the ability for the project to be completed within one year of initiation, and the cost, which was not to exceed $20,000.

NCI’s Center for Biomedical Informatics and Information Technology (CBIIT) is providing technical support for three of the projects, with contractors or collaborators providing support for the other three projects.

Out of ten proposals that were submitted, the six described below were selected for funding:

Cancer Mortality Maps – A major update

David P. Check, Susan S. Devesa, Ph.D., Joseph F. Fraumeni, Jr., M.D., Carl McCabe, Ph.D. (DCEG); Ruth Parsons, Dave Hacker, James Cucinelli, Jeremy Lyman, Tim McNeel (IMS); Zaria Tatalovich (DCCPS)

The Cancer Mortality Maps (CMM) website uses mortality data from the National Center for Health Statistics and population data from the U.S. Census to enable researchers to generate maps of cancer mortality. Cancer maps have led researchers to conduct informative in-depth epidemiologic studies of geographic areas with unusual patterns in cancer mortality. In coordination with the NCI’s Division of Cancer Control and Population Sciences (DCCPS), this project will merge the CMM website’s long-term historical data with DCCPS’s GeoViewer map application, improving on the CMM site’s graphics and functionality.

Online risk assessment tool for lung cancer screening

Hormuzd A. Katki, Ph.D., Anil K. Chaturvedi, Ph.D., Carl McCabe, Ph.D. (DCEG); Robert Shirley, Sue Pan (CBITT); William Klein, Gila Neta (DCCPS); Christine Berg (Outside NIH)

Following results of the National Lung Screening Trial (NLST) and the US Preventive Services Task Force (USPSTF) recommendations on screening for lung cancer, the  proposal aims are to develop an interactive tool for researchers to estimate a patient’s risk of death from lung cancer, and potential benefit or harm from an NLST-like computed tomography (CT) screening program, based on the patient’s demographic information and smoking history. As a scientific work-in-progress, the tool would be intended for use by other researchers and potentially by health policy groups evaluating risk-based CT lung screening guidelines.

A tool to calculate organ doses for patients undergoing radiographic or fluoroscopic examinations

Choonsik Lee, Ph.D. and Steven L. Simon, Ph.D. (DCEG)

In this project, researchers develop a computer program to calculate radiation doses to body organs in pediatric and adult patients undergoing radiographic examinations or fluoroscopically-guided procedures. Radiation exposures from medical sources are the largest and fastest growing source of radiation to the U.S. population. The proposed tool will be a critical asset for researchers studying cancer risk from these medical procedures. Initially the program will be used by DCEG dosimetrists on epidemiological studies in the U.S. Radiologic Technologists cohort, and later disseminated to medical physicists and radiologists in clinical practice.

LD Link: A tool to simplify the task of finding correlated alleles of SNPs in high linkage disequalibrium

Mitchell Machiela, Sc.D. M.P.H (DCEG); Robert Shirley, Sue Pan (CBIIT)

Fine mapping and functional studies are widely employed by DCEG investigators to identify causal variants from genome-wide association studies hits. Several tools exist to help investigators select single nucleotide polymorphism (SNP) markers for highly correlated regions of linkage disequilibrium (LD) in the human genome; however, no simple online interface is available for identifying alleles that are commonly inherited together when two SNPs are in high LD. The current approach is time-consuming and requires programming expertise and substantial computational resources. LD Link will simplify this task while saving time and computational resources.

CrossTalk: A web tool for comparative age-period-cohort analysis

Philip S. Rosenberg, Ph.D., David P. Check, William F. Anderson, M.D., M.P.H. (DCEG); Robert Shirley, Sue Pan (CBIIT)

The web tool will be used to conduct comparative age-period-cohort analysis of “two-hazard” problems, that is to say, a pair of hazards, each with its own set of age-period-cohort parameters. An example would be the same cancer in different populations. The tool as envisioned could be used for hypothesis testing as well as rate-ratio estimation. This tool builds on the NCI Age Period Cohort Analysis Web Tool recently developed in collaboration with experts in CBIIT (manuscript in review at CEBP). Significant efficiencies in development of the comparative tool will be achieved because much of the web code and features will be based on code that was previously developed for the first tool. (Update: Rosenberg PS, Check DP, Anderson WF. A web tool for age-period-cohort analysis of cancer incidence and mortality rates. Cancer Epidemiol Biomarkers Prev 2014 Nov;23:2296-2302. E-pub Aug 2014)

A statistical tool for Poisson regression analysis of data from case-cohort studies

Sholom Wacholder, Ph.D., Orestis Panagiotou, Ph.D., Amanda Black, Ph.D., Robert N. Hoover, M.D., Sc.D. (DCEG); Y Li, D Li (U of Maryland)

To enable more routine use of case-cohort design, DCEG investigators will develop a tool for Poisson regression analysis of data from case-cohort studies. Case-cohort design offers cost savings by measuring biomarkers on a subcohort, or random sample of individuals from a cohort, and all the cases as opposed to measuring biomarkers on a full cohort. Using case-cohort design with Poisson regression analysis also offers the ability to estimate absolute risk of disease with the added benefit of allowing for multiple complex time variables. The proposed tool will be created in the R programming language commonly used by statisticians.

Proposals were reviewed for technical feasibility by Robert Shirley (CBIIT) and for utility to epidemiologic and genetic research by a DCEG review committee, which included Stephen Chanock, Robert Hoover, Geoffrey Tobias, Margaret Tucker, and Shelia Zahm.

< Older Post

New genetic risk variants identified in multiethnic analysis of prostate cancer

Newer Post >

Gadalla and Trabert Selected as 2014 NIH Earl Stadtman Investigators