The Breast Cancer Surveillance Consortium (BCSC) database contains data from 1994-2009 on breast cancer screening, quality, and outcomes, and its main goal is to advance breast cancer research. This project is funded to maintain the BCSC database and share the data with investigators. Our objectives are to:

• Maintain the BCSC pooled database using secure methods that protect the confidentiality of women, radiologists, facilities, and BCSC registries;
• Encourage and facilitate sharing of BCSC research resources, and continue developing and implementing our communications plan to expand resource dissemination;
• Provide statistical and scientific expertise to researchers, distribute customized limited datasets, and analyze BCSC data;
• Develop and support a new web-based dynamic query tool and other electronic interfaces to efficiently share aggregated BCSC data with the public.

The BCSC is a collaboration of breast imaging registries formally established in 1994 with support from the National Cancer Institute (NCI). As the BCSC SCC, we coordinate standardized data collection, pool data from the registries, develop statistical methodology, conduct statistical analyses for research projects, and share data with within and beyond the consortium. The current BCSC database consists of data collected by five active mammography registries and two inactive registries:

• Carolina Mammography Registry in North Carolina (CMR)
• Group Health Registry in Seattle, Washington (GHC)
• New Hampshire Mammography Network (NHMN)
• San Francisco Mammography Registry (SFMR)
• Vermont Breast Cancer Surveillance System (VBCSS)
• New Mexico Mammography Project in Albuquerque (NMMP)
• Colorado Mammography Advocacy Project (CMAP)

These registries have been actively collecting data since 1994 or 1996 (depending on registry) including patient risk factors, clinical history, breast-imaging procedures (screening and diagnostic), benign and malignant pathology results, cancer outcomes, vital status, radiologist characteristics, and facility information. Some registries have data going back to the 1970s; however, this information is mostly used to determine prior breast cancer diagnoses and mammography exposure. Most analysis limit data to 1994 forward.

Registry procedures:

BCSC data will be shared using 4 main approaches:

1. Completed deidentified datasets and summary statistics made publicly available on the internet.
2. A web-based data query tool for automatically creating user-specified completely deidentified data sets and summary statistics.
3. Customized data sets created by the Statistical Coordinating Center (SCC) and delivered to investigators using secure methods, after BCSC steering committee approval of a research proposal.
4. Statistical analyses performed by the SCC, after BCSC steering committee approval of a research proposal.

Quality factors:

The data used by the BCSC have been previously collected directly from each BCSC registry. The BCSC database has no personal identifiers such as names, phone numbers, addresses, or social security numbers and no one in the population will be contacted for additional data collection. The data set only contains HIPAA limited identifiers: dates, ages >89, and ZIP codes.

It is possible that new data will be collected under this contract as determined by the BCSC Steering Committee. This could include corrections to the current data, updates to existing data elements (e.g., additional radiology data, updated cancer and vital status outcomes), or new data elements (i.e., information about advanced imaging modalities, cancer biomarkers). New data will not contain personal identifiers. Individual-level data will never be shared unless rigorously examined and approved by the Steering Committee (which will include the principal investigators from each BCSC site) and the appropriate IRB safeguards are in place.

Our major objective for this contract is to maintain the current high standards of the BCSC database with no interruption in service to the scientific community. We will continue to be a resource to the research community, providing scientific and statistical expertise and conducting high-quality statistical analyses to advance breast cancer research. We will expand outreach and dissemination, so that more investigators are aware of this valuable resource and use it for their research. Our second objective is to develop new electronic interfaces and publicly available datasets that meet the highest possible standards for caBIG® compatibility. This will facilitate efficient sharing of BCSC data and expand potential for collaboration, increasing its value to the broader research community. Our ultimate aim is to facilitate research that improves breast cancer screening and reduces cancer burden and mortality.

Inclusion Criteria:

• 18 years of age

Exclusion Criteria:

• none

The BCSC pooled database includes data from seven BCSC registries, of which five are currently active. Information comes from a large study population that is ethnically, geographically, and socioeconomically diverse. The database currently includes 9 million mammograms from 2.3 million women interpreted by over 1400 radiologists at 400 facilities. There are 1.1 million digital mammograms; 660,000 ultrasounds; 180,000 benign and malignant biopsies; and records of over 155,000 deaths. Over 107,000 women have a breast cancer diagnosis, of which 83% are first diagnosed in 1994 or later. The data are collected in the course of routine, clinical care to accurately represent community practice.