The goal of the EIF/WCRF Biomarker Discovery Project is to develop biomarkers (proteins) that reveal the presence of breast cancer in a blood test and then to determine which are effective markers for early disease detection, prediction of metastatic potential or therapeutic response. Effective biomarkers for any one of these applications would save lives. Many women with early stage disease are currently missed by mammography if they have dense breasts. Some women with node negative disease have undetected metastatic disease and many women without metastatic disease are over treated with chemotherapy because they cannot be adequately diagnosed. Finally, the development of more effective therapeutics for breast cancer would be empowered by biomarkers that reveal which women in a clinical trial are responding to a new therapy. Success in this project would not only improve the outcomes for breast cancer patients but the approaches developed would be directly applicable to other cancer sites as well as other types of human disease.

This project is formulated with the belief that recent developments in technology together with the information from the human genome project permit a major advance in biomarker discovery. However, it is not clear at the present time exactly how to best employ new technologies in biomarker discovery and an aspect of this project is to uncover that most effective path. We will tap the enormous combined expertise of world leaders in the fields of proteomics, informatics, and clinical breast cancer care. Identification of biomarkers for breast cancer is a significant and ambitious goal which requires not only the application of outstanding scientific expertise, but also careful attention to scientific oversight and management so that the collective progress of multiple laboratories is coordinated and synchronized for maximum effectiveness. This project brings together all these components in a dynamic, interactive consortium environment to facilitate success. Consortium members will examine the same biological materials, share results and compare different approaches to the problem. We will ensure tight cohesiveness among the consortium members by an emphasis on frequent communication (including regular conference calls and meetings) and by strong project leadership (exercised by an active Steering Committee). We will adjust the organization of the project and the distribution of funds, in accordance with needs and opportunities as the project progresses. In short, the Consortium is squarely focused on achieving its core mission.

Consortium members with expertise in proteomics include Dr. Ruedi Abersold at the Institute for Systems Biology, Dr. Dick Smith and the Pacific Northwest National Laboratories, Dr. Sam Hanash at the Fred Hutchinson Cancer Research Center and Dr. Eric Lander at the Broad Institute. Clinical experts are Dr. Helena Chang at UCLA, Dr. Larry Norton at Memorial Sloan Kettering, and Dr. Gabriel Hortobagyi at MD Anderson. Informatics support will be led by Dr. Martin McIntosh at the Fred Hutchinson Cancer Research Center with additional personnel from each of the proteomics sites. This will be the first study of its kind conducted across expert laboratories sharing methodologies, data and samples focused on identifying the most effective path to breast cancer biomarker discovery.

The basic strategy of the core project will be to employ two discovery modes. One is a candidate approach which first identifies the most abundant and disease-specific proteins in the tumor or its proximal fluids and then compares the blood of patients to normal individuals for the presence of these candidates. Since the candidates are likely to be present in blood at very low concentration we will employ a suite of techniques to identify very low abundance proteins suspected of being useful diagnostic agents. A second approach is to compare directly the blood of cancer patients with that of normal controls using unique instrumentation available only to the consortium members. Any markers found to be diagnostic of cancer will be identified and quantitated. In addition we will support a series of pilot projects, each of which has promise to provide new insights to biomarker discovery.

Not only do we want to assess technology performance by comparing data across platforms, but we must also aggregate data from diverse sources and from samples prepared in different ways so that a significant fraction of the proteome can be monitored. It is these dual needs for assessment of technology performance and aggregation of data from diverse sources that will dictate the characteristics of an effective program. New tools must be built for data sharing, standards held fast for data collection, the proteome landscape and sample types must be divided up and assigned to different investigators, and the resulting "features" must be identified and quantified. Since improvements and additions are likely to accrue from many sources, the tools and database must be open source and publicly available, and effective collaborations with industry and intramural partners sought. Realizing that ultimate value will come from commercialization of platforms and technology, the whole enterprise must be undertaken with a view to the needs for regulatory approval and commercial adaptation.