Benign Breast Disease Team Project

Abbreviated Name
BBD Benign Breast
Lead Investigator
Engstrom, PaulFox Chase Cancer Center
Coordinating Investigator
Feng, Ziding Fred Hutchinson Cancer Center
Involved Investigators


To identify women diagnosed with atypical ductal hyperplasia (ADH) who are at increased risk of developing invasive breast cancer (IBC) and who might benefit from risk reduction with the use of chemoprevention agents such as Tamoxifen. (Note: A companion protocol will study women with DCIS and their risk for invasive breast cancer.)


Biomarkers expressed in benign breast disease tissue of women who progresse to invasive breast cancer are quantitatively and/or qualitatively different from those expressed in tissue of matched women who do not progress to invasive breast cancer. A nested case-control design will be used to determine if available proteomic and methylation biomarkers predict risk for future invasive cancer (IBC). Women with ADH who are disease-free will be matched 2:1 to women with ADH who progress to IBC.

Analytic Method

The performance of each biomarker tested will be undertaken in three steps: A.   Relative Risks Associated with Biomarker Values 1.    Cox regression will be used. Time measured from BBD diagnosis. 2.    Relative risks beyond those conferred by other predictors including age, family history, and histology will be examined. 3.    Separate models will be fit for ADH versus UH versus NP if there are sufficient numbers; Analyses will be combined with stratification if appropriate. B.   Capacity for Discrimination between Cases and Controls 1.   Primary comparison groups: subjects who developed invasive breast cancer by 7 years versus controls who are alive and without cancer at 7 years after BBD 2.   Secondary comparison groups: incident DCIS, invasive cancer after 7 years 3.   ROC curves will be used to compare cases with primary controls. Calculations are complicated because they must handle varying follow-up and the quota method of selecting non-cases 4.   Separate ROCs for BBD with and without atypia will be estimated if possible and compared. We will compare ROC curves for different biomarkers 5.   We will develop a combination biomarker score based on markers that appear to perform well using Cox regression. We will calculate its ROC curve and compare with ROC for best individual biomarker C.   Absolute Risk of IBC for Individual Decision Making 1.   Calculate individual risk of IBC .5-7.0 years after BBD with and without biomarker 2.   Compare risk distributions. How many people classified as high risk (>.75%X4.5) with and without the biomarker 3.   Calculate risk distributions for subjects who develop IBC by 7 years (cases) and for subjects alive without IBC by 7years (controls) 4.   Of subjects who develop IBC by 7 years, how many classified as high risk? This is the sensitivity (TPR) 5.   Of subjects who are alive without IBC by 7 years, how many classified as low risk? This is the specificity (1-FPR) 6.   Calculate the standardized net benefit that combines TPR and FPR into a single index 7.   Compare for different biomarkers and for the most discriminating biomarker combination


  • No publications available at this time for this protocol.


Data Collections

There is one collection of data that has not yet been made public. Logging in may you give you access to the names of the non-public data collections.

 Team Project
Start Date
Aug 13 2010
Estimated Finish Date
Jun 13 2013
Finish Date
Feb 18 2022
Protocol ID
Protocol Type
Reference Set
Field of Research
Collaborative Group
Breast and Gynecologic Cancers Research Group
Cancer Types
  • Malignant neoplasm of breast
Phased Status

Associated Forms