Gretchen Williams, Rong Hu, Rulla Tamimi, Laura Collins. Beth Israel Deaconess Medical Center, Boston; Channing Laboratory, Brigham and Women's Hospital, Boston
Background: The distribution of ER staining in breast cancers by immunohistochemistry (IHC) has been shown in some studies to be largely bi-modal with tumors being either negative or strongly and diffusely positive, which is likely a function of the methodology used. When evaluated using continuous measures such as rtPCR, ER expression shows a normal distribution. Therefore, we sought to evaluate the proportion of tumor cells demonstrating ER positivity in a large cohort of breast cancers using image analysis as a means of accurately quantifying tumor cell positivity.
Design: Tissue microarray (TMA) sections of breast cancers from women enrolled in the Nurses' Health Study were immunostained for ER. Slides were then scanned using ScanScope slide scanning system (Aperio Technologies Inc., Vista, CA) and ER staining was quantified using a specific algorithm, Nuclear v9, for nuclear size, intensity, roundness, curvature, compactness, and elongation using Aperio Image scope system. For this portion of the analysis, tumor cell positivity from the TMA core with the greatest ER positivity was related to tumor grade and menopausal status.
Results: Image analysis data of ER was available for 708 breast cancers. The mean percent of tumor cell ER positivity was 46%, (median=42%, max=99.6%) with an even distribution of positivity seen from <1 to 100% (see figure).
Results of ER positivity according to tumor grade and menopausal status are shown in the table.