Abstract
Estrogen receptor (ER) assays were performed by sucrose gradient centrifuga-tion method at multiple sites in large breast cancers. Intersite variation of ER in a tumor was observed in 24 out of 35 cases. 16 tumors with relatively low ER levels showed different ER status with multiple assays. The results suggest that an assay performed on a small random part of a large tumor may not yield the true ER status. ER value at the largest cross-section was almost the same as the average ER values in each tumor. In addition, 21 cases were examined in relation to ER values at multiple sites in the large tumors and response to endocrine therapy. As the ER value at the largest cross-section was highly correlated with the therapeutic response to endocrine therapy of breast cancer, it would represent true ER status and level. The results suggest that the ER assay at the largest cross-section of a large tumor is an appropriate method to predict response to endocrine therapy.