Comparison of methods for detection of fulvestrant-induced changes in breast tumor estrogen and progesterone receptor expression in a neoadjuvant trial (NEWEST)

Background: Fulvestrant downregulates breast tumor estrogen and progesterone receptor (ER and PgR) levels in a dose-dependent manner. Using an Automated Cellular Imaging System (ACIS), NEWEST reported that 4-wks’ treatment with fulvestrant high-dose (HD, 500mg/month+500mg on Day 14 of Month 1) significantly reduced ER levels in primary breast tumors v approved-dose (AD, 250mg/month). However, no significant difference was detected between the two doses on PgR levels. To allow comparison with previous studies, a non-automated H-score assessment was performed and compared with ACIS. Methods: ER and PgR H-scores were derived by manual assessment of % tumor cells in each of 5 staining categories (negative, very weak, weak, moderate, strong) in the same sections scored by ACIS. This microscopic assessment was performed by 2 experienced observers blind to ACIS and clinical data. Mean % changes in H-scores were then calculated (table). Results: Both scoring methods showed a greater effect for fulvestrant HD v AD on ER expression at Wk 4, but H-score provided better dose discrimination. ACIS showed no difference between fulvestrant HD v AD on PgR expression at Wk 4, however, a significantly greater reduction in PgR expression was detected with fulvestrant HD using H-score. Conclusions: The choice of scoring method for determining ER and PgR expression in pharmacodynamic studies such as NEWEST is critical. Compared with H-scores, ACIS has a narrower dynamic range and reduced ability to discriminate fulvestrant HD v AD, particularly on PgR expression.