Comparative Analysis of Four Methods of MIB-1 Quantitation in Estrogen Receptor-Positive Primary Breast Carcinomas
MD Darrow, C Chisolm, D Williams, B Chastain, A Page, A Adams, C Cohen. Emory University, Atlanta, GA
Background: MIB-1, a proliferation marker, is one of the weighted parameters in the twenty-one gene Oncotype Breast Cancer Assay, which quantifies the likelihood of recurrence and assesses the potential benefit of chemotherapy. It is also included in several ASCO and NCCN treatment protocols. Thus, accurate quantitation of the MIB-1 labeling index is essential. The aim of this study is to compare MIB-1 proliferation indices of ER-positive primary breast carcinomas generated by different image cytometric techniques to assess consistency and accuracy of results.
Design: Estrogen receptor-positive primary breast carcinomas from 189 patients with Oncotype DX Recurrence Scores were examined. Using the ACIS III Image Cytometer (Dako), five micron MIB-1 immunostained slides were digitally scanned and assayed for nuclear staining in three 20X "hotspot" fields (recommended by DAKO), and in five, ten (most accurate with the CAS200 image analyzer), and fifteen random 20X fields. Results by these four techniques were compared using the Pearson's correlation coefficient.
Results: When compared to the MIB-1 proliferative indices generated by use of the three 20X "hotspot" field technique recommended by DAKO, the values obtained by the other three techniques all exhibited strong, statistically significant correlation (p <0.001), with correlation coefficients ranging from 0.95 for the 15-field technique to 0.98 for the 5-field technique. In addition, the indices determined by using the random 5-, 10-, and 15-field methods, when compared to each other, showed very strong correlation (p<0.001), with correlation coefficients ranging from 0.976 (5-fields vs. 15-fields) to 0.988 (10-fields vs. 15-fields).
Conclusions: In ER-positive primary breast carcinomas, generation of MIB-1 proliferative indices using ACIS III Image Cytometry in three "hotspots", and in five, ten, and fifteen random 20X fields yields very similar results with minimal variation. This finding suggests that, unlike HER2 where quantitation in 10 random fields is optimal, use of any of these four methods in determining MIB-1 values is equivalent and that technique variation in various immunohistochemistry laboratories will not significantly impact resultant MIB-1 indices and subsequent carcinoma stratification.
Monday, March 22, 2010 1:00 PM
Poster Session II # 256, Monday Afternoon