[2069] Comparison of Breast Cancer Proliferation Rates between Core Biopsy and Surgical Excision Using Visual Mitotic Count and Automated Analysis of MIB-1 and PPH3

Ryan McCormick, Cynthia Cohen, Meriah Schoen, Jason Wang, Amy Adams. Emory University, Atlanta, GA

Background: Moderate correlation in the histologic grade of breast cancer between core biopsy (CB) and corresponding surgical excision (SE) has been demonstrated, the difference often due to underestimation of mitotic count in the CB. Studies have shown better concordance with regard to proliferation rate between CB and SE using visual assessment of the immunohistochemical markers MIB-1 and PPH3. Using visual mitotic count (VMC) on SE as the gold standard, our aim is to assess whether automated image analysis of MIB-1 and PPH3 yields improved agreement with regard to proliferative rate.
Design: All breast cancer cases with available material from both CB and SE specimens were identified over a 14 month period. Cases from patients receiving neoadjuvant therapy were excluded. A VMC was performed on H&E sections for both the CB and SE, using cutoffs based on CAP protocol for microscope field diameter. Slides from the CB and SE specimens were also stained for MIB-1 and PPH3. To determine the average percentage positivity for MIB-1 and PPH3, 3 and 10 random fields, respectively, were assessed using an Automated Cellular Imaging System III (DAKO Corp) from “hot spot” fields identified at 20x. MIB-1 staining was categorized as low (<10%), intermediate (10-20%), or high (>20%). PPH3 positivity was categorized as low (<2%), intermediate (2% to 5%), or high (>5%). Results for VMC (on CB), MIB-1, and PPH3 were compared to the gold standard of VMC on SE.
Results: 28 cases were identified with both a CB and corresponding SE. Agreement between the methods is illustrated in Table 1.

Table 1. Agreement Between Various Methods of Assessing Proliferation Compared to the Visual Mitotic Count on Surgical Excision
Weighted KappaCBSE
VMC=visual mitotic count; SE=surgical excision; CB=core biopsy

Conclusions: Consistent with prior reports, our results show fair agreement for VMC between the CB and SE. The low rate of agreement between MIB-1 and the gold standard on both CB and SE is not all that surprising given that MIB-1 labels cells in G1, S, and G2 phases in addition to M phase; however, agreement is improved with use of the SE specimen versus the CB. We had predicted better agreement between the gold standard and PPH3-stained CB and SE sections (weighted kappa 0.033 and -0.032, respectively). The results are likely due to the cutoff values for PPH3 established for the automated imaging system. Investigation to further refine these values may yield improved agreement.
Category: Techniques

Monday, March 4, 2013 9:30 AM

Poster Session I Stowell-Orbison/Surgical Pathology/Autopsy Awards Poster Session # 310, Monday Morning


Close Window