Mitotic Grade and MIB-1 Are Not Redundant Measures of Proliferation in Breast Cancer
J Carter, J Hanson, R Greiner, N Asgarian, J Deschenes, J Mackey, J Hugh. University of Alberta, Edmonton, AB, Canada
Background: In breast cancer, the proliferation rate measured by MIB-1 immunohistochemistry allows sub-typing of estrogen receptor positive breast cancers that closely approximates genetic-based tests and predicts response to therapy. Whether MIB-1 data complements or replaces traditional light-microscopic assessments of mitotic grade (MG) is not established. We evaluated the interaction between MIB-1 and MG using disease-free survival as the outcome variable.
Design: Tumor samples were available for 1329 patients (89%) randomized to the Breast Cancer International Research Group 001 phase III trial comparing docetaxel, doxorubicin and cyclophosphamide (TAC) with fluorouracil, doxorubicin, and cyclophosphamide (FAC) in women with operable, node-positive breast cancer. Disease-free survival was assessed at a median follow-up of 55 months. Both the MIB-1 index and MG were assessed by a single observer blinded to treatment and outcome.
Results: Tumors were classified MG 1, 2, or 3 (52%, 25%, and 22%, respectively). The distribution of MIB1 scores was mathematically modeled for two normal populations and the best discriminatory cut-point of greater than or equal to 30% positively staining cells was adopted. This defined MIB-1Lo (49%) and MIB-1Hi (51%) tumors. Both MG and MIB-1 were significant prognostic indicators (univariate analysis) and highly correlated (Spearman correlation coefficient, p<0.0001). 59% of cases were perfectly concordant as MG1 and MIB-1Lo (n=507) or MG3 and MIB-1Hi (n=276). Only 2% (n=22) were MG3 and MIB-1Lo and excluded from further analysis. Using the most concordant group (MG3, MIB-1Hi) as the basis for comparison, there was a statistically significant interaction between MG and MIB-1. MIB-1Lo and MIB-1Hi defined significantly different prognostic groups within the MG1 and MG2 tumors (p=0.0004 and p=0.007 respectively). Similarly, MG divided the MIB-1Hi population into different prognostic groups (p=0.006). In a pairwise comparison, patients with MG3 tumors had a relative risk of relapse of 1.68 compared with MG1 (p=0.0019).
Conclusions: MG is not redundant with MIB-1 and gives complementary prognostic information. There is a possibility that the two markers may be integrated to provide cell cycle information that predicts chemotherapy response.
Wednesday, March 24, 2010 1:00 PM
Poster Session VI # 21, Wednesday Afternoon