Different methods of pretreatment Ki-67 labeling index evaluation in core biopsies of breast cancer patients treated with neoadjuvant chemotherapy and their relation to response to therapy.

Increased proliferation activity of breast cancer cells evaluated by Ki-67 immunohistochemistry, i.e. a high Ki-67 labeling index (Ki-67 LI), may predict better tumor regression in case of neoadjuvant chemotherapy. Despite recommendations for the evaluation of Ki-67 LI, there are variations in methodology. We assessed the effect of different evaluation methods on the Ki-67 LI in patients with different response to neoadjuvant chemotherapy. Thirty pretreatment core-biopsy samples of patients receiving neoadjuvant docetaxel-epirubicin chemotherapy with or without capecitabine were evaluated for their Ki-67 LI. Pathologic regression was categorized as no regression, partial regression and complete regression, with 10 cases in each category. Three antibodies (MIB1, B56, SP6), 4 observers and 4 methods (counting or estimating on glass slides and counting or estimating on representative digital images) were compared. The Kruskal-Wallis test and analyses of variance were performed to investigate the differences in Ki-67 LIs between different clinical outcomes (tumor regression categories). Breast carcinomas with pathological complete regression had a higher mean Ki-67 LI than tumors not achieving complete regression with any methods, observers and antibodies investigated, although there was a variation between different evaluations in what may represent high proliferation. Estimating the Ki-67 LI on digital images representing the highest proliferation in the core biopsy seemed the best in separating complete responders from non-responders. High Ki-67 LI values were more likely associated with pathological complete regression independently of the method of evaluation used, although the definition of high proliferation is problematic. Estimating the Ki-67 LI may be an adequate method of evaluation.