Prognostic factor Ki67 for breast cancer patients in each subgroup.

Abstract

565 Background: Immunohistochemical (IHC) Ki67 has described it as a prognostic and predictive marker for breast cancer. The St. Gallen Consensus Meeting determined that Ki67 labeling index is chiefly important for distinguishing between “Luminal A” and “Luminal B (HER2 negative)” subtypes and is a predictive marker for chemotherapeutic efficacy. However, the high and low cutoff points remain controversial. Our objective is to compare survival in patients with low, intermediate, and high Ki67 levels in each subgroup. Methods: We retrospectively identified all the patients in the Tokai University breast cancer database for whom IHC Ki67 data were available between January 1, 2000, and December 31, 2010. Ki67 was defined as low if <10% Ki67 was detected, as Intermediate if 10–20% Ki67 was detected, and as high if >20% Ki67 was detected. To assess Ki67 levels and survival outcomes, survival curves were calculated using the Kaplan–Meier method and compared using the log-rank test. Results: We identified 1331 primary breast cancer patients without metastasis, of whom 686 received neoadjuvant or adjuvant chemotherapy. Patients with high Ki67 had poorer relapse-free survival (RFS) than patients with intermediate (p = 0.009) and low Ki67 (p < 0.001). Patients with intermediate Ki67 had poorer RFS than patients with low Ki67 (p < 0.001). In ER-positive cases (n = 1059), patients with high and intermediate Ki67 had poorer RFS than patients with low Ki67 (p < 0.001 and p = 0.002, respectively). In HER2-positive and ER-negative cases (n = 103), patients with high Ki67 had poorer RFS than patients with low Ki67 (p = 0.002). In triple-negative cases (n = 164), patients with high Ki67 tended to have poorer RFS than patients with low Ki67 (p = 0.064). Conclusions: Our data demonstrated that low, intermediate, and high Ki67 levels may be used to differentiate prognosis in ER-positive cancer patients as well as HER2-positive and triple-negative cancer patients.