New insights into the prognostic value of Ki-67 labeling index in patients with triple-negative breast cancer

Abstract

The clinicopathological importance of the Ki-67 labeling index (LI) in breast cancer has been studied intensely; however, its prognostic significance in triplenegative breast cancer (TNBC) is unclear. We aimed to determine the optimal Ki-67 cut-off point to demonstrate its prognostic relevance for breast-cancerspecific survival (BCSS) in TNBC patients. A total of 571 female TNBC patients underwent diagnosis and surgery at our institution from January 2002 to June 2011. Clinicopathological information for all patients was available and categorized by Ki-67 LI and age at diagnosis. The cut-off values for Ki-67 LI and age were selected using the medians. A varying-coefficient Cox model was used to describe the effect of Ki-67 LI on BCSS outcomes changing with age after adjustment for disease characteristics. For survival analysis, the Kaplan-Meier method and the log-rank test were used. Cox proportional hazards models were applied to determine the association of Ki-67 LI and age with BCSS outcomes after adjustment for disease characteristics. Median age was 50 years, and median Ki-67 LI was 35% (range, 0 -97.5%). There was no prognostic significance of stratification by Ki-67 LI in all patients. When analyzing age at diagnosis as a continuous variable, the log-transformed HRKi67 > 35% vs. ≤ 35% for BCSS increased in an S-shaped curve with increasing age up to about 50 years-old and remained higher-risk for high Ki-67 LI. After adjusting for clinicopathological risk factors, low Ki-67 LI was a poor prognostic factor for BCSS (HR: 0.36, 95% CI: 0.14-0.96, P = 0.042) in patients of ≤ 50 years, but not in patients diagnosed at > 50 years (hazard ratio [HR]: 1.57, 95% CI: 0.76-3.22, P = 0.241). In conclusion, lower Ki-67 LI has poor prognosis relevance in TNBC patients diagnosed at ≤ 50 years-old. Further validation of the clinical significance of Ki-67 LI in TNBC is required.