Assessment of the relationships between Ki67 expression and neoadjuvant treatment response and prognosis in breast cancer using two types of response evaluation systems

Background: Although it is not a good prognostic marker for all breast cancers, pathological complete response (pCR) is an endpoint in evaluating neoadjuvant chemotherapy (NAC) efficacy. Ki67, a proliferation marker, has a complex role as a predictive marker in determining the NAC response. This study aimed to investigate the relationship between pretreatment and posttreatment Ki67 levels, posttreatment Ki67 change and treatment response using the Miller–Payne (MP) and Residual Cancer Burden (RCB) response assessment systems. Methods : A total of 178 invasive breast carcinoma patients who underwent NAC were included in the study. Ki67 levels were evaluated by immunohistochemical method in trucut biopsy and surgical excision specimens. Treatment response in the surgical excision specimen was classified according to both MP and RCB classifications. We investigated the relationships between pretreatment Ki67 level, posttreatment Ki67 level and posttreatment change in Ki67 with NAC response and survival. Additionally, the cut-off value of the pretreatment Ki67 level for pCR and nonpathological response (pNR) was investigated. Results . The pretreatment Ki67 level was significantly higher in the pCR group than in the partial response (pPR) and pNR groups (p<0.001) in both the MP and RCB systems. The posttreatment Ki67 level was significantly higher in the pNR group than in the pPR group (p<0.001) in both systems. There was a negative correlation between pretreatment Ki67 and disease-free survival (DFS) in the luminal B HER2-negative subtype (r=-0.377, p=0.036) and a significant negative correlation between posttreatment Ki67 and overall survival (OS) in the HER2-positive subtype (r=-0.544, p= 0.0209). A significant correlation was found between the posttreatment Ki67 change and the degree of response in the luminal B HER2-positive and HER2-positive subtypes (p<0.05). The Ki67 cut-off value was 37.5 for pCR in both the MP and RCB classification systems (95% CI 0.673-0.833 and 95% CI0.66-0.827) (p<0.001). There was a significant moderate agreement between the MP and RCB systems (p<0.001). Conclusions : Pre-and posttreatment Ki67 levels may be used to assess treatment response in various molecular subtypes of breast cancer. In this study, the cut-off value of Ki67 for pCR was 37%. More aggressive treatments may be considered in luminal B HER2-negative tumours with high pretreatment Ki67 levels and tumours that do not show a Ki67 decrease.