Purpose

The value of Ki67 measured on residual disease after neoadjuvant chemotherapy is not sufficiently described.

Experimental design

Participants of the GeparTrio study with primary breast cancer randomly received neoadjuvant response-guided [8 cycles TAC (docetaxel/doxorubicin/cyclophosphamide) in responding and TAC-NX (vinorelbine/capecitabine) in nonresponding patients] or conventional (6 cycles TAC) chemotherapy according to interim response assessment. Ki-67 levels were centrally measured immunohistochemically after neoadjuvant treatment if tumor tissue was available. Here, we analyze 1,151 patients having a pathologic complete response (pCR; n, 484), or residual disease with low (0-15%), intermediate (15.1-35%), or high (35.1-100%) posttreatment Ki67 levels in 488, 77, and 102 patients, respectively.

Results

Patients with high posttreatment Ki67 levels showed higher risk for disease relapse (P < 0.0001) and death (P < 0.0001) compared with patients with low or intermediate Ki67 levels. Patients with low Ki67 levels showed a comparable outcome to patients with a pCR (P = 0.211 for disease-free and P = 0.779 for overall survival). Posttreatment Ki67 levels provided more prognostic information than pretreatment Ki67 levels or changes of Ki67 from pre- to posttreatment. Information on pCR plus posttreatment Ki67 levels surmount the prognostic information of pCR alone in hormone-receptor-positive disease [hazard ratios (HR), 1.82-5.88] but not in hormone-receptor-negative disease (HR: 0.61-1.73). Patients with conventional and response-guided treatment did not show a different distribution of posttreatment Ki67 (P = 0.965).

Conclusions

Posttreatment Ki67 levels provide prognostic information for patients with hormone-receptor-positive breast cancer and residual disease after neoadjuvant chemotherapy. Levels were not prognostic for outcome after response-guided chemotherapy. High posttreatment Ki67 indicates the need for innovative postneoadjuvant treatments.