A prognostic model based on nodal status and Ki 67 predicts the risk of recurrence and death in breast cancer patients with residual disease after preoperative chemotherapy

Introduction and Aims: The achievement of a pathologic complete response (pCR) after preoperative chemotherapy (PCT) is a validated surrogate end point for long term outcome. On the other side, patients with residual disease in the breast and/or axilla are an heterogeneous group with very different prognosis, including both patients with truly chemo-resistant disease as well as patients with an important tumor downstaging even if not in pCR. Aim of this analysis is to identify, in patients with residual disease, potential markers able to discriminate patients at higher risk of relapse. Patients and Methods:. The following parameters were evaluated on the surgical specimen in patients with less than pCR following PCT: residual breast disease, number of involved nodes, proliferation (Ki 67), hormone receptor, HER2, p53, EGFR, VEGFR2. Survival curves were estimated with the Kaplan-Meier method and the log rank test was used to test for differences between groups. Hazard Ratios and their confidence intervals were estimated by using Cox model. Results: 195 breast cancer patients were included. Median age 51 yrs (range: 27-73); 71% of the patients had ER+ tumors at diagnosis, 20% were HER2+. After PCT, 55% of the patients received mastectomy, 45% underwent conservative surgery. 57% of the patients had residual breast disease 10 nodes; 50% of the patients had Ki 67 >/= 15%. Among the examined parameters, nodal positivity and Ki 67>/=15% were significantly related with a higher risk of relapse (HR 2.5 , p=0.014 and HR 3.4, p /=15% was also predictive of a higher risk of death (HR 4.1, p=0.007). On the basis of these two parameters, patients were classified in three groups: 1) low risk (negative nodes and Ki 67/= 15%): 54.4% of the patients; 3) high risk (nodal positivity and Ki 67>/= 15%): 31.2% of the patients. Five-year DFS rates were 90%, 72%, and 43% respectively (log rank test p