Abstract P6-07-36: A Prognostic Model for Triple-Negative Breast Cancer Patients Based on Node Status, Cathpesin-D and Ki-67
L Huang,Yb Liu,S Chen,Rj Zhou,Y Liu,Z-M Shao
DOI: https://doi.org/10.1158/0008-5472.sabcs12-p6-07-36
IF: 11.2
2012-01-01
Cancer Research
Abstract:Abstract Background: Triple-negative breast cancer (TNBC) had been classified as a breast cancer subgroup that is negative for ER, PR and HER2. The purpose of this study was to identify prognostic factors affecting outcome in women with TNBC and to construct a prognostic model which can predict recurrence. Material and methods: We retrospectively analyzed the clinicopathologic characteristics and outcome of 284 patients diagnosed with a stage I to III triple negative breast cancer in Shanghai Cancer Hospital, including age, menopausal status, tumor size, node status, grade, ER, PR, HER-2, Cathpesin-D, PS2, Nm23, P53, Bax, Bcl-2, P27, Topo-IIa, GSTn, MDR, Cyclin D1 and Ki67. 196 patients in 2000, 2001 and 2002 belonged to training set. Prognostic significance of factors was assessed using univariate and multivariate analyses, which were combined to build a model. DFS and OS were estimated by the Kaplan-Meier method and compared among groups via the log-rank test. The prognostic value of the model was further validated in separate testing set containing 88 patients in 2003 and 2004. The median age of patients was 50 years (range, 23–85 years). 95.1% patients received cytotoxic chemotherapy after surgery; 216 (76.1%) with anthracycline based, 54 (19.0%) with CMF regimen. Results: The median follow-up duration was 70.6 months, 102 patients (35.9%) experienced recurrence or metastasis. In Cox regression, ≥4 lymph node metastasis (HR = 3.7, P < 0.001), Cath-D positive (HR = 2.1, p = 0.034), and Ki-67 index≥20%, (HR = 2.0, p = 0.006) were identified as independent factors for DFS, only ≥4 lymph node metastasis (HR = 4.9, P < 0.001) showed prognostic significance for OS. The prognostic model is based on the sum total of factors, in which ≥4+ nodes, Cath-D + or Ki-67 index≥20% would contribute 1 point separately to the risk score. According to the prognostic score, patients were assigned to low-risk group(0 and 1 point) and high-risk group(2 and 3 point). The 5-year DFS rates in two groups were 82.3% and 47.8%, respectively (p < 0.001). The model provided a high predictive accuracy of recurrence, which showed an AUC of 0.71 in the training set and an AUC of 0.70 in validation set. Conclusion: Based on nodal status, Ki-67 and Cathpesin-D, we built a model that had considerable clinical value in predicting recurrence in TNBC patients, which help clinicians guide the appropriate level of adjuvant treatment and scheduling of surveillance visits. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P6-07-36.