Neoadjuvant epirubicin, gemcitabine, and docetaxel for primary breast cancer: Survival and prognostic factors in two consecutive neoadjuvant phase I/II trials.
1096 Background: We previously reported primary end points of two consecutive phase I/II trials which evaluated two different schedules of neoadjuvant gemcitabine (G), epirubicin (E) and docetaxel (Doc) for primary breast cancer. Here we report mature survival data and evaluate prognostic factors for disease-free (DFS) and overall survival (OS). Methods: 151 patients were treated in two phase I/II trials of G, E and Doc as neoadjuvant chemotherapy (NAC) for T2-4 N0-2 M0 PBC between Feb. ’02 and Dec. ’04. Patients were treated with six cycles of GEDoc (G 800mg/m2 day (d) 1+8, E 60-90mg/m2 d 1, Doc 60-75mg/m2 d 1 every three weeks) or five cycles of G 1250mg/m2 plus E 90-100mg/m2 every two weeks followed sequentially by four cycles of Doc 80-100mg/m2 every two weeks (GEsDoc). Pathologic complete response (pCR), clinical/pathological factors were correlated with DFS and OS. Results: There was no significant difference in DFS or OS between patients in the GEDoc and GEsDoc trial (DFS: Hazard ratio (HR) 1.13, p=0.67; OS: HR 1.06, p=0.88) with a 5-year DFS and OS of 72 vs 74% and 85 vs 86%, respectively. In an univariate analysis pCR unexpectedly was associated with a worse OS (HR 3.11; p= 0.007). HR for DFS showed a similar but non-significant trend (HR 1.78; p=0.1). Molecular subtypes (OS: HR [lum B] 3.17; [triple negative] 5.81; [HER2] 11.5; p=0.002), negative estrogen receptor (ER) status (OS: HR 3.14; p=0.002) and Ki-67 >20% (OS: HR 5.41; p=0.001) were all significantly associated with DFS and OS. The recently published CPS-EG score (Mittendorf 2011) was also significantly correlated with OS (p=0.006) and DFS (p=0.0006). In a multivariate analysis high Ki-67 was the only significant predictor of OS (HR 10.4; p=0.0026) whereas molecular subtype (p=0.05) and Ki-67 (p=0.04) were significantly associated with DFS. Conclusions: These results raise caution on the reliability of pCR as a single surrogate marker for survival in trials with small sample sizes. Our results emphazise the role of additional factors, esp. Ki-67 and subtypes. Integrative scores based on clinical and pathologic stage as well as tumor biology, might be more reliable predictors of survival.