Background: Anthracycline-containing regimens are considered superior to cyclophosphamide, methotrexate, and fluorouracil (CMF) as adjuvant therapy for early breast cancer, but predictors of anthracycline-derived benefit are lacking. We compared CMF and epirubicin (E) in different sequences with CMF alone in a phase III trial on rapidly proliferating breast cancer (RPBC), observing no differences in disease-free (DFS) or overall survival (OS). We performed a post hoc analysis on efficacy in relation to tumor biologic profiles. Methods: Patients (pts) with N-, T > 1 cm or ≤ 3 N⁺ and any T RPBC, defined by thymidine labeling index or grade or S-phase or Ki67 (≥20), were randomized to receive E (100 mg/m2 i.v. q 21 d, 4 cycles) followed by CMF (600, 40, 600 mg/m2 i.v. d 1 and 8 q 28, 4 cycles) (arm A, 440 pts) or CMF followed by E (arm B, 438 patients) or CMF for 6 cycles (arm C, 188 pts, closed after the EBCTCG 2000 meta-analysis). Biological features were fully assessable in 755 pts (arm A/B/C: 303/309/143) and included estrogen (ER) and progesterone (PgR) receptors (cut off ≥10%), Ki-67 (very high: ≥40%) and HER2 (positive: 3+). DFS and OS were estimated with Kaplan-Meier curves and compared by log rank test and Cox models. Results: Among pts with triple negative RPBC, CMF yielded significantly worse 5-year DFS and OS rates compared with CMF and E sequences (Table). A very high Ki-67 (>40%) predicted a significantly worse 5-year DFS (59% vs 83%, p=.041) in pts treated with CMF, especially within the triple negative subset (38% vs 90%, p=.044), but not in the other treatment group. The interaction between treatment and tumor type (triple negative vs HER2-positive) was significant (p=0.001 for DFS, p=0.01 for OS). Conclusions: In pts with triple negative RPBC, CMF appears to be inferior to E and CMF combinations. A very high Ki-67 (≥40%) seems to identify pts among those with triple-negative phenotype who benefit from E.