Purpose: We evaluated the efficacy of cyclophosphamide, methotrexate, and fluorouracil (CMF) versus cyclophosphamide, doxorubicin, and fluorouracil (CAF) in node-negative breast cancer patients with and without tamoxifen (TAM), overall and by hormone receptor (HR) status.
Patients and methods: Node-negative patients identified by tumor size (> 2 cm), negative HR, or high S-phase fraction (n = 2,690) were randomly assigned to CMF, CAF, CMF + TAM (CMFT), or CAF + TAM (CAFT). Cox regression evaluated overall survival (OS) and disease-free survival (DFS) for CAF versus CMF and TAM versus no TAM separately. Two-sided CIs and one-sided P values for planned comparisons were calculated.
Results: Ten-year estimates indicated that CAF was not significantly better than CMF (P = .13) for the primary outcome of DFS (77% v 75%; HR = 1.09; 95% CI, 0.94 to 1.27). CAF had slightly better OS than CMF (85% v 82%, HR = 1.19 for CMF v CAF; 95% CI, 0.99 to 1.43); values were statistically significant in the planned one-sided test (P = .03). Toxicity was greater with CAF and did not increase with TAM. Overall, TAM had no benefit (DFS, P = .16; OS, P = .37), but the TAM effect differed by HR groups. For HR-positive patients, TAM was beneficial (DFS, HR = 1.32 for no TAM v TAM; 95% CI, 1.09 to 1.61; P = .003; OS, HR = 1.26; 95% CI, 0.99 to 1.61; P = .03), but not for HR-negative patients (DFS, HR = 0.81 for no TAM v TAM; 95% CI, 0.64 to 1.03; OS, HR = 0.79; 95% CI, 0.60 to 1.05).
Conclusion: CAF did not improve DFS compared with CMF; there was a slight effect on OS. Given greater toxicity, we cannot conclude CAF to be superior to CMF. TAM is effective in HR-positive disease, but not in HR-negative disease.