Background: An Early Breast Cancer Trialists' Collaborative Group (EBCTCG) meta-analysis indicates that adjuvant bisphosphonates increase time to bone recurrence and survival in postmenopausal breast cancer patients, but results of individual trials have been inconclusive. Retrospective analyses of AZURE, a trial of adjuvant zoledronic acid, showed MAF (a transcription factor of the AP-1 family) amplification status predicted bisphosphonate benefit independently of menopause for invasive disease-free survival (IDFS) and overall survival (OS). Validation of MAF amplification status as a potential companion diagnostic for adjuvant bisphosphonates was confirmed using NSABP-B34 specimens. Methods: The randomized, placebo-controlled NSABP B-34 study of women with stage 1-3 breast cancer were assigned to adjuvant systemic therapy plus oral clodronate 1600 mg daily or placebo for 3 years. The primary endpoint was disease-free survival (DFS) with overall survival (OS) as a secondary outcome. MAF amplification was assessed by fluorescence in-situ hybridization on anonymized sections of breast tumor tissue in all patients with tumor samples and performed in a laboratory blind to treatment assignment. Protocol and analysis plans were pre-specified. Disease outcomes were analysed using intention to treat principles. Results: 2496 B-34 patients contributed tumor samples (from 2001-2004), of whom 1883 (75%) were evaluable (947 placebo and 936 clodronate). 1515 (80%) tumors were MAF negative (766 placebo and 749 clodronate) and 368 were MAF positive. At median follow-up of 108 months, MAF was prognostic for DFS, OS and bone-metastasis-free survival in the control group (MAF-positive vs MAF-negative: HRDFS=1·39, 95%CI 1·01-1·92; p=0.045; HROS=1·59, 95%CI 1·08-2·33; p=0.018; HRBM=2·03, 95%CI 1·13-3·68; p=0.016). In patients with MAF-negative tumors, clodronate gave higher DFS and OS than controls at 60 months (HRDFS=0·70, 95%CI 0·51-0·94; p=0.020 and HROS=0·59, 95%CI 0·37-0·93; p=0.024), the latter maintained through follow-up (HROS=0·74, 95%CI 0·54-1.00; p=0.047), but not in patients with MAF-positive tumors - consistent with previous AZURE results. Conclusions: MAF benefit prediction from adjuvant bisphosphonates was confirmed using specimens from 2 randomized clinical trials (AZURE and NSABP-B-34) conducted and analyzed in similar manner using the same validated tests and clinical endpoints. These results are evidence towards introducing MAF testing into clinical practice.