Background
NSABP B-33. Phase III RCT of 1,598 postmenopausal patients with HR+ breast cancer who were disease-free after 5 years of tamoxifen. During recruitment, the MA.17 results demonstrated benefit from extended letrozole, prompting early termination of accrual, unblinding, and an offer of exemestane to placebo patients.
Interventions and follow up
Arm A: Exemestane 25mg/d × 5yr after 5yr of tamoxifen (72% continued exemestane)
Arm B: Placebo × 5yr after 5yr of tamoxifen (44% crossed over to exemestane)
Primary endpoint: DFS
mFollow up: 30mo
Arm B: Placebo × 5yr after 5yr of tamoxifen (44% crossed over to exemestane)
Primary endpoint: DFS
mFollow up: 30mo
Results
4-yr DFS: 91% (exemestane) vs 89% (placebo); RR 0.68; P=.07 (borderline, not significant)
4-yr RFS (relapse-free survival): 96% vs 94%; RR 0.44; P=.004
4-yr RFS (relapse-free survival): 96% vs 94%; RR 0.44; P=.004
Adverse events
Overall: Toxicity assessed up to unblinding was acceptable for the adjuvant setting; early closure and crossover limited mature safety reporting
Musculoskeletal: Arthralgia and bone loss with exemestane
Vasomotor: Hot flashes with exemestane
Musculoskeletal: Arthralgia and bone loss with exemestane
Vasomotor: Hot flashes with exemestane
Conclusions
Extended adjuvant exemestane after 5 years of tamoxifen produced a significant improvement in RFS and a non-significant trend in DFS. Premature closure and substantial crossover diluted the estimated benefit.
Key Limitations
Trial closed early after MA.17, markedly reducing power; 44% of placebo patients crossed to exemestane, biasing toward the null. Short follow-up (30mo). No OS conclusion possible. The primary DFS endpoint was not statistically significant.
Clinical Context
Supportive (alongside MA.17) of extended adjuvant AI after 5 years of tamoxifen in postmenopausal HR+ disease. ASCO/ESMO endorse considering extended endocrine therapy in higher-risk patients; B-33's design limitations mean MA.17/MA.17R carry greater weight.