We sought to determine retrospectively whether extracapsular spread (ECS) might identify a subgroup that could benefit from radiotherapy after mastectomy, especially patients with 1 to 3 positive lymph nodes (LN1-3+). We randomized 1,475 premenopausal women with node-positive breast cancer to three, six, or nine courses of "classical" CMF (cyclophosphamide, methotrexate, and fluorouracil). After a review of all pathology forms, 933 patients (63%) had information on the presence or absence of ECS. ECS was present in 49.5%. The median follow-up was 10 years. In univariate analyses, ECS was associated with worse disease-free survival (DFS) and overall survival (OS). In multivariate analyses adjusting for tumor size, vessel invasion, surgery type, and age group, ECS remained significant (DFS: hazard ratio, 1.61; 95% CI, 1.34 to 1.93; P < .0001; OS: 1.67; 95% CI, 1.34 to 2.08; P < .0001). However, ECS was not significant when the number of positive nodes was added. The locoregional failure rate +/- distant failure (LRF +/- distant failure) within 10 years was estimated at 19% (+/- 2%) without ECS, versus 27% (+/- 2%) with ECS. The difference was statistically significant in univariate analyses, but not after adjusting for the number of positive nodes. No independent effect of ECS on DFS, OS, or LRF could be confirmed within the subgroup of 382 patients with LN1-3+ treated with mastectomy without radiotherapy. Our results do not support an independent prognostic value of ECS, nor its use as an indication for irradiation in premenopausal patients with LN1-3+ treated with classical CMF. However, we could not examine whether extensive ECS is of prognostic importance.