PURPOSE: To evaluate locoregional recurrence according to nodal status in women with T1 to T2 breast cancer and zero to three positive nodes (0-3N+) treated with breast-conserving surgery (BCS). METHODS AND MATERIALS: The study subjects comprised 5,688 women referred to the British Columbia Cancer Agency between 1989 and 1999 with pT1 to T2, 0-3N+, M0 breast cancer, who underwent breast-conserving surgery with clear margins and radiotherapy (RT) of the whole breast. The 10-year Kaplan-Meier local, regional, and locoregional recurrence (LR, RR, and LRR, respectively) were compared between the N0 (n = 4,433) and 1-3N+ (n = 1,255) cohorts. The LRR was also examined in patients with one to three positive nodes (1-3N+) treated with and without nodal RT. Multivariate analysis was performed using Cox regression modeling. RESULTS: Median follow-up was 8.6 years. Systemic therapy was used in 97% of 1-3N+ and 41% of N0 patients. Nodal RT was used in 35% of 1-3N+ patients. The 10-year recurrence rates in N0 and 1-3N+ cohorts were as follows: LR 5.1% vs. 5.8% (p = 0.04); RR 2.3% vs. 6.1% (p < 0.001), and LRR 6.7% vs. 10.1% (p < 0.001). Among 817 1-3N+ patients treated without nodal RT, 10-year LRR were 13.8% with age <50 years, 20.3% with Grade III, and 23.4% with estrogen receptor (ER)-negative disease. On multivariate analysis, 1-3N+ status was associated with significantly higher LRR (hazard ratio [HR], 1.85; 95% confidence interval, 1.34-2.55, p < 0.001), whereas nodal RT significantly reduced LRR (HR, 0.59; 95% confidence interval, 0.38-0.92, p = 0.02). CONCLUSION: Patients with 1-3N+ and young age, Grade III, or ER-negative disease have high LRR risks approximating 15% to 20% despite BCS, whole-breast RT and systemic therapy. These patients may benefit with more comprehensive RT volume encompassing the regional nodes.