BACKGROUND:The objective of this study was to evaluate and identify patterns of failure and prognostic factors for locoregional recurrence (LRR) that could justify postmastectomy radiotherapy after modified radical mastectomy in patients with early stage triplenegative breast cancer. METHODS: Between January 2000 and July 2007, the authors retrospectively analyzed 390 patients who had triple-negative breast cancer with T1/T2 tumors and from zero to 3 positive lymph nodes (pathologic T1-T2N0-N1) who underwent modified radical mastectomy without postmastectomy radiotherapy at the author's institution. The 5-year cumulative incidence for events was calculated using Kaplan-Meier analysis, and subgroups were compared using the log-rank test. Multivariate analysis was performed using a Cox proportional hazards model. RESULTS: Overall, 86.4% of patients received chemotherapy. At a median follow-up of 60.5 months, the 5-year cumulative rates of local recurrence, regional recurrence, LRR, and distant metastasis were 5.4%, 4.7%, 8%, and 13.4%, respectively. On multivariate analysis, age <50 years, the presence of lymphovascular invasion, grade 3 tumor, and 3 involved lymph nodes were associated significantly with an increased risk of LRR. The 5-year LRR rate for patients who had 0 or 1 risk factor, 2 risk factors, and 3 or 4 risk factors was 4.2%, 25.2%, and 81% (P <.0001), respectively. The presence of lymphovascular invasion and having 3 involved lymph nodes were statistically significant predictors of regional recurrence, and the patients who had regional recurrence had a significantly greater risk of distant metastases compared with patients who had local recurrence (59.1% vs 20.9%; P < .0001). CONCLUSIONS: Several risk factors were identified in this study that correlated independently with a greater incidence of LRR in patients who had early stage triple-negative breast cancer. The current results indicated that postmastectomy radiotherapy should be considered for those patients who have 2 or more of these factors. Cancer 2013;119:2366-74.