Objective: There is much evidence for the detrimental effect of treatment interruptions on tumor control, particularly in head and neck cancer. In order to determine the outcome of the treatment interruptions in postoperative irradiation of breast cancer, 853 female patients treated between 1990 and 1999 inclusive were retrospectively analyzed. Methods: Locally advanced breast cancer patients who received neoadjuvant chemotherapy were not included in the study. Five hundred and forty-six patients (64%) treated with mastectomy and 307 patients (36%) with breast-conserving surgery were analyzed. A total dose of 50 Gy (46-54 Gy) was given to the chest wall/breast and regional lymph nodes in 1.8- to 2-Gy daily fractions, 5 times per week. A 14-Gy (10- to 20-Gy) photon or electron boost was given to the tumor bed of the patients with breast-conserving surgery. Unplanned treatment interruptions occurred in 741 (87%) of the patients and the median duration of the gaps was 13 days (1-91 days). A total of 348 patients (41%) had no treatment break or interruptions of 1 week or less, whereas 505 patients (59%) had treatment interruptions of more than 1 week. The locoregional control (LC) and overall survival (OS) rates were estimated with the Kaplan-Meier method. A Cox proportional hazard regression model was used to evaluate the influence of host- and treatment-related factors on LC and OS (age, menopausal status, histological subtype, grade, hormonal receptor status, pT stage, pN stage, type of surgery, adjuvant treatment, number of gaps and duration of gaps). Results: For all patients LC rates for 5 and 10 years were 95 and 87%, respectively, and OS rates were 78% for 5 years and 62% for 10 years. LC rates for the group of patients with no treatment break or interruptions of 1 week or less, for 5 and 10 years were 94 and 90%, whereas the LC rates for 5 and 10 years were 89 and 86%, for the group of patients with interruptions of more than 1 week (p=0.019). Treatment interruptions of more than 1 week and premenopausal status appeared to be independent adverse prognostic factors in multivariate analyses affecting the LC (p=0.043 and p=0.005, respectively). The OS rates for the patients without treatment interruptions or interruptions of 1 week or less were also significantly better than for the patients with treatment interruptions of more than 1 week (p=0.026) in multivariate analyses. Conclusion: Interruptions more than 1 week during postoperative irradiation of breast cancer adversely affect the treatment outcome. Copyright (C) 2005 S. Karger AG, Basel.