PURPOSE To present local control, complications, and cosmetic outcomes of intraoperative radiation therapy (IORT) for early breast cancer, as well as technical aspects related to the use of a nondedicated linear accelerator. METHODS AND MATERIALS This prospective trial began in May of 2004. Eligibility criteria were biopsy-proven breast-infiltrating ductal carcinoma, age >40 years, tumor <3 cm, and cN0. Exclusion criteria were in situ or lobular types, multicentricity, skin invasion, any contraindication for surgery and/or radiation therapy, sentinel lymph node involvement, metastasis, or another malignancy. Patients underwent classic quadrantectomy with intraoperative sentinel lymph node and margins evaluation. If both free, the patient was transferred from operative suite to linear accelerator room, and IORT was delivered (21 Gy). Primary endpoint: local recurrence (LR); secondary endpoints: toxicities and aesthetics. Quality assurance involved using a customized shield for chest wall protection, applying procedures to minimize infection caused by patient transportation, and using portal films to check collimator-shield alignment. RESULTS A total of 152 patients were included, with at least 1 year follow-up. Median age (range) was 58.3 (40-85.4) years, and median follow-up time was 50.7 (12-110.5) months. The likelihood of 5-year local recurrence was 3.7%. There were 3 deaths, 2 of which were cancer related. The Kaplan-Meier 5-year actuarial estimates of overall, disease-free, and local recurrence-free survivals were 97.8%, 92.5%, and 96.3%, respectively. The overall incidences of acute and late toxicities were 12.5% and 29.6%, respectively. Excellent, good, fair, and bad cosmetic results were observed in 76.9%, 15.8%, 4.3%, and 2.8% of patients, respectively. Most treatments were performed with a 5-cm collimator, and in 39.8% of the patients the electron-beam energy used was ≥12 MeV. All patients underwent portal film evaluation, and the shielding was repositioned in 39.9% of cases. No infection or anesthesia complications were observed. CONCLUSIONS Local control with IORT was adequate, with low complication rates and good cosmetic outcomes. More than one-third of patients benefited from the "image-guidance" approach, and almost 40% benefited from the option of higher electron beam energies.