Background Whole breast irradiation delivered once per day over 3–5 weeks after breast conserving surgery reduces local recurrence with good cosmetic results. Accelerated partial breast irradiation (APBI) delivered over 1 week to the tumour bed was developed to provide a more convenient treatment. In this trial, we investigated if external beam APBI was non-inferior to whole breast irradiation. Methods We did this multicentre, randomised, non-inferiority trial in 33 cancer centres in Canada, Australia and New Zealand. Women aged 40 years or older with ductal carcinoma in situ or node-negative breast cancer treated by breast conserving surgery were randomly assigned (1:1) to receive either external beam APBI (38·5 Gy in ten fractions delivered twice per day over 5–8 days) or whole breast irradiation (42·5 Gy in 16 fractions once per day over 21 days, or 50 Gy in 25 fractions once per day over 35 days). Patients and clinicans were not masked to treatment assignment. The primary outcome was ipsilateral breast tumour recurrence (IBTR), analysed by intention to treat. The trial was designed on the basis of an expected 5 year IBTR rate of 1·5% in the whole breast irradiation group with 85% power to exclude a 1·5% increase in the APBI group; non-inferiority was shown if the upper limit of the two-sided 90% CI for the IBTR hazard ratio (HR) was less than 2·02. This trial is registered with ClinicalTrials.gov, NCT00282035. Findings Between Feb 7, 2006, and July 15, 2011, we enrolled 2135 women. 1070 were randomly assigned to receive APBI and 1065 were assigned to receive whole breast irradiation. Six patients in the APBI group withdrew before treatment, four more did not receive radiotherapy, and 16 patients received whole breast irradiation. In the whole breast irradiation group, 16 patients withdrew, and two more did not receive radiotherapy. In the APBI group, a further 14 patients were lost to follow-up and nine patients withdrew during the follow-up period. In the whole breast irradiation group, 20 patients were lost to follow-up and 35 withdrew during follow-up. Median follow-up was 8·6 years (IQR 7·3–9·9). The 8-year cumulative rates of IBTR were 3·0% (95% CI 1·9–4·0) in the APBI group and 2·8% (1·8–3·9) in the whole breast irradiation group. The HR for APBI versus whole breast radiation was 1·27 (90% CI 0·84–1·91). Acute radiation toxicity (grade ≥2, within 3 months of radiotherapy start) occurred less frequently in patients treated with APBI (300 [28%] of 1070 patients) than whole breast irradiation (484 [45%] of 1065 patients, p<0·0001). Late radiation toxicity (grade ≥2, later than 3 months) was more common in patients treated with APBI (346 [32%] of 1070 patients) than whole breast irradiation (142 [13%] of 1065 patients; p<0·0001). Adverse cosmesis (defined as fair or poor) was more common in patients treated with APBI than in those treated by whole breast irradiation at 3 years (absolute difference, 11·3%, 95% CI 7·5–15·0), 5 years (16·5%, 12·5–20·4), and 7 years (17·7%, 12·9–22·3). Interpretation External beam APBI was non-inferior to whole breast irradiation in preventing IBTR. Although less acute toxicity was observed, the regimen used was associated with an increase in moderate late toxicity and adverse cosmesis, which might be related to the twice per day treatment. Other approaches, such as treatment once per day, might not adversely affect cosmesis and should be studied. Funding Canadian Institutes for Health Research and Canadian Breast Cancer Research Alliance.
