Phase II trial of brachytherapy alone after lumpectomy for select breast cancer: Toxicity analysis of RTOG 95-17

Purpose: Accelerated partial breast irradiation (APBI) can be delivered with brachytherapy within 4-5 days compared with 5-6 weeks for conventional whole breast external beam radiotherapy. Radiation Therapy Oncology Group 95-17 is the first prospective phase I-II cooperative group trial of APBI alone after lumpectomy in select patients with breast cancer. The toxicity rates are reported for low-dose-rate (LDR) and high-dose-rate (HDR) APBI on this trial. Methods and Materials:: The inclusion criteria for this study included invasive nonlobular tumors ≤3 cm after lumpectomy with negative surgical margins and axillary dissection with zero to three positive axillary nodes without extracapsular extension. The patients were treated with either LDR APBI (45 Gy in 3.5-5 days) or HDR APBI (34 Gy in 10 twice-daily fractions within 5 days). Chemotherapy (≥2 weeks after APBI) and/or tamoxifen could be given at the discretion of the treating physicians. Results: Between August 1997 and March 2000, 100 women were enrolled in this study, and 99 were evaluated. Of the 99 women, 33 were treated with LDR and 66 with HDR APBI. The median follow-up for all patients was 2.7 years (range, 0.6-4.4 years) and was 2.9 years for LDR and 2.7 years for HDR patients. Toxicities attributed to APBI included erythema, edema, tenderness, pain, and infection. Of the 66 patients treated with HDR APBI, 2 (3%) had Grade 3 or 4 toxicity. Of the 33 patients treated with LDR, 3 (9%) had Grade 3 or 4 toxicity during brachytherapy. Late toxicities included skin thickening, fibrosis, breast tenderness, and telangiectasias. No patient experienced late Grade 4 toxicity; the rate of Grade 3 toxicity was 18% for the LDR and 4% for the HDR groups. Conclusion: Acute and late toxicity for this invasive breast radiation technique was modest and acceptable. Patients receiving chemotherapy, a nonprotocol therapy, had a greater rate of Grade 3 toxicity. The study design did not allow for this to be tested statistically