RTOG 9804: A prospective randomized trial for “good risk” ductal carcinoma in situ (DCIS), comparing radiation (RT) to observation (OBS)

Abstract

1004 Background: Whole breast RT following conservation surgery (BCS) for low risk DCIS has remained controversial despite several large trials comparing RT to OBS, all showing significant benefit in local control with RT. RTOG 9804 compares RT to OBS for mammographically detected disease, of low or intermediate nuclear grade, <2.5 cm size, and surgical margins ≥ 3 mm. Tamoxifen (TAM) use for 5 years was allowed but not required. Methods: The primary endpoint was ipsilateral breast local failure (LF). LF and contralateral breast failures (CBF) were estimated by the cumulative incidence method and treatment arms compared by log-rank test. Disease-free (DFS) and overall survival (OS) were estimated by the Kaplan-Meier method and treatment arms compared by log-rank test. Patients were stratified by age, margin width, grade, TAM use, and primary size. With 1790 patients, 80% power and using a 2-sided log rank test at 0.05, the study was designed to detect a reduction in 5-year local recurrence from 6% to 3.5% with RT. Results: Accrual goals for the planned 1790 patients were not met; the study was closed early. From December 1999 to July 2006, 636 women were randomized to receive 50 Gy in 5 weeks vs. OBS. 43 women were ineligible on review and 8 withdrew consent. Median follow-up (F/U) time was 6.46 years. Mean age was 59; TAM was used in 62% of women. There were 2 LF in the RT arm vs. 15 in the OBS arm: at 5 years 0.4% RT vs. 3.2% OBS (p=0.0023, HR [95%CI] = 0.14 [0.03, 0.61]). With limited events, LF is not correlated with size, grade, margin status, or age. The rate of CBF at 5 years was 3.0% for the RT arm vs. 1.9% for the OBS arm (p=0.42, HR [95%CI] = 1.46 [0.59, 3.62]) and does not appear to be influenced by TAM use (3.6 versus 2.7% TAM). The DFS and OS results were excellent. Rate of grade 1-2 toxicity was 76% in the RT arm vs. 30% in the OBS arm, and the rate of ≥ 3 grade toxicities was 4% on both arms. Conclusions: In this “good risk” subset of DCIS, the LF rate was decreased significantly with the addition of RT. Longer follow-up is planned.