Background
Phase III RCT (RTOG 98-04). 636 patients with "good-risk" DCIS treated with breast-conserving surgery: nuclear grade 1-2, lesion ≤2.5 cm, margins ≥3 mm. Tamoxifen was allowed at physician discretion (~62% of participants). Randomized to observation alone vs whole breast irradiation. This is the 7-year outcomes publication.
Interventions and follow up
Arm A: Observation alone (± tamoxifen)
Arm B: Whole breast RT (50 Gy in 25 fractions) (± tamoxifen)
Primary endpoint: Ipsilateral breast event (IBE) — invasive or in situ recurrence
mFollow up: 7.2 years (median)
Arm B: Whole breast RT (50 Gy in 25 fractions) (± tamoxifen)
Primary endpoint: Ipsilateral breast event (IBE) — invasive or in situ recurrence
mFollow up: 7.2 years (median)
Results
7-year cumulative IBE: 6.7% (observation) vs 0.9% (RT), P<.001 — absolute difference 5.8%
7-year invasive IBE: 3.5% (obs) vs 0.4% (RT)
Overall survival: No significant difference
7-year invasive IBE: 3.5% (obs) vs 0.4% (RT)
Overall survival: No significant difference
Adverse events
Main adverse events: RT was well tolerated with expected skin toxicities. Grade 3+ adverse events were uncommon. No excess cardiac events reported. Tamoxifen-associated toxicities (DVT, endometrial cancer) were small but present in the tamoxifen arm.
Conclusions
For patients with low-to-intermediate nuclear grade DCIS meeting strict criteria, RT reduced the 7-year IBE rate from 6.7% to 0.9% (absolute reduction ~6%, relative reduction 87%). Despite the low absolute recurrence risk in both arms, RT substantially reduced the risk of both invasive and non-invasive local recurrence. OS was not impacted.
Key Limitations
Key Limitations: The control arm IBE rate of 6.7% at 7 years is lower than many historical series — reflecting the highly favorable selection criteria (grade 1-2, ≤2.5cm, margins ≥3mm). Longer follow-up would likely show higher IBE rates (per ECOG 5194, rates continue to rise through 12 years). Tamoxifen use was not randomized. The trial enrolled "good-risk" DCIS only, limiting applicability to the broader DCIS population.
Clinical Context
RTOG 98-04 and ECOG 5194 together form the evidentiary basis for RT omission discussions in low-risk DCIS. However, neither study is a head-to-head comparison of RT vs observation per se; ECOG 5194 is observational. Together, these trials support shared decision-making: for older patients with grade 1-2 DCIS, small lesion, and wide margins, observation may be acceptable given low absolute recurrence rates, particularly if no RT-sensitive endpoint (mortality) has been demonstrated to benefit.
References