Randomized Trials Demonstrate Efficacy of Radiotherapy for Treatment of DCIS

The efficacy of radiotherapy at reducing in-breast recurrence after lumpectomy for DCIS has been demonstrated in five phase III prospective randomized trials ( Table 47.1 ). The first four of these to be conducted—the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-17 trial, the European Organization for Research and Treatment of Cancer (EORTC) 10853 trial, the United Kingdom Coordinating Committee on Cancer Research (UKCCCR) DCIS trial, and the Swedish multicenter (SweDCIS) trial)—now have 15 to 17.5 years follow-up. The first trial to be opened was the NSABP B-17 clinical trial, which enrolled 818 women from 1985 through 1990 who had undergone lumpectomy for DCIS with microscopically clear margins and were randomized postoperatively to observation versus whole breast radiotherapy. The most recent analysis done after 17.25 years median follow-up demonstrates a sustained benefit of breast radiotherapy with a 52% reduction in the risk of invasive ipsilateral breast cancer recurrence (hazard ratio [HR] 0.48 95% confidence interval [CI] 0.33–0.69, p < .001) and a 47% relative reduction in the risk of DCIS ipsilateral in-breast recurrence (HR 0.53, 95% CI 0.35–0.8, p < .001) compared with those randomized to lumpectomy alone.

The EORTC 10853 trial enrolled 1010 women postlumpectomy with DCIS ≤5 cm in size randomizing cases to observation versus breast radiotherapy over a similar time period (1986–1996). In comparison to NSABP B-17, microscopically clear resection margins were not stipulated for eligibility in this trial, and fewer cases were mammography detected ( Table 47.2 ). However, 15-year outcomes from the EORTC 10853 trial are similar to the NSABP B-17 trial demonstrating a sustained 47% relative reduction in ipsilateral local recurrence and approximately equal reduction in DCIS and invasive cancer recurrences.

The Swedish Breast Cancer group enrolled 1067 women from 1987 to 1999 who had been invited to participate in a mammography screening program and had undergone lumpectomy for DCIS occupying a quadrant or less of the breast. Microscopically clear surgical margins were not required, but a specimen radiograph was done for 97%. At a mean of 8 years of follow-up, a 60% reduction in local recurrence (corresponding relative risk of 0.40; 95% CI 0.30–0.54) was seen with the addition of radiotherapy with similar reductions in risk for ipsilateral invasive and DCIS recurrences. Extended follow-up on this trial was attained by using the unique Swedish national registration number to link the trial database to the Swedish Cancer Registry. With a median follow-up of 17.5 years, 20-year outcomes were reported. In contrast to the other trials, a somewhat smaller overall relative risk reduction of 37.5% was reported in the radiotherapy arm. The cumulative incidence of recurrence was 32.0% in the observed arm (95% CI 28.0–36.0) and 20.0% in the radiotherapy arm (95% CI 16.0–24.0). A much larger relative risk reduction of 67% was seen for DCIS compared with 13.0% for invasive breast cancer recurrence.

The UK/ANZ DCIS Trial accrued 1701 women with DCIS detected in the National Breast Screening Program who had undergone lumpectomy with cancer-free surgical margins between 1990 and 1998. The trial used a 2 × 2 factorial design to assess radiotherapy, tamoxifen, or both in patients with completely excised DCIS. Patients could elect either to enter into the four-way randomization or into one of two separate two-way randomizations. Among the various randomization schemes, 1030 patients were randomized to radiotherapy or observation after lumpectomy. A 68% relative risk reduction and 12.3% absolute risk reduction in ipsilateral cancer recurrences were reported at 10 years (12.7 year median follow-up) with 19.4% recurrences in the observed arm versus 7.1% with radiotherapy (HR 0.32, 95% CI 0.22–0.47, p < .0001).

These four trials were included in a meta-analysis by the Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) that included 3729 women with a median follow-up of 8.9 years. Radiotherapy approximately halved the rate of ipsilateral breast events (rate ratio 0.46, standard error [SE] 0.05, 2 p < .00001) with no evidence of heterogeneity between the trials in the proportional reduction. Radiotherapy resulted in a larger proportional reduction in the rate of ipsilateral breast recurrence for women more than 50 years of age compared with younger women (rate ratios: age <50 years 0.69, SE 0.12; ≥50 years 0.38, SE 0.06, 2 p = .0004 for the difference between these proportional reductions). The proportional reduction in recurrence by radiotherapy did not differ significantly according to any other clinical or pathologic factor. There was no significant difference in the meta-analysis for breast cancer or overall mortality between treatment arms. There were 50 of 1878 (2.7%) breast cancer deaths for the radiotherapy groups and 44 of 1851 (2.3%) for observation postlumpectomy. Importantly, there was no significant difference in heart disease deaths in those irradiated versus observed.

The four randomized trials just reviewed were accrued during an era when DCIS was still clinically detected in some, often larger in size, frequently higher grade with significant comedo necrosis and without consistent attention to completeness of excision by negative surgical margins (see Table 47.2 ). On closer examination, the rise in DCIS incidence seen in the SEER analysis is related to more detection of “noncomedo” DCIS versus higher-risk “comedo” histology that has been relatively stable in incidence. This led to the fifth phase III randomized trial, the Radiation Therapy Oncology Group (RTOG) 9804 trial in “good-risk” DCIS to evaluate whether radiotherapy benefits those who undergo lumpectomy for DCIS with good risk features, that is, mammographically detected, low- or intermediate-grade DCIS, 2.5 cm or less in size, with 3 mm or greater surgical margins. The study did not meet accrual goal and was closed early with 636 enrolled of a planned 1790. The population accrued was uniformly mammography detected, surgical margins consistently negative, with smaller lesion size on average without high nuclear grade compared with the prior phase III trials (see Table 47.2 ). Roughly two-thirds were treated with tamoxifen. With a median follow-up was 7.17 years, there were 19 in-breast recurrences in the observation arm versus two in the radiotherapy arm. The cumulative incidence of recurrence was 6.7% (95% CI 3.2%–9.6%) in the observation arm versus 0.9% (95% CI 0.0%–2.2%) in the RT arm (HR 0.11; 95% CI 0.03–0.47; p < .001).

Results With Excision Alone in Selected Patients

The majority of women with DCIS are now diagnosed with small, nonpalpable lesions that are detected by mammography. With improved diagnostic, surgical, and pathologic techniques and a better understanding of which factors can affect local recurrence, many investigators have proposed that excision alone may be adequate treatment for DCIS in select women. There have been a series of studies exploring omission in select subsets.

Silverstein and colleagues published results for a nonrandomized series of 706 patients treated in Van Nuys, California, or at the University of Southern California (USC) in Los Angeles (426 by excision alone and 280 by excision and radiotherapy), which showed that the addition of radiotherapy did not decrease the overall rate of local recurrence (17.5% and 16.4% in the two groups, respectively). The authors found that tumor size, margin width, grade, and patient age were independent predictors of local recurrence. They combined these parameters to create the “USC/Van Nuys Prognostic Index.” They suggested that patients with intermediate and high scores benefit from the addition of radiotherapy, whereas patients with low scores could be adequately treated with breast conserving surgery alone. Of great importance, the surgical specimens of patients included in this series underwent complete serial sectioning at 2- to 3-mm intervals. Sections were then arranged and processed in sequence. Unfortunately, this kind of painstaking analysis is not available at most treatment centers. Several recent publications that have attempted to validate the USC/Van Nuys Prognostic Index or earlier Van Nuys Prognostic Index have not demonstrated that the score accurately predicts the risk of local recurrence.

Two prospective single-arm trials have been conducted in recent years in the United States to test the hypothesis that excision with wide surgical margins may be adequate treatment for patients with small, low-grade lesions. One of these was conducted at the teaching hospitals affiliated with the Harvard Medical School from 1995 to 2002. Eligibility criteria included maximum size of 25 mm, predominant histology nuclear grade 1 or 2, final margins of 10 mm or more or no tumor on reexcision, and no suspicious calcifications on postoperative mammograms. After 158 patients were enrolled, the study was stopped early after a high number of local recurrences were observed. At a median follow-up of 40 months, 13 patients developed an ipsilateral recurrence, for a 5-year actuarial rate of local recurrence of 12%. The authors concluded that radiotherapy could not be safely omitted, even in the population of patients considered to be at lowest risk of recurrence. Updated study results were recently published. With a minimum follow-up of 8 years and a median of 11 years, the 10-year cumulative incidence of local recurrence was 15.6 % indicating substantial and ongoing risk of local recurrence even among women with favorable DCIS undergoing excision alone.

Eastern Cooperative Oncology Group (ECOG) conducted a prospective phase II trial, E5194, of wide local excision alone without radiotherapy for treatment of DCIS. There were two study arms: low- or intermediate-grade DCIS 2.5 cm or less in size and high-grade DCIS 1 cm or less in size. In the low/intermediate-grade arm, 561 women were enrolled, with a median age of 60 and median DCIS size of 6 mm. Thirty-one percent received tamoxifen. In the high-grade arm, 104 women were enrolled before early termination, with a median age of 58 and median DCIS size of 7 mm. Twenty-four percent received tamoxifen. With a median follow-up of 12.3 years, 12-year ipsilateral breast events were 14.4% in the low/intermediate-grade arm and 24.6% in the high-grade arm.

ECOG/ACRIN E5194 and RTOG 9804 suggest that for low- or intermediate-grade DCIS spanning less than 2.5 cm with greater than 3-mm surgical margins, in-breast recurrence rates steadily increase with time but may be acceptably low to justify omission of radiation.

Factors Associated With Local Recurrence