Long-Term Outcome after Breast-Conservation Treatment with Radiation for Mammographically Detected Ductal Carcinoma in Situ of the Breast
Reviewer: John P. Plastaras, MD, PhD
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: October 16, 2005
Source: Cancer. Vol. 103, p. 1137-1146 (2005). Authors: Solin LJ, Fourquet A, Vicini FA, Taylor M, Olivotto IA, Haffty B, Strom EA, Pierce LJ, Marks LB, Bartelink H, McNeese MD, Jhingran A, Wai E, Bijker N, Campana F, Hwang WT. Affiliations: University of Pennsylvania, Philadelphia (LJS, WTH) and 9 other centers.
The role of radiotherapy as part of breast-conserving treatment of ductal carcinoma in situ (DCIS) of the breast has been somewhat controversial, despite 3 large randomized-controlled trials (NSABP B-17, EORTC, UK/ANZ DCIS trials) which showed that addition of radiation decreases local recurrence (LR) by about 50%
Despite the impact of radiotherapy on local control, these trials have not demonstrated a statistically-significant impact on overall survival
The use of screening mammography has dramatically increased the incidence of non-palpable DCIS, which is only detected by mammography
Compiling patients from 10 institutions, this paper represents the largest long-term (15 yr) study of patients with mammographically-detected DCIS treated with lumpectomy and radiation
Design: Retrospective, multi-institutional review of patients treated similarly
Patients: 1003 women with unilateral, mammographically detected DCIS who underwent lumpectomy and definitive breast irradiation (at least 40 Gy) treated before 1995 from 10 institutions (Univ of Pennsylvania, William Beaumont Hospital, Mallinckrodt Insitute of Radiology, Duke Univ, Univ of Michigan, M.D. Anderson Cancer Center, Yale Univ, British Columbia Cancer Agency, Institut Curie, and Netherlands Cancer Institute)
median f/u = 8.5 yrs with 363 patients at 10 yr and 68 at 15 yr
median age = 53 yr, 84% presented with microcalcifications alone
47% had re-excisions (usually to attempt to obtain negative margins); 60% had negative margins (defined by each institution; 8 institutions used 2 mm as cutoff between negative and close)
Median whole breast dose = 50 Gy
72% had boosts; median total dose = 60 Gy. Of the patients treated with lower total doses (75 were treated with <50 Gy), most were treated with larger fractions
Because the patients were treated before 1995, the use of adjuvant hormonal therapy (e.g. tamoxifen) was not commonplace
OS @ 15 y: 89%; Cause Specific Survival @ 15 y: 98%; LF @ 15 y: 19%
Median time to failure = 5.3 yr; of the patients whose first event was local failure, 55% had invasive recurrences
The authors provide detailed accounts of type of first failure: 85 had local failure, 71 had contralateral breast cancer, and 56 had second malignant neoplasms
In Cox multivariate analysis, both age <=50 (p=0.00057) and negative margins (p=0.0026) were very significantly associated with lower risk of LR
Both positive margins and CLOSE margins were associated with increased risk for local failure (HR 3.4 and 1.9, respective) compared to negative margins
Cause-specific survival for patients with mammographically-detected DCIS treated with lumpectomy/radiation is very high, therefore routine long-term health maintenance strategies should be emphasized
Because margin status is potentially controllable, negative margins should be obtained whenever possible, and especially in the high-risk younger patients
Although lacking central path review and definition of close/negative margin status varied some in this study (most used 2 mm), both positive and close margins had higher risk of LR
Given that local recurrences and contralateral breast cancers were among the most common first events, bilateral screening mammography should be continued;
Tamoxifen has shown to decrease the incidence of both local recurrence and contralateral breast cancer in a prospective trial (NSABP B-24); Tamoxifen could decrease the rate of these events compared to this study which pre-dated the tamoxifen era
Post-biopsy mammograms should be obtained prior to definitive radiotherapy when calcifications were seen at presentation
Given the usual limitations of retrospective studies, this is a very large cohort of similarly treated patients with mammographically-detected DCIS with 68/1003 followed at least 15 yr. By pooling patients from 10 institutions, enough patients were available for robust univariate and multivariate analyses.
Size: Unlike the Van Nuys Prognostic Index which uses tumor size in the scoring system, this study did not show an effect of tumor size (when known) – in fact there were more LRs in patients with smaller tumors. In the majority of patients (58%), the size was unknown, which makes it difficult to use in prognosis
The lack of central path review and the era of the study left certain potentially important prognostic factors out of the analysis such as grade, presence of comedo necrosis, let alone any type of molecular marker including ER/PR/Her-2 status.
Age: Previous smaller retrospective studies have linked young age with increased LR, however, this study establishes age as an INDEPENDENT risk factor for LR
Younger age and negative margins were associated with fewer local recurrences. This study highlights the importance of trying to obtain negative margins with re-excision when necessary.
Now that adjuvant hormonal therapy is common for DCIS, it is unknown if this will change the impact of margin and age seen in the pre-tamoxifen era
The rates of OS and CSS in this cohort of patients with DCIS treated with breast-conserving therapy compare favorably to patients historically-treated with mastectomy.