Breast-conserving therapy for ductal carcinoma in situ of the breast: the French Cancer Centers' experience.

Purpose: To assess the long-term outcome for women with ductal carcinoma in situ of the breast treated in current clinical practice by conservative surgery with or without definitive breast irradiation.
Methods and Materials: We analyzed 705 cases of ductal carcinoma in situ treated between 1985 and 1995 in nine French regional cancer centers; 515 underwent conservative surgery and radiotherapy (CS+RT) and 190 CS alone. The median follow-up was 7 years.
Results: The 7-year crude local recurrence (LR) rate was 12.6% (95% confidence interval [CI] 9.4-15.8) and 32.4% (95% CI 25-39.7) for the CS+RT and CS groups, respectively (p <0.0001). The respective 10-year results were 18.2% (95% CI 13.3-23) and 43.8% (95% CI 30-57.7). A total of 125 LRs occurred, 66 and 59 in the CS+RT and CS groups, respectively. Invasive or microinvasive LRs occurred in 60.6% and 52% of the cases in the same respective groups. The median time to LR development was 55 and 41 months. Nine (1.7%) and 6 (3.1%) nodal recurrences occurred in the CS+RT and CS groups, respectively. Distant metastases occurred in 1.4% and 3% of the respective groups. Patient age and excision quality (final margin status) were both significantly associated with LR risk in the CS+RT group: the LR rate was 29%, 13%, and 8% among women aged < or =40, 41-60, and > or =61 years (p <0.001). Even in the case of complete excision, we observed a 24% rate of LR (6 of 25) in women <40 years. Patients with negative, positive, or uncertain margins had a 7-year crude LR rate of 9.7%, 25.2%, and 12.2%, respectively (p = 0.008). RT reduced the LR rate in all subgroups, especially in those with comedocarcinoma (17% vs. 59% in the CS+RT and CS groups, respectively, p <0.0001) and mixed cribriform/papillary tumors (9% vs. 31%, p <0.0001). In the multivariate Cox regression model, young age and positive margins remained significant in the CS+RT group (p = 0.00012 and p = 0.016). Finally, the relative LR risk in the CS+RT group compared with the CS group was 0.35 (95% CI 0.25-0.51, p = 0.0001). Subsequent contralateral breast cancer occurred in 7.1% and 7.5% of the patients in the CS+RT and CS groups, respectively.
Conclusion: Despite the absence of randomization, our results are extremely consistent with the updated National Surgical Adjuvant Breast Project B17 and European Organization for Research and Treatment of Cancer 10853 trials. We also noted that the LR risk was very high in women <40 years and/or in the case of incomplete excision.