Ductal carcinoma in situ (DCIS) of the breast: what is the safest treatment?

Purpose/Objective: Evaluating the outcome of women treated for DCIS by mastectomy (M), conservative surgery without (CS) or with radiotherapy (CS+RT) with a detailed analysis of local recurrence (LR) risk factors in the last subgroups.Materials/Methods: From 1985 to 1996, 1223 women with pure DCIS were treated in nine French Cancer Centers by M, CS or CS+RT. Median age was 52 years and median follow-up was 94 months. The rates of non-palpable cases were, respectively, 53.3%, 63.4% and 73.8% in the three groups. Family history of breast cancer was also found in 27%, 20.3% and 21.7%.Results: The crude LR rates were 26.4%, 13% and 1.7% (p<0.0001) with median delays to LR of 37, 53 and 38 months in the CS, CS+RT and M groups (see table).No LR risk factors were found in M group, but age and quality excision influenced most LR rates in both conservative groups. Among women treated by CS, the 8-year LR rates were 28.6%, 29.2% and 29.4% in women younger than 40, from 41 to 60, and older than 60 years respectively (NS). Among women treated by CS+RT, LR rates were 32.2%, 13% and 6.9% in the same subgroups (p < 0.001). In the CS group, the 8-year LR rates were 22%, 60.2% and 28.8% according to complete, incomplete/doubtful or unknown excision quality (p = 0.001). In the same subgroups treated by CS+RT, LR rates were 8.8%, 27.3% and 16.7% (p=0.0001). In all histological subgroups RT reduced the LR rates, especially in comedocarcinoma (43.1% in CS vs 16.1% in CS+RT, p < 0.0001). In a multivariate analysis on the 865 women treated by CS or CS+RT, treatment group, quality excision and age were all significant parameters. According to treatment groups, we only identified excision quality as a LR risk factor in CS group (p=0.0001) with 3.86 (95% IC = 2–7.4) and 1.96 (95% IC = 1.42–2.72) relative risk (RR) of LR among patients with incomplete/doubtful and unknown excision versus complete. In the CS+RT group, women under 40 and aged from 40 to 60 had a LR RR of 4.34 (95% IC = 1.9–9.8) and 2.08 (95% IC = 1.38–3.13) versus women older than 60 (p= 0.0004). The second significant parameter was excision quality (p < 0.0001) with 2.8 (95% IC = 1.7–4.6) and 1.67 (95% IC = 1.3–2.1) RR of LR among patients with incomplete/doubtful and unknown excision versus complete. Only one patient with a non-invasive LR developed subsequent metastases (1/60 = 1.7%), but after an invasive LR, we observed a 17% total metastases rate in the CS and CS+RT groups (15/88). The most important risk factors of contralateral BC was the occurrence of a LR in the first treated breast (26/154, 16.9% versus 57/1069, 5.33%; p < 0.0001). The presence of comedocarcinoma subtype also increased the risk of contralateral BC in the CS and CS+RT groups.Conclusions: M remains the safest treatment for DCIS with 98.3% 8-year local control. After conservative surgery, RT reduced LR rate by 66%, according to NSABP-B17 and EORTC 10853 trials’ results. Age under 40 and incomplete/doubtful excision are the strongest LR risk factors. These results suggest an increased follow-up among high LR risk groups treated conservatively in order to minimize the number of invasive LR. Purpose/Objective: Evaluating the outcome of women treated for DCIS by mastectomy (M), conservative surgery without (CS) or with radiotherapy (CS+RT) with a detailed analysis of local recurrence (LR) risk factors in the last subgroups. Materials/Methods: From 1985 to 1996, 1223 women with pure DCIS were treated in nine French Cancer Centers by M, CS or CS+RT. Median age was 52 years and median follow-up was 94 months. The rates of non-palpable cases were, respectively, 53.3%, 63.4% and 73.8% in the three groups. Family history of breast cancer was also found in 27%, 20.3% and 21.7%. Results: The crude LR rates were 26.4%, 13% and 1.7% (p<0.0001) with median delays to LR of 37, 53 and 38 months in the CS, CS+RT and M groups (see table).No LR risk factors were found in M group, but age and quality excision influenced most LR rates in both conservative groups. Among women treated by CS, the 8-year LR rates were 28.6%, 29.2% and 29.4% in women younger than 40, from 41 to 60, and older than 60 years respectively (NS). Among women treated by CS+RT, LR rates were 32.2%, 13% and 6.9% in the same subgroups (p < 0.001). In the CS group, the 8-year LR rates were 22%, 60.2% and 28.8% according to complete, incomplete/doubtful or unknown excision quality (p = 0.001). In the same subgroups treated by CS+RT, LR rates were 8.8%, 27.3% and 16.7% (p=0.0001). In all histological subgroups RT reduced the LR rates, especially in comedocarcinoma (43.1% in CS vs 16.1% in CS+RT, p < 0.0001). In a multivariate analysis on the 865 women treated by CS or CS+RT, treatment group, quality excision and age were all significant parameters. According to treatment groups, we only identified excision quality as a LR risk factor in CS group (p=0.0001) with 3.86 (95% IC = 2–7.4) and 1.96 (95% IC = 1.42–2.72) relative risk (RR) of LR among patients with incomplete/doubtful and unknown excision versus complete. In the CS+RT group, women under 40 and aged from 40 to 60 had a LR RR of 4.34 (95% IC = 1.9–9.8) and 2.08 (95% IC = 1.38–3.13) versus women older than 60 (p= 0.0004). The second significant parameter was excision quality (p < 0.0001) with 2.8 (95% IC = 1.7–4.6) and 1.67 (95% IC = 1.3–2.1) RR of LR among patients with incomplete/doubtful and unknown excision versus complete. Only one patient with a non-invasive LR developed subsequent metastases (1/60 = 1.7%), but after an invasive LR, we observed a 17% total metastases rate in the CS and CS+RT groups (15/88). The most important risk factors of contralateral BC was the occurrence of a LR in the first treated breast (26/154, 16.9% versus 57/1069, 5.33%; p < 0.0001). The presence of comedocarcinoma subtype also increased the risk of contralateral BC in the CS and CS+RT groups. Conclusions: M remains the safest treatment for DCIS with 98.3% 8-year local control. After conservative surgery, RT reduced LR rate by 66%, according to NSABP-B17 and EORTC 10853 trials’ results. Age under 40 and incomplete/doubtful excision are the strongest LR risk factors. These results suggest an increased follow-up among high LR risk groups treated conservatively in order to minimize the number of invasive LR.