Comparing Treatment for Ductal Carcinoma In Situ (DCIS) With or Without Sentinel Lymph Node Biopsy [Internet]

Excerpt

Objectives: This project had 2 primary goals. Aim 1 examined the associations between sentinel lymph node biopsy (SLNB) and complications among older patients who underwent breast-conserving surgery (BCS) for ductal carcinoma in situ (DCIS), and aim 2 examined whether SLNB improves long-term outcomes.

Data Source: From the Medicare-Surveillance, Epidemiology, and End Results (SEER data set, we identified women aged 67-94 diagnosed with DCIS who underwent BCS during 1998-2011 (in aim 1) and 2001-2013 (in aim 2).

Methods: In aim 1, we assessed the incidence of complications, including lymphedema, any infection, seroma, or pain within 9 months of DCIS diagnosis. We used the Mahalanobis distance (measured by patient demographics and tumor characteristics) to match the SLNB group to the non-SLNB group. We then applied generalized linear models to estimate the associations between SLNB and complications. In aim 2, we assessed the long-term risk of treated recurrence (measured by subsequent mastectomy), ipsilateral invasive breast cancer (IBC) occurrence (as reported by SEER), breast cancer mortality, and adverse effects. We used Mahalanobis matching and competing-risk survival analyses to estimate the associations between SLNB and long-term outcomes.

Results: The aim 1 sample consisted of 15 515 Medicare beneficiaries, 2409 (15.5%) of whom received SLNB. Overall, 16.8% of women who received SLNB had complications, compared with 11.3% of women who did not receive SLNB (p < .001). After matching, the groups with and without SLNB differed significantly regarding lymphedema (2.5% vs 0.5%, respectively; p < .001), any infection (12.3% vs 9.6%, respectively; p < .001), seroma (6.4% vs 3.8%, respectively; p < .001), or pain (9.8% vs 7.7%, respectively; p = .003). The use of SLNB correlated with mastectomy for the index DCIS but not radiotherapy. There were 12 776 beneficiaries in the aim 2 sample, 1992 (15.6%) of whom received SLNB. Median follow-up time after DCIS was 69 months. After matching, the 2 groups with and without SLNB did not differ significantly regarding treated recurrence (3.9% vs 3.7%; p = .62), IBC occurrence (1.4% vs 1.7%; p = .33), or breast cancer mortality (1.0% vs 0.9%; p = .86). In contrast, SLNB correlated with an increased risk of lymphedema (6.2% vs 4.0%; p < .001). Multivariate analyses of the matched sample showed that SLNB did not correlate significantly with treated recurrence, IBC occurrence, or breast cancer mortality (P values ≥ .27). However, SLNB correlated significantly with the incidence of lymphedema (adjusted hazard ratio: 1.53; 99% CI, 1.12-2.11).

Conclusions: Among older women with DCIS who received BCS, SLNB correlated with higher risks of short-term complications. Furthermore, SLNB did not correlate with improved long-term outcomes but did correlate with higher risks of long-term adverse effects. Our findings do not support using SLNB for older women with DCIS amenable to breast conservation.

Limitations: Limitations included potential unobserved confounding due to nonrandomized design as well as reliance on administrative data and billing codes.