Abstract
BACKGROUND: Axillary lymph node status remains the most powerful prognostic indicator in invasive breast cancer. Ductal carcinoma in situ (DCIS) is a non-invasive disease and does not spread to axillary lymph nodes. The presence of an invasive component to DCIS mandates nodal evaluation through sentinel lymph node biopsy (SLNB). Quantification of the necessity of upfront SLNB for DCIS requires investigation. The aim was to establish the likelihood of having a positive SLNB (SLNB+) for DCIS and to establish parameters predictive of SLNB+. METHODS: A systematic review was performed as per the PRISMA guidelines. Prospective studies only were included. Characteristics predictive of SLNB+ were expressed as dichotomous variables and pooled as odds ratios (o.r.) and associated 95 per cent confidence intervals (c.i.) using the Mantel-Haenszel method. RESULTS: Overall, 16 studies including 4388 patients were included (mean patient age 54.8 (range 24 to 92) years). Of these, 72.5 per cent of patients underwent SLNB (3156 of 4356 patients) and 4.9 per cent had SLNB+ (153 of 3153 patients). The likelihood of having SLNB+ for DCIS was less than 1 per cent (o.r. <0.01, 95 per cent c.i. 0.00 to 0.01; P < 0.001, I2 = 93 per cent). Palpable DCIS (o.r. 2.01, 95 per cent c.i. 0.64 to 6.24; P = 0.230, I2 = 0 per cent), tumour necrosis (o.r. 3.84, 95 per cent c.i. 0.85 to 17.44; P = 0.080, I2 = 83 per cent), and grade 3 DCIS (o.r. 1.34, 95 per cent c.i. 0.80 to 2.23; P = 0.270, I2 = 0 per cent) all trended towards significance in predicting SLNB+. CONCLUSION: While aggressive clinicopathological parameters may guide SLNB for patients with DCIS, the absolute and relative risk of SLNB+ for DCIS is less than 5 per cent and 1 per cent, respectively. Well-designed randomized controlled trials are required to establish fully the necessity of SLNB for patients diagnosed with DCIS. REGISTRATION NUMBER: CRD42021284194 (https://www.crd.york.ac.uk/prospero/).