RESUMO
INTRODUCTION: High-risk or B3 breast lesions are considered lesions of uncertain malignant potential and comprise between 5 and 12% of initial biopsy results. We sought to perform a systematic review and meta-analysis of studies published within the last twenty years to determine the pooled Positive Predictive Value (PPV) of VAB in selected B3 lesions. METHODS: The study report is based on the guidelines of PRISMA and Meta-Analysis of Observational Studies in Epidemiology. OUTCOMES: The primary outcome of this study was to determine the PPV of VAB in determining final histological diagnosis in B3 breast lesions using pooled estimates. The secondary outcomes were to determine if needle gauge or the re-classification of Lobular Carcinoma in Situ(LCIS) introduced in 2012 influenced pooled estimates. RESULTS: 78 studies incorporating 6,377 B3 lesions were included in this review, 1214 of which were upgraded to DCIS or invasive malignancy following surgical excision(19%). The pooled PPV of VAB in Atypical Ductal Hyperplasia(ADH) and Lobular Neoplasia(LN) were 0.79(CI 0.76-0.83) and 0.84(CI 0.8-0.88). VAB of Flat Epithelial Atypia(FEA), radial scar and papillary lesions with/without atypia all had a pooled PPV >90% (underestimation rates 7%, 1%, 5% and 3% respectively). Needle gauge size and the change in LCIS classification did not appear to influence underestimation rates on subgroup analysis. CONCLUSION: Results from this meta-analysis suggests it is reasonable to perform VAB as definitive treatment for certain B3 lesions, specifically LN, FEA, radial scar, and papillary lesions when specific criteria are fulfilled. Surgical excision should continue as the mainstay of treatment for ADH.