Abstract P6-02-01: Meta-Analysis of Ductal Carcinoma In Situ (DCIS) on Core Needle Biopsy — Underestimation and Predictors of Invasion

Abstract Background: A diagnosis of ductal carcinoma in situ (DCIS) on core needle biopsy (CNB) may represent an underestimate (or understaging) of invasive breast cancer (IBC). This is identified on excision histology in 4%-35% of CNB diagnoses of DCIS. It is of clinical significance as it may prevent informed discussion about management options for IBC and may mean that more than one surgical procedure is required. We systematically review the literature on CNB to report pooled estimates for underestimation of IBC following CNB diagnosis of DCIS and identify preoperative predictors of invasion. Methods: Studies were identified by searching MEDLINE and were evaluated against predetermined inclusion criteria: data were extracted independently by 2 authors. We calculated study-specific proportions for DCIS underestimates. Using meta-regression (random effects logistic models) we investigated the association between study-level preoperative variables and understaged IBC. Results: 47 studies met eligibility criteria reporting 6,213 cases of DCIS on CNB with excision histology as the reference standard. Table 1: Pooled underestimation rates for There were 1,482 underestimates (DCIS on CNB and IBC on excision); pooled estimate 25.9%(95% CI 22.2, 29.9). Preoperative variables associated with underestimation included ultrasound-guidance (vs stereotaxis), 14Gautomated (vs 11GVAB), high grade CNB lesion (vs nonhigh grade), size >20mm, BIRADS 4/5 (vs BIRADS3), mammographic mass (vs calc only) and palpability. Conclusion: Using available evidence we estimate that around 26% of CNB DCIS diagnoses represent understaged IBC. We identified preoperative variables associated with higher risk of understaging. This information can be used for preoperative discussion of management options, including possible sentinel node biopsy at the time of excision. Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P6-02-01.