Prediction of underestimated invasiveness in patients with ductal carcinoma in situ of the breast on percutaneous biopsy as rationale for recommending concurrent sentinel lymph node biopsy.

AIM: To develop a model to predict invasion and improve the indication of concurrent sentinel lymph node biopsy (SLNB) for patients with ductal carcinoma in situ (DCIS) on minimally invasive biopsy.
METHODS: We evaluated the data of 205 patients with DCIS in minimally invasive biopsy specimens. Clinical, radiological and histological variables were assessed in order to identify predictors of invasive carcinoma in final pathology using logistic regression analyses. We developed and retrospectively tested an algorithm to indicate concurrent SLNB.
RESULTS: Invasiveness was underestimated in 18.0% (37 of 205). Univariate analysis revealed the following significant risk factors: lesion palpability, a mass lesion on ultrasound, the presence of a mammographically detectable mass, architectural distortion or density, a BI-RADS score of 5, a lesion diameter ≥50 mm, and ≥50% of histologically affected ducts. With a palpable mass, which remained the only independent predictor of invasion after multivariate adjustment, and the presence of at least three of the remaining five risk factors, the probability of invasion was 56.0%. If the prediction model had been used to indicate SLNB 9.8% (20 of 205) of patients could have been benefited (i.e. spared unnecessary or correctly recommended concurrent SLNB) compared to the factual performed SLNB procedures. Those patients with pure DCIS treated with breast conserving surgery (BCS) benefited most with a relative risk reduction of nearly 50% for unnecessary SLNB.
CONCLUSION: The prediction model could rationally guide an informed discussion about risks and benefits of concurrent SLNB in patients with DCIS on minimally invasive biopsy.