The diagnostic methods for examining suspicious lesions in the breast are becoming less invasive, like core-needle biopsy. Yet, the risk of invasion has been reported to be up to 47% for patients with ductal carcinoma in situ (DCIS) initially diagnosed by core-needle biopsy. The value of sentinel lymph node biopsy (SLNB) for DCIS has not been clearly proved. We searched for the factors associated with invasiveness of preoperatively diagnosed DCIS, and we determined the indications for performing SLNB for patients with preoperatively diagnosed DCIS.

Between October 1997 and December 2008, we retrospectively reviewed 135 patients with DCIS that was initially diagnosed by core-needle biopsy or other biopsy methods. We compared the invasive breast cancer group, which was finally diagnosed with the pure DCIS group in regards to clinical, radiological, and pathological factors.

21.5% of the patients with initial diagnosis of DCIS were finally diagnosed with invasive breast cancer. On univariate analysis, the statistically meaningful factors for invasiveness were palpable lesion (P<0.0001), core-needle diagnosis (P=0.007), large tumor size (P=0.028), high nuclear grade (P=0.002), and negative estrogen receptor (P=0.005). On multivariate analysis, a palpable lesion was the only independent risk factor (odds ratio 3.9 (1.1 to 13.8); P=0.035). Axillar lymph node metastases were found in three patients in the invasive cancer group. There was no lymph node metastasis in the DCIS group.

We recommend that SLNB be considered in initially diagnosed DCIS with palpable lesion or high nuclear grade due to the high risk of invasiveness.