Identifying a subset of patients with DCIS who have a low likelihood of residual disease at surgical excision following a core needle biopsy.

Background and Objectives: Current randomized controlled trials are investigating the outcomes of non-surgical treatment for patients with ductal carcinoma in situ (DCIS). We sought to evaluate pre-operative factors associated with no residual disease at definitive resection following a core needle biopsy (CNB) diagnosis of DCIS. Methods: Eight hundred and thirty-four operations for DCIS were performed at our institution between January 2004 and October 2014. We evaluated patient and biopsy tumor characteristics to determine pre-operative factors associated with no residual disease at surgical resection using uni- and multivariable analyses. Results: Sixty-nine patients (8%) had no residual disease on final pathology. On multivariable analysis, low- or intermediate-grade lesions, < 1cm in size on mammography, and lesions where >= 90% of calcifications were removed correlated with finding no residual disease on final pathology, c-statistic 0.84. Of the 14 patients with all three low-risk factors, 36% had no residual disease on final pathology. Conclusions: Although our multivariable analysis performed well, its clinical utility would be limited as we were unable to identify a subset of patients with DCIS in whom the probability of finding no residual disease is low enough to consider routine use of non-surgical management.