Risk Factors for Positive Resection Margins Following Breast-Conserving Surgery.

Background The incidence of margin re-excision following breast conserving surgery (BCS) is a quality measure in the National Health Service. The threshold is less than 20% of all BCS procedures. Despite three decades of studies and a wealth of literature identifying multiple factors associated with increased risk for margin involvement, an accepted threshold rate affecting one in five procedures remains high. Aim The aim of the study was to identify adverse features that continue to compromise successful margin clearance despite the recognition of risk factors and the implementation of strategies designed to minimise those risks. Methods All margin re-excisions following BCS for invasive breast carcinoma and ductal carcinoma in situ (DCIS) performed from October 2013 to September 2018 were retrieved from the database of a single institution. A total of 1379 patients underwent BCS during the period considered, 194 of which needed margin re-excision. Radiological investigations and histopathology reports for each patient were retrieved. Lesion size and focality on mammogram, ultrasound (US) scan, and magnetic resonance imaging (MRI), and histopathologic tumour characteristics were recorded and analysed. Results The overall re-excision rate was 14.06% (194/1379 patients). Margin re-excisions cleared 69% (134/194) of wide local excision cavities that had at least one involved margin. 53% (103/194) of patients had no further disease after one attempt at re-excision and 15.9% (31/of 194) had further disease, which was cleared after re-excision. Another 15.9% (31/194) had disease within the shave with involved margins. In this sub-group the presence of DCIS at the new resection margin accounted for 90.3% (28/31) of cases, 3% (1/31) were invasive ductal carcinoma (IDC) and 6% (2/31) were unrecorded. In the sub-group of patients who had an excised margin with pathology and a new clear margin (15.9% of all re-excisions), DCIS was found in 61% (19/31) of cases, IDC in 12.9% (4/31), invasive lobular carcinoma (ILC) in 6% (2/31) of cases, lobular neoplasia (LN) in 12.9% (4/31), mixed IDC and DCIS in 6% (2/31)of cases. The correlation between imaging size and actual histopathological size has shown a statistically significant discrepancy in this cohort. The median size on histology was 22 mm, compared to a median size of 16 mm on mammography, 14 mm on ultrasound, and 17 mm on MRI.  Conclusion According to our cohort of patients, the most consistent factor associated with a re-excision was the presence of DCIS at the resection margin, whether pure DCIS or IDC admixed with DCIS. The comparison between tumour size on imaging and final histopathological size revealed the best correlation with mammogram followed by US. The weakest correlation was with MRI.