Magnetic resonance imaging of breast cancer: factors affecting the accuracy of preoperative lesion sizing

Abstract

Accurate preoperative sizing of breast cancer with imaging modalities has a great importance in the surgical planning. To assess the influence of tumor size and histology on the accuracy of measurement of cancer local extension by magnetic resonance imaging (MRI). One hundred and eighty-six patients with primary breast cancer, for a total of 221 lesions, were included in this retrospective study. Tumors were divided into five histological groups: invasive ductal carcinoma (IDC), IDC with extensive intraductal component (EIC), invasive lobular carcinoma (ILC), ductal carcinoma in situ (DCIS), and "other histology" (mucinous, papillary, medullary, tubular, and apocrine breast cancer). Microscopic measurement of the largest diameter of tumors at pathology was chosen as reference standard and compared with MRI measurement. Concordance was defined as a difference ≤ 5 mm between MRI and pathology. The mean size of tumors at pathology was 24.8 ± 19.4 mm, while at MRI it was 29.7 ± 20 mm (P < 0.05), with a significant overestimation of MRI. MRI-pathology concordance was found in 98/221 cases (44.3%), while MRI overestimated the size of 81/221 tumors (36.7%). The extent of overestimation was significantly different among the five histological groups (P < 0.05). At multivariate analysis, DCIS histology was the factor more significantly associated with MRI-pathology discordance (P = 0.0005), while the influence of tumor dimension at pathology was less significant (P = 0.0073). DCIS histology is strongly associated with discordance between MRI and pathology sizing of breast cancer. Lesion size can also influence the accuracy of MRI measurements, but to a lesser extent.