Accuracy of Breast Magnetic Resonance Imaging Compared to Mammography in the Preoperative Detection and Measurement of Pure Ductal Carcinoma In Situ: A Retrospective Analysis

Abstract

Ductal carcinoma in situ (DCIS) hinders imaging detection due to multifocal appearance and discontinuous growth. Preoperative determination of its extent is therefore challenging. Aim of this study was to investigate the additional benefit of breast magnetic resonance imaging (MRI) to mammography (MG) in the diagnosis of DCIS according to size and grading. Retrospective analysis of 295 patients with biopsy-proven, pure DCIS. Mean patient age was 57.0 years (27-87 years). All patients obtained MG. Additional MRI was performed in 41.7% (123/295). Mammographic breast density, background parenchymal enhancement (BPE), tumor size and grading were analysed. Tumor size on MG and MRI were compared to histopathological size of the surgical specimen. Mean tumor size was 39.6 mm. DCIS was occult on MG in 24.4% (30/123) and on MRI in 1.6% (2/123). Size was underestimated by 4.6 mm (mean) mammographically. DCIS was high grade in 54.5% (67/123), intermediate grade in 40.7% (50/123) and low grade in 4.9% (6/123). MG was exact regarding tumor size in low grade DCIS, underestimated intermediate grade DCIS by 1 mm (median) and high grade DCIS by 10.5 mm. MRI overestimated low grade DCIS by 1 mm (median), was exact regarding intermediate grade DCIS and underestimated high grade DCIS by 1 mm. BPE did not influence tumor detection and measurement. MRI outperforms MG in the detection and size estimation of DCIS and can reduce positive margin rates.