Breast density does not affect breast cancer tumor size assessment: A comparison of radiologic versus pathologic measurement by different imaging modalities across breast densities

Background: Tumor size is an important parameter in breast cancer staging. Definitive tumor size is determined by measurement of the pathologic specimen. However, prior to surgery, size must be assessed by imaging with mammography (MMG), ultrasound (US), or magnetic resonance imaging (MRI). Discrepancies between imaging-assessed and pathologic size are not uncommon. Breast density decreases the sensitivity of MMG, and may affect image-based tumor size assessment. Aim: To compare tumor size assessed by the different imaging modalities to pathologic size across breast densities. Material & methods: This was a retrospective analysis of 183 female patients (197 breast cancers) diagnosed and operated for primary breast cancer at a single center. Tumor size measurements were collated for each available imaging modality and compared with measurements from pathologic specimens. Breast density was assessed on MMG using the Breast Imaging Reporting and Data System. Results: Mean pathologic tumor size was 23.0 ± 19.3 mm. Mean tumor size did not differ significantly with MMG (22.3 ± 16.6 mm; P = 0.165) or MRI (23.4 ± 19.2 mm; P = 0.620). However, US significantly underestimated mean tumor size (15.2 ± 8.6 mm; P = 0.0001 vs pathology). Breast density did not affect the accuracy of tumor size assessment by any imaging modality. Conclusions: US may underestimate breast tumor size. Treatment decisions that take into account tumor size can be made equally reliably in patients with high or low breast density.