Abstract
METHODS: A cross-sectional study identified cases of mucinous breast carcinoma from pathology records (2004-2012). Two radiologists classified imaging features by consensus and two pathologists classified cases into pure or mixed subtypes. Bi-variable analyses were performed using relevant statistical tests. RESULTS: We identified 80 lesions in 77 female patients (median age 65 years, range 29-88): 58 lesions on mammograhy, 72 on ultrasound, and 25 on MRI. Statistically significant findings (p < 0.05) are as follows. On mammography, tumour margins tended to be indistinct (12, 48%) and spiculated (11, 44%) for pure and mixed lesions, respectively. Pure mucinous masses were less microcalcified (23, 77%) and mixed masses equally so. On ultrasound, pure tumours tended towards an irregular or oval shape (44, 42%) with mixed tumours having an irregular shape (78%). More pure tumours (53%) had posterior acoustic enhancement than mixed lesions (33%), and all pure tumours lacked posterior acoustic shadowing. Pure lesions had a heterogeneous echo pattern more than mixed tumours (78% vs 39%). On MRI, pure tumours tended towards a persistent kinetic curve (42%) whereas mixed tumours predominantly had a washout pattern (75%). Most pure tumours were T(2) hyperintense (83%) whereas mixed lesions were T2 isointense or hyperintense (61%, 23%), respectively. CONCLUSION: An analysis of imaging features can help to infer underlying histology of pure and mixed forms of mucinous breast carcinoma. ADVANCES IN KNOWLEDGE: Pure mucinous carcinomas present less suspicious imaging features than mixed mucinous carcinomas and could be mistaken for non malignant lesions. An imaging analysis of mucinous breast carcinoma can help infer their underlying histology.