Initial diagnosis and treatment of women with breast cancer is based on the imaging findings. Anecdotal experience suggests that the quality of breast imaging reports is variable; however, systematic evaluation of the content of reports has not been documented to date. We present an audit of the breast imaging reports of all new breast cancer cases referred to a multidisciplinary breast centre during 2004, based on 244 imaging reports from 253 cases. We focus on the quality of imaging reports from the perspective of completeness, concordance with standards, and provision of information considered relevant to clinical decision-making. The audit shows that many reports do not provide key information, and that there are substantial variations in the quality of reports between breast screening services (as part of a coordinated national programme) and community-based radiology services. About one-quarter of all reports do not provide an imaging diagnosis, and only half of all imaging reports are concordant with standards for structured reporting. The least reported variables were breast density category (reported in 24%), lesion depth (37%), lesion shape (55% for mammography, 39% for ultrasound), and location (59%). The most frequently provided information was mammography lesion type (99.6%), sonographic lesion size (90.4%), and recommendation for further investigation (89%). The vast majority of reports from screening services used structured reporting, and these were more likely to provide the information recommended in standards than were reports from community-based radiologists. This work indicates that the quality (content and completeness) of breast imaging reports, particularly community-based radiology reports, is not in line with standards. The clinical implications of these findings warrant further study.