Diagnostic reference levels for digital mammography in New South Wales.

INTRODUCTION: This work aims to explore radiation doses delivered in screening mammography in Australia, with a focus on whether compressed breast thickness should be used as a guide when determining patient derived diagnostic reference levels (DRLs).

METHODS: Anonymized mammograms (52,405) were retrieved from a central database, and DICOM headers were extracted using third party software. Women with breast implants, breast thicknesses outside 20-110 mm and images with incomplete exposure or quality assurance (QA) data were excluded. Exposure and QA information were utilized to calculate the mean glandular dose (MGD) for 45,054 mammograms from 61 units representing four manufacturers using previously well-established methods. The 75th and 95th percentiles were calculated across median image MGDs obtained for all included data and according to specific compressed breast thickness ranges.

RESULTS: The overall median image MGD, minimum, maximum were: 1.39, 0.19 and 10.00 mGy, respectively, the 75th and 95th percentiles across all units' median image MGD for 60 ± 5 mm compressed breast thickness were 2.06 and 2.69 mGy respectively. Median MGDs, minimum, maximum, 75th and 95th percentiles were presented for nine compressed breast thickness ranges, DRLs for NSW are suggested for the compressed breast thickness range of 60 ± 5 mm for the whole study and three detector technologies CR, DR, and photon counting to be 2.06, 2.22, 2.04 and 0.79 mGy respectively.

CONCLUSION: MGD is dependent upon compressed breast thickness and it is recommended that DRL values should be specific to compressed breast thickness and image detector technology.