Dose reduction in mammography by using imaging plate technology: A retrospective analysis.

Conventional mammography is a decisive tool in detecting breast cancer. Continuous efforts are undertaken in order to further improve the image quality as well as to reduce the applied doses. The purpose of our study was to compare diagnostic image quality of dose reduced computed mammography with a new needle-based detector system to full dose powder imaging plates. We retrospectively compared 360 randomly chosen mammographies performed on a GE Senographe DMR running the Agfa DX-M needle-based imaging plate system (NIP) with their preliminary examinations which were acquired at standard dose with the same GE mammography device and an Agfa CR85-X powdered storage phosphor imaging plate system (PIP). NIP-based mammographies were about 29.8 % dose-reduced. The preliminary examinations had to be performed not earlier than 2 years before the recent investigations. Exclusion criteria were changes in ACR level and appearance of the scored targets and not optimally positioned and exposed mammographies. The images were blinded and read separately twice by 2 mammography experts according to a 3-point score on diagnostic image quality and the visualization of parenchyma, cysts, fibroadenomas, physiologic lymph nodes, solitary microcalcifications and macrocalcifications. Dose reduced NIPs showed a significantly better visualization of parenchyma at ACR II/III and solitary microcalcifications at ACR I-III mammographies (p < 0.001) whereas the difference in scoring macrocalcifications, cysts, fibroadenomas and physiologic lymph nodes was not significant. The reading showed an excellent intra- (r = 0.97/0.94) and interobserver agreement (r = 0.92). With computed mammography using the needle-based detector system a significant dose reduction is possible without loss of diagnostic image quality.

Informations de copyright

Copyright © 2020 Elsevier B.V. All rights reserved.

Déclaration de conflit d'intérêts

Declaration of Competing Interest The authors declare that there are no conflicts of interest.