eduzhai > Applied Sciences > Engineering >

Deep learning reconstruction of digital breast tomosynthesis images for accurate breast density and patient-specific radiation dose estimation

  • Save

... pages left unread,continue reading

Document pages: 17 pages

Abstract: The two-dimensional nature of mammography makes estimation of the overallbreast density challenging, and estimation of the true patient-specificradiation dose impossible. Digital breast tomosynthesis (DBT), a pseudo-3Dtechnique, is now commonly used in breast cancer screening and diagnostics.Still, the severely limited 3rd dimension information in DBT has not been used,until now, to estimate the true breast density or the patient-specific dose.This study proposes a reconstruction algorithm for DBT based on deep learningspecifically optimized for these tasks. The algorithm, which we name DBToR, isbased on unrolling a proximal-dual optimization method. The proximal operatorsare replaced with convolutional neural networks and prior knowledge is includedin the model. This extends previous work on a deep learning-basedreconstruction model by providing both the primal and the dual blocks withbreast thickness information, which is available in DBT. Training and testingof the model were performed using virtual patient phantoms from two differentsources. Reconstruction performance, and accuracy in estimation of breastdensity and radiation dose, were estimated, showing high accuracy (density<+ -3 ; dose <+ -20 ) without bias, significantly improving on the currentstate-of-the-art. This work also lays the groundwork for developing a deeplearning-based reconstruction algorithm for the task of image interpretation byradiologists.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×