Can breast models be simplified to estimate scattered radiation in breast tomosythesis? (original) (raw)

Medical Imaging 2019: Physics of Medical Imaging, 2019

Abstract

Scattered radiation can represent a large portion of the total signal recorded at the image receptor in certain x-ray breast imaging systems, such as digital breast tomosynthesis (DBT). For many years, Monte Carlo (MC) simulations have represented the golden approach to estimate the scatter field, initially with simple models and more recently with anthropomorphic phantoms. However, it is unclear how the scattered radiation varies between such models. Further knowledge of the scatter behaviour can help to develop faster and simpler scatter field estimation approaches, which are highly demanded in virtual clinical trial (VCT) strategies. In this work, the scattered radiation estimated for several homogeneous breast models is compared against that from textured breast phantoms. By means of MC simulations, scatter fields are investigated under the same DBT scenario. Results for a quasi-realistic breast model suggest that homogeneous models with same shape and glandularity can approximate the scattered radiation produced by a heterogeneous phantom with a median error of 2%. Simpler models with semi-circular shapes, which reduces the complexity in the scatter field estimation and decrease the computational time, show good approximation in the central region of the breast although larger discrepancies are observed in the peripheral region of the breast image.

Oliver Diaz hasn't uploaded this paper.

Let Oliver know you want this paper to be uploaded.

Ask for this paper to be uploaded.