Monte Carlo study of an innovative CBCT oscillating scanning geometry for image quality improvement.

Cone-Beam Computed Tomography (CBCT) has its main limitations in the abundance of scatter radiation and the conic beam confined on a circular trajectory which affect precision and accuracy in quantitative analysis. This work aims at investigating, by means of a validated Monte Carlo code, an innovative geometry with collimated beam and oscillating scanning geometry. The code was based on Geant4 simulation toolkit and GGEMS platform for computation time acceleration by employing GPU architecture. Simulations showed that the SPR reduced of 13.25% passing from a full field irradiation to a 3 cm height collimated beam with a phantom of 60 mm thickness in projected images. The same reduction of the field of view increased CNR between two simulated inserts made of adipose and glandular tissue of 50% and improved microcalcification visibility. Moreover, we developed a digital phantom for evaluating the accuracy and precision of HU evaluations in 3D images and image conspicuity, largely affected by cone-artifacts in conventional CBCT. Such a phantom is constituted of some PMMA slabs sandwiched between thin air layers and contains inclusions made of water, PC, PE and PVC.