SU‐FF‐I‐111: Exploration of the Scatter Distribution in Cone‐Beam CT Using Monte Carlo Techniques

Abstract

Purpose: To validate our current Monte Carlo (MC) model and explore the scatter distribution in cone‐beam CT (CBCT) projection images. The MC data will be used to garner a better understanding of the relationship between imaging parameters and the resulting scatter distribution. Method and Materials: Measured images to validate the CBCT MC model where collected using a flexible bench‐top CBCT system. The x‐ray tube was modeled using the BEAMnrc MC code system and the imaging geometries using a modified DOSXYZnrc program that differentiates primary from scattered particles. Two objects were investigated, a 16.4 and 30.6 cm diameter water cylinder. Projections were simulated for three different cone angles (1.4°, 5.7°, and 11.3°), three different source‐to‐axis distances (SAD: 50, 75, and 100 cm), and four different axis‐to‐detector distances (ADD: 9, 18, 30, and 56 cm). All simulations and measurements were done with an energy of 100 kVp. In order to validate the simulated scatter distributions experiments were conducted using beam‐blocking techniques to estimate the scatter for a 16.4 cm and 30.6 cm water cylinder for a single geometry (SAD=100, ADD=56 cm). Results: The simulated projections had a mean local absolute difference of 3.7 +/− 2.1% and 8.7 +/− 3.8% with the measured projections for the 16.4 and 30.6 cm diameter water cylinder respectively. Scatter simulations had a mean local absolute difference of 9.6 +/− 4.1% and 5.4 +/− 3.3% with the scatter measurements for the 16.4 cm and 30.6 cm diameter water cylinders respectively. Conclusion: The results support the use of MC to gain a further understanding of scatter and develop new techniques to correct for scatter induced image quality artifacts in CBCT. The scatter database created will be a valuable tool in exploring the relationships between imaging geometry parameters and the scatter distribution.