A hybrid approach for accurate estimation of the scatter component in x-ray CT combining experimental measurements and Monte Carlo simulations

The corruption of projection data in x-ray CT with scattered radiation decreases low contrast detectability (LCD) in reconstructed images. Scatter removal is therefore mandatory in CT imaging because of the need to generate images presenting clinically acceptable LCD. The assessment of the scatter component in CT is an active research area in quantitative imaging, however very few contributions have been made in this domain. Thorough knowledge on the scatter distribution is essential for development of scatter correction algorithms. Conventional experimental measurement (EM) based methods using one lead bar as a blocker are not able to extract scattered profiles in a single exposure. A novel method previously proposed by our group using an array of blockers seems to be a reasonable way to efficiently measure scatter profiles. However, the lead blocker is a secondary source of scattered radiation which propagates substantial errors in the EM procedure. Monte Carlo (MC) simulation is the only way enabling to estimate and correct the errors arising from EM. In this study, a hybrid approach combining EM and MC simulations is proposed to accurately calculate scatter profiles in x-ray CT scanners. The hybrid method involves two steps: (a) EM using lead arrays to calculate the scattered radiation profile and (b) MC simulations to estimate and correct scattered radiation induced from lead arrays. The MC simulator was validated through comparison with EM. MC simulations of a 64-slice CT scanner geometry showed that the magnitude of scattered radiation arising from the lead array used for measurement of scatter profile is 54% of total scattered radiation, which can not be ignored. The measured scatter profiles in different tube voltages were corrected using the suggested hybrid approach. In conclusion, the hybrid method seems to be a precise and robust method for accurate estimation of scattered radiation profile. The method is still undergoing some improvements to reduce the- - computational time and to optimize the combination of MC and measured data.