Room: AAPM ePoster Library
Purpose: Determine the effect of iterative CBCT (iCBCT) reconstruction and noise suppression on CBCT SNR, CNR, and spatial resolution to enable optimization of imaging protocols for improved low contrast or dose sparing.
Methods: CBCT images were acquired with a version 2.7 TrueBeam linac. All images were acquired of a Catphan 504 phantom centered on the interface of the low contrast and geometry modules. The standard Head, Pelvis, and Image Gently protocols were acquired. The default kV and mAs settings were used, then reduced to evaluate the potential of reducing dose while maintaining the various imaging metrics with iCBCT reconstruction. After acquisition and reconstruction with the standard algorithm, images were retrospectively reconstructed using the iCBCT algorithm with the medium and high noise suppression options. Images were analyzed using an in-house ImageJ plugin that calculated MTF, CNR, and SNR.
Results: iCBCT improves CNR and SNR when compared to the standard reconstruction based on filtered back projection. The magnitude of this effect is modified by the choice of noise suppression option. This improved the SNR and 1% CNR six fold for the head protocol and by four and eight fold, respectively, for the pelvis protocol. There was no appreciable improvement in the image gently protocol as the total signal was too low. The kV and mAs variations showed what minimum photon output levels were necessary to resolve different low contrast levels.
Conclusion: iCBCT is a new reconstruction algorithm available for use in IGRT. The resulting improvement in CNR and SNR can enable more confident registrations and potentially lead to more specialized protocols for dose reduction or improvement of low contrast visibility. This algorithm can help overcome CBCTs inherent lower contrast, and, provided sufficient photon fluence for the size/contrast of a target, can enable the visualization of targets previously difficult to resolve.