Room: Track 1
Purpose: In external beam radiotherapy (EBRT), on-treatment kV-CBCT with traditional detector sizes gives a limited field of view (FOV) diameter, resulting in truncated images of large patient anatomy. By combining projection data from the kV and MV imagers of the EBRT system, non-truncated images might be achievable while avoiding the costs of large-area detectors.
Methods: Our method uses an acquisition geometry in which a 43 cm MV detector scans the central region of the subject with a 2.5 MV beam, while a 40 cm kV detector with a 32 cm lateral offset scans the periphery at 125 kVp. Image noise is mitigated using several techniques. First, a high-efficiency multi-layer detector design diminishes noise in the MV data. Second, the use of the higher energy MV beam to scan the center of the FOV improves photon penetration through thicker anatomy. Third, an edge-preserving noise reduction algorithm combined with the poly-energetic correction technique of Joseph & Spital (1978, J. Comp. Assist. Tomogr., 2:100-108) serves both to reduce noise and to mono-energize the projections. The method was tested in simulation on a 56 cm x 40 cm abdomen phantom, and was compared both to a clinically standard half-fan kV-CBCT geometry and to a hypothetical 104 cm kV detector with the same FOV as our technique (both at 125 kVp). Simulated exposure levels corresponded to a CTDI (32 cm) of 25 mGy.
Results: Reconstructions with the proposed technique achieved a non-truncated FOV with approximately a third of the background noise of the 104 cm kV-CBCT, and superior soft tissue differentiation in the liver. The baseline half-fan kV-CBCT exhibited 5 times the noise as well as a more restricted FOV.
Conclusion: By combining kV and MV data, non-truncated on-treatment CBCT images are achievable with significant soft tissue differentiation performance and at clinically applicable doses.
Funding Support, Disclosures, and Conflict of Interest: This work was supported by Varian Medical Systems