Room: Exhibit Hall | Forum 9
Purpose: To explore the current state of CBCT scan protocol selection with regard to insufficient photon transmission and patient size by analyzing projection images.
Methods: At our institution, scan protocols are selected by therapists without customization. Projection images from 4914 scans from a LINAC mounted CBCT system were manually categorized by anatomic site. A subset of projections at 45Â° intervals from each scan was analyzed. Because low image intensity can result in photon-starvation artifacts in the reconstructed images, pixels with intensities <10 were identified. To account for bad detector elements, only projections with at least 20 low-intensity pixels were considered underexposed. The maximum water equivalent pathlength, measured as the equivalent thickness of water to yield the same integral attenuation, was used as a surrogate for patient size. To guide future protocol selection, pathlength cutoffs were determined that would limit the rate of underexposed projections to <5%.
Results: In current practice many projections are underexposed, particularly with large patients or when the anatomical variation is substantial. Underexposure was identified in 16% of all projections. Of those underexposed using the â€˜Headâ€™ protocol, 88% were categorized as â€˜Neckâ€™ and the underexposure is likely caused by the shoulders. Of the projections acquired with the â€˜Pelvisâ€™ protocol, 9% were underexposed. Even the highest exposure protocol (â€˜Pelvis Obeseâ€™) may be insufficient for the largest patients resulting in underexposure in 8% of projections. The maximum water equivalent pathlength cutoffs to reduce underexposure to <5% were calculated to be 25, 34.5, 40, and 43cm for the â€˜Headâ€™, â€™Thoraxâ€™, â€™Pelvisâ€™, and â€™Pelvis Obeseâ€™ protocols, respectively.
Conclusion: This work indicates that a subset of CBCT scans suffer underexposure due to suboptimal scan parameter selection. Future work to automate the scan protocol selection process could potentially increase efficiency while improving imaging quality for patients with a wide range of body sizes.
Funding Support, Disclosures, and Conflict of Interest: Funding was provided in part by NIH T32EB002103 and Varian Medical Systems. Hania Al-Hallaq receives royalties and licensing fees for computer-aided diagnosis technology through the University of Chicago.