Image extrapolation for patient-specific CT dose determination based on scout images

Monte Carlo (MC) simulation based on a patient's computed tomography (CT) images is a promising way to retroactively determine patient-specific dose from CT scans. CT scans generally include only a portion of the patient's body; however, photon interactions in regions adjacent to the scan region contribute to the overall scan dose. Thus, the CT images alone, which do not include these adjacent regions, are insufficient for determining dose. In fact, dose underestimation, especially at the scan edge, occurs when the adjacent regions are not accounted for. In this work, the dose underestimation without the scatter region is demonstrated, and the size of the scatter region required to provide sufficient simulated scatter was determined with mathematical phantom studies. For a simple cylindrical water phantom, up to a 25% underestimation of dose was found when no scatter region was used, and a 40 mm scatter region was determined to be required to eliminate this error for scans performed with both the 40 mm and 5 mm collimation sizes. In addition, four different image extrapolation methods based on CT images and scout images were proposed for a chest CT scan of an anthropomorphic phantom. The dose was calculated with the chest images only and the chest images plus the four types of extrapolated images. The results were compared with the dose calculated using whole body images of the anthropomorphic phantom. The image extrapolation methods, especially the ones based on scout images, were shown to improve the dose calculation accuracy under both step-shoot scan mode and helical scan mode.