Radiation dose dependence on subject size in abdominal computed tomography: Water phantom and patient model...

Development of patient organ dose evaluation method in computed tomography (CT) needs to model the correlation between organ dose and patient size, under various conditions of scan length, tube current lineshape, and organ location. To facilitate this task, this work was to perform a comprehensive study of the relationship between the dose to water phantom and its diameter under various settings of phantom axis, scan length, and the location across or beyond the scanned range.

A dose calculation algorithm and the published data by Li et al. [Med. Phys. 39, 5347-5352 (2012); 40, 031903 (2013); 43, 5878-5888 (2016)] were used to calculate longitudinal dose distribution D_L (z) in 10- to 50-cm diameter water phantoms undergoing constant tube current scans. The relationship between dose and phantom diameter was examined on three phantom axes (center, cross-sectional average, periphery), at seven scan lengths from 15 to 70 cm, and at eight longitudinal locations within or beyond each scan range. The water phantom results were compared to those of patient models of eight previous studies.

For the water phantoms matching the abdominal perimeters (36.3-124.5 cm) of the GSF family of voxelized phantoms, the median and range of D_L (z)(water) across scan range were consistent with those of the organ doses from the GSF phantom abdominal scans of a previous study. In 41 water phantoms (diameters 10-50 cm), D_L (z)(water) at locations inside scan range decreased with increasing phantom diameters. Exponential regression analysis of the above trend yielded regression parameters approximately consistent with those of phantom or patient models of eight previous studies. However, the usual exponential function might not be optimal for modeling the dose dependence on subject size. Inside scan range, the log(dose) vs diameter curve was non-linear on a semilogarithmic graph. Outside of scan range, dose might increase with larger subject sizes, contradicting to the exponential attenuation law. In the CT examinations of a patient population, direct modeling of organ dose dependence on patient size would be more challenging due to varying scan lengths and changing organ distances to the scan range centers.

An efficient approach to take into account the abdominal organ dose dependences on other factors is to calculate D_L (z)(water) with the water equivalent diameter, scan length, and tube current lineshape from the patient examinations, and to evaluate the organ dose to D_L (z)(water) ratio, where z is at the organ's longitudinal location. The ratio may be used for abdominal organ dose evaluation in the patient examinations. How to make use of D_L (z)(water) for organ dose evaluation in other body regions may be explored in the future.