Monte Carlo CT Dose Calculation: A Comparison Between Experiment and Simulation Using ARCHER-CT
T. Liu,X. Du,L. Su,Y. Gao,W. Ji,D. Zhang,J. Shi,B. Liu,M. Kalra,X. Xu
DOI: https://doi.org/10.1118/1.4889155
IF: 4.506
2014-01-01
Medical Physics
Abstract:Purpose: To compare the CT doses derived from the experiments and GPU-based Monte Carlo (MC) simulations, using a human cadaver and ATOM phantom. Methods: The cadaver of an 88-year old male and the ATOM phantom were scanned by a GE LightSpeed Pro 16 MDCT. For the cadaver study, the Thimble chambers (Model 10×5−0.6CT and 10×6−0.6CT) were used to measure the absorbed dose in different deep and superficial organs. Whole-body scans were first performed to construct a complete image database for MC simulations. Abdomen/pelvis helical scans were then conducted using 120/100 kVps, 300 mAs and a pitch factor of 1.375:1. For the ATOM phantom study, the OSL dosimeters were used and helical scans were performed using 120 kVp and x, y, z tube current modulation (TCM). For the MC simulations, sufficient particles were run in both cases such that the statistical errors of the results by ARCHER-CT were limited to 1%. Results: For the human cadaver scan, the doses to the stomach, liver, colon, left kidney, pancreas and urinary bladder were compared. The difference between experiments and simulations was within 19% for the 120 kVp and 25% for the 100 kVp. For the ATOM phantom scan, the doses to the lung,more » thyroid, esophagus, heart, stomach, liver, spleen, kidneys and thymus were compared. The difference was 39.2% for the esophagus, and within 16% for all other organs. Conclusion: In this study the experimental and simulated CT doses were compared. Their difference is primarily attributed to the systematic errors of the MC simulations, including the accuracy of the bowtie filter modeling, and the algorithm to generate voxelized phantom from DICOM images. The experimental error is considered small and may arise from the dosimeters. R01 grant (R01EB015478) from National Institute of Biomedical Imaging and Bioengineering.« less