Premium
SU‐E‐T‐689: Semi‐Automated GPU‐Based Monte Carlo Dose Calculation for Total Body Irradiation
Author(s) -
Folkerts M,
Ouyang L,
Jia X,
Jiang S,
Gu X
Publication year - 2015
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4925052
Subject(s) - computer science , monte carlo method , software , attenuation , computational science , nuclear medicine , computer graphics (images) , computer vision , artificial intelligence , physics , optics , mathematics , medicine , statistics , programming language
Purpose: To enable fast and accurate TBI dose calculations for research and clinical applications. Methods: We adapted GPU‐based Monte Carlo (MC) dose calculation library written for patient specific IMRT quality assurance to enable us to compute dose delivered to a patient during total body irradiation (TBI). The full CT dataset of the patient is loaded into the system along with organ contours, specifically, the body and lung contours, as well as rectangular contours denoting compensator sizes and locations. The CT data is automatically cropped by setting all values outside of the body contour to air HU value. The lung and compensator contours are automatically voxelized, projected and rotated by 45 degrees onto the 40×40 cm fluence map plane and weighted by the appropriate attenuation value. The mirror image of the fluence map is also computed for the opposite beam configuration (P‐A for A‐P, or R‐L for L‐R). The fluence maps are then included as input, along with the auto‐segmented material types and density values from the CT dataset, into the GPU‐based MC calculation. The Result is the full body MC calculated TBI dose for the patient which can then be used for various research studies. The users upload their data to the GPU server using FTP. The interface to the calculation software is in the form of an iPython Notebook, enabling the user to access the GPU server through a common web browser to input various parameters, monitor the process, and analyze results. Results: The developed software enables researchers and clinicians to compute MC TBI dose results in less than 2 minutes given lung and body contours and compensator transmission coefficients. Conclusion: We successfully designed and developed a semi‐automated MC TBI dose calculation software to enable research and clinical studies.