Premium
SU‐D‐BRB‐06: G4DBR: A Fast Geant4‐Based Monte Carlo Dosimetry Platform for Brachytherapy
Author(s) -
Afsharpour H,
Carrier J,
DˈAmours M,
Enger S A.,
Landry G,
Poon E,
Verhaegen F,
Beaulieu L
Publication year - 2011
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 - 10.1118/1.3611524
Subject(s) - dosimetry , brachytherapy , monte carlo method , dicom , nuclear medicine , computer science , medical physics , physics , medicine , radiation therapy , mathematics , statistics , artificial intelligence , radiology
Purpose: To present, G4DBR, a fast Geant4‐based Monte Carlo (MC) dosimetry platform for brachytherapy. The special case of low dose rate (LDR) brachytherapy is considered here. Methods: Geant4 9.3 has been used for designing a new MC platform for calculating the dose distribution in brachytherapy called G4DBR. This code is capable of dealing with the DICOM RT format to build a virtual representation of each patient with the full multi‐seed configuration. The dose is scored both to medium and to water using track‐length estimator. The dose distributions are extracted in 3ddose format for visualization or to calculate the DVHs. Results: One prostate permanent 125I seed (PPSI) and one breast permanent 103Pd seed implant (BSPI) patient have been selected for evaluating the performance of G4DBR on a 2.93 GHz Intel Xeon Nehalem single core. Post‐implant dosimetry of those cases are performed in a 2 mm3 mesh for comparison with BrachyDose and MCPI. 45 seconds were required for G4DBR to reach a statistical uncertainty of 2% on PTV dose in PPSI. Note that for a similar precision, BrachyDose requires 30 seconds on 3 GHz Woodcrest (Thomson et al. Med. Phys. 2010) while MCPI needs 59 seconds on a single 2.4 GHz Pentium 4 CPU (Chibani et al. Med. Phys. 2005). G4DBR takes 114 seconds to attain a statistical uncertainty of less than 2% in the BPSI case. Conclusions: G4DBR is accurate and fast enough for clinical purposes. G4DBR is able to achieve good calculation speeds comparable with BrachyDose and MCPI. Indeed, a statistical uncertainty of less than 2% is attained in 45 seconds in a prostate case while 118 seconds were needed in BPSI to achieve 0.5% of uncertainty. Further developments will include the incorporation of high dose rate (HDR) dosimetry and a user‐friendly GUI for G4DBR.