MO‐E‐AUD B‐04: Fast, Accurate Photon Beam Accelerator Modeling Using BEAMnrc and VMC++: A Systematic Investigation of Variance Reduction and Efficiency Enhancing Methods and Cross‐Section Data

M. Fragoso, I. Kawrakow, B. Faddegon, T. Solberg, I. Chetty

Research output: Contribution to journalArticle

Abstract

Purpose: To report on the accuracy of cross‐section data in BEAMnrc and on the performance of variance reduction and efficiency enhancing techniques for fast, accurate linac simulations using the BEAMnrc and VMC++ code systems. Method and Materials: BEAMnrc and VMC++ were used to simulate a 6 MV photon beam from a Siemens Primus linac. Phase space (PHSP) files were generated for a range of field sizes, from 10×10 to 40×40 cm2. BEAMnrc parameters under investigation were grouped by: i) photon and bremsstrahlung cross‐sections; ii) approximate efficiency improving techniques (AEIT); iii) variance reduction techniques (VRT); iv) VRT (bremsstrahlung splitting) with AEIT (range rejection). Efficiencies were obtained for the mean energy, fluence, angular and spectral distributions and PHSP files were subsequently used as input for DOSXYZnrc‐based phantom dose calculations; these calculations were verified against measurements. Results: Efficiencies were calculated for the various VRT/AEIT combinations in BEAMnrc, relative to simulations without VRT/AEIT, namely: (a) 935 (∼111 min. on a single 2.6 GHz CPU) and 200 for 10×10 and 40×40 resp. using directional bremsstrahlung splitting (DBS) and no electron splitting, (b) 420 and 175 for 10×10 and 40×40 resp. using DBS and electron splitting combined with augmented range rejection, a technique recently introduced in BEAMnrc. Calculations with VMC++ produced efficiencies of 1400 (∼6 min. on a single CPU) for 10×10 versus BEAMnrc (no VRT/AEIT). Noteworthy differences (±1–3%) were observed with the NIST bremsstrahlung cross‐sections compared with those of Bethe‐Heitler (default). However, MC calculated dose distributions (using all combinations of VRT/AEIT and cross‐section data) agreed within 2%/2 mm of measurements. Conclusion: VRT/AEIT related to DBS significantly improves the efficiency of BEAMnrc PHSP simulations. VMC++ can be used to perform simulations of the entire linac and phantom within minutes on a single processor. Further investigation of bremsstrahlung cross‐section data is warranted. Acknowledgement: NIH‐R01CA106770.

Original languageEnglish (US)
Pages (from-to)2874-2875
Number of pages2
JournalMedical Physics
Volume35
Issue number6
DOIs
StatePublished - 2008

Fingerprint

Photons
Electrons
Space Simulation
Data Accuracy

ASJC Scopus subject areas

  • Biophysics
  • Radiology Nuclear Medicine and imaging

Cite this

MO‐E‐AUD B‐04 : Fast, Accurate Photon Beam Accelerator Modeling Using BEAMnrc and VMC++: A Systematic Investigation of Variance Reduction and Efficiency Enhancing Methods and Cross‐Section Data. / Fragoso, M.; Kawrakow, I.; Faddegon, B.; Solberg, T.; Chetty, I.

In: Medical Physics, Vol. 35, No. 6, 2008, p. 2874-2875.

Research output: Contribution to journalArticle

@article{f0bde78be2754cffb444b336025c1693,
title = "MO‐E‐AUD B‐04: Fast, Accurate Photon Beam Accelerator Modeling Using BEAMnrc and VMC++: A Systematic Investigation of Variance Reduction and Efficiency Enhancing Methods and Cross‐Section Data",
abstract = "Purpose: To report on the accuracy of cross‐section data in BEAMnrc and on the performance of variance reduction and efficiency enhancing techniques for fast, accurate linac simulations using the BEAMnrc and VMC++ code systems. Method and Materials: BEAMnrc and VMC++ were used to simulate a 6 MV photon beam from a Siemens Primus linac. Phase space (PHSP) files were generated for a range of field sizes, from 10×10 to 40×40 cm2. BEAMnrc parameters under investigation were grouped by: i) photon and bremsstrahlung cross‐sections; ii) approximate efficiency improving techniques (AEIT); iii) variance reduction techniques (VRT); iv) VRT (bremsstrahlung splitting) with AEIT (range rejection). Efficiencies were obtained for the mean energy, fluence, angular and spectral distributions and PHSP files were subsequently used as input for DOSXYZnrc‐based phantom dose calculations; these calculations were verified against measurements. Results: Efficiencies were calculated for the various VRT/AEIT combinations in BEAMnrc, relative to simulations without VRT/AEIT, namely: (a) 935 (∼111 min. on a single 2.6 GHz CPU) and 200 for 10×10 and 40×40 resp. using directional bremsstrahlung splitting (DBS) and no electron splitting, (b) 420 and 175 for 10×10 and 40×40 resp. using DBS and electron splitting combined with augmented range rejection, a technique recently introduced in BEAMnrc. Calculations with VMC++ produced efficiencies of 1400 (∼6 min. on a single CPU) for 10×10 versus BEAMnrc (no VRT/AEIT). Noteworthy differences (±1–3{\%}) were observed with the NIST bremsstrahlung cross‐sections compared with those of Bethe‐Heitler (default). However, MC calculated dose distributions (using all combinations of VRT/AEIT and cross‐section data) agreed within 2{\%}/2 mm of measurements. Conclusion: VRT/AEIT related to DBS significantly improves the efficiency of BEAMnrc PHSP simulations. VMC++ can be used to perform simulations of the entire linac and phantom within minutes on a single processor. Further investigation of bremsstrahlung cross‐section data is warranted. Acknowledgement: NIH‐R01CA106770.",
author = "M. Fragoso and I. Kawrakow and B. Faddegon and T. Solberg and I. Chetty",
year = "2008",
doi = "10.1118/1.2962386",
language = "English (US)",
volume = "35",
pages = "2874--2875",
journal = "Medical Physics",
issn = "0094-2405",
publisher = "AAPM - American Association of Physicists in Medicine",
number = "6",

}

TY - JOUR

T1 - MO‐E‐AUD B‐04

T2 - Fast, Accurate Photon Beam Accelerator Modeling Using BEAMnrc and VMC++: A Systematic Investigation of Variance Reduction and Efficiency Enhancing Methods and Cross‐Section Data

AU - Fragoso, M.

AU - Kawrakow, I.

AU - Faddegon, B.

AU - Solberg, T.

AU - Chetty, I.

PY - 2008

Y1 - 2008

N2 - Purpose: To report on the accuracy of cross‐section data in BEAMnrc and on the performance of variance reduction and efficiency enhancing techniques for fast, accurate linac simulations using the BEAMnrc and VMC++ code systems. Method and Materials: BEAMnrc and VMC++ were used to simulate a 6 MV photon beam from a Siemens Primus linac. Phase space (PHSP) files were generated for a range of field sizes, from 10×10 to 40×40 cm2. BEAMnrc parameters under investigation were grouped by: i) photon and bremsstrahlung cross‐sections; ii) approximate efficiency improving techniques (AEIT); iii) variance reduction techniques (VRT); iv) VRT (bremsstrahlung splitting) with AEIT (range rejection). Efficiencies were obtained for the mean energy, fluence, angular and spectral distributions and PHSP files were subsequently used as input for DOSXYZnrc‐based phantom dose calculations; these calculations were verified against measurements. Results: Efficiencies were calculated for the various VRT/AEIT combinations in BEAMnrc, relative to simulations without VRT/AEIT, namely: (a) 935 (∼111 min. on a single 2.6 GHz CPU) and 200 for 10×10 and 40×40 resp. using directional bremsstrahlung splitting (DBS) and no electron splitting, (b) 420 and 175 for 10×10 and 40×40 resp. using DBS and electron splitting combined with augmented range rejection, a technique recently introduced in BEAMnrc. Calculations with VMC++ produced efficiencies of 1400 (∼6 min. on a single CPU) for 10×10 versus BEAMnrc (no VRT/AEIT). Noteworthy differences (±1–3%) were observed with the NIST bremsstrahlung cross‐sections compared with those of Bethe‐Heitler (default). However, MC calculated dose distributions (using all combinations of VRT/AEIT and cross‐section data) agreed within 2%/2 mm of measurements. Conclusion: VRT/AEIT related to DBS significantly improves the efficiency of BEAMnrc PHSP simulations. VMC++ can be used to perform simulations of the entire linac and phantom within minutes on a single processor. Further investigation of bremsstrahlung cross‐section data is warranted. Acknowledgement: NIH‐R01CA106770.

AB - Purpose: To report on the accuracy of cross‐section data in BEAMnrc and on the performance of variance reduction and efficiency enhancing techniques for fast, accurate linac simulations using the BEAMnrc and VMC++ code systems. Method and Materials: BEAMnrc and VMC++ were used to simulate a 6 MV photon beam from a Siemens Primus linac. Phase space (PHSP) files were generated for a range of field sizes, from 10×10 to 40×40 cm2. BEAMnrc parameters under investigation were grouped by: i) photon and bremsstrahlung cross‐sections; ii) approximate efficiency improving techniques (AEIT); iii) variance reduction techniques (VRT); iv) VRT (bremsstrahlung splitting) with AEIT (range rejection). Efficiencies were obtained for the mean energy, fluence, angular and spectral distributions and PHSP files were subsequently used as input for DOSXYZnrc‐based phantom dose calculations; these calculations were verified against measurements. Results: Efficiencies were calculated for the various VRT/AEIT combinations in BEAMnrc, relative to simulations without VRT/AEIT, namely: (a) 935 (∼111 min. on a single 2.6 GHz CPU) and 200 for 10×10 and 40×40 resp. using directional bremsstrahlung splitting (DBS) and no electron splitting, (b) 420 and 175 for 10×10 and 40×40 resp. using DBS and electron splitting combined with augmented range rejection, a technique recently introduced in BEAMnrc. Calculations with VMC++ produced efficiencies of 1400 (∼6 min. on a single CPU) for 10×10 versus BEAMnrc (no VRT/AEIT). Noteworthy differences (±1–3%) were observed with the NIST bremsstrahlung cross‐sections compared with those of Bethe‐Heitler (default). However, MC calculated dose distributions (using all combinations of VRT/AEIT and cross‐section data) agreed within 2%/2 mm of measurements. Conclusion: VRT/AEIT related to DBS significantly improves the efficiency of BEAMnrc PHSP simulations. VMC++ can be used to perform simulations of the entire linac and phantom within minutes on a single processor. Further investigation of bremsstrahlung cross‐section data is warranted. Acknowledgement: NIH‐R01CA106770.

UR - http://www.scopus.com/inward/record.url?scp=85024826571&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85024826571&partnerID=8YFLogxK

U2 - 10.1118/1.2962386

DO - 10.1118/1.2962386

M3 - Article

AN - SCOPUS:85024826571

VL - 35

SP - 2874

EP - 2875

JO - Medical Physics

JF - Medical Physics

SN - 0094-2405

IS - 6

ER -