Building clinical Monte Carlo code from

Geant4, PENELOPE or EGSnrc

Joao Seco Joao Seco 11, Christina Jarlskog , Christina Jarlskog 11, Hongyu Jiang , Hongyu Jiang 22 and Harald Paganetti and Harald Paganetti 11

1 Francis H. Burr Proton Therapy Center Massachusetts General Hospital,Harvard

Medical School, 30 Fruit Street, Boston, Massachusetts 02114 USA

2 University of Arkansas for Medical Sciences, 4301 W.Markham Street, Little Rock, Arkansas 72202 USA

GEANT4, PENELOPE and EGSnrc Monte Carlo Codes:

• Do you need to use Monte Carlo codes to solve your problem…?

Can you use a faster PC or “clever” math's, physics or numerical tricks to solve your problem;

• How accurate is your Monte Carlo physics …? Possible issues are:

1. Multiple scattering

2. Bremsstrahlung

3. Photo-electric/Compton/Pair-Production/Rayleigh/Stopping Powers etc…

• How fast is you Monte Carlo code….?

Can it run on 1 PC or does it need a large cluster … keeping in mind that clinical plans must not take more than 1 hour to generate;

• How versatile is your MC code… can it do photons, electrons, protons, carbon ions, etc… ?

Verhaegen and Seuntjens 2003, PMB, 48 R107-R164

Review of Physics within each code

Inter-comparison of electron Monte Carlo dose calculations for EGSnrc, GEANT and

PENELOPE (2004)

Joao Seco, Alex Howard and Frank Verhaegen

AND

Accuracy of the photon and electron physics in GEANT4 for radiotherapy applications (2005)

Emily Poon and Frank Verhaegen

Our results with

EGSnrc, STD_EM, LowE and PENELOPE …..

1 MeV H2O - Electrons EGSnrc vs GEANT

0.00E+00

1.00E-18

2.00E-18

3.00E-18

4.00E-18

5.00E-18

6.00E-18

7.00E-18

0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 0.400 0.450 0.500

Depth (cm)

Dose

/Incid

ent P

artic

le/Inc

i. Are

a (G

y/cm^

2)

EGSnrc

STD_EM

Penelope

LowE

1 MeV BONE - Electrons EGSnrc vs GEANT

0.00E+00

1.00E-18

2.00E-18

3.00E-18

4.00E-18

5.00E-18

6.00E-18

7.00E-18

0.000 0.050 0.100 0.150 0.200 0.250

Depth (cm)

Dose

/Incid

ent P

articl

e/Inc

i. Area

(G

y)/cm

^2

EGSnrc

STD_EM

Penelope

LowE

1 MeV LUNG - ELECTRON EGSnrc vs GEANT

0.00E+00

1.00E-18

2.00E-18

3.00E-18

4.00E-18

5.00E-18

6.00E-18

7.00E-18

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

Depth (cm)

dose

/incid

ent p

artic

le/In

ci.Ar

ea (G

y/cm

^2)

EGSnrc

STD_EM

Penelope

LowE

SLAB_THICKNESS =CSDA

100*

CSDA - continuos slowing down approximation in g/cm2

– material density in g/cm3

1 MeV incident electrons on homogeneous material

100 keV incident electrons on homogeneous material

10 MeV incident electrons on homogeneous material

POSSIBLE PROBLEM IN THE ELASTIC MULTIPLE SCATTERING ALGORITHM….

Consistency test of the electron transport algorithm in

the GEANT4 Monte Carlo code

Emily Poon, Jan Seuntjens and Frank Verhaegen

Medical Physics Unit, McGill University, 1650 Cedar Avenue, Montreal, Quebec, H3G 1A4, Canada

Abstract

In this work, the condensed history algorithm in GEANT4 (version 4.6.2.p01) is examined. Simulations of an ionization chamber composed of water for 1.25 MeV incident photon beams under Fano conditions, and evaluated the consistency of the cavity response for several combinations of electron transport parameters.

GEANT4 permits electrons to reach geometric boundaries in large steps, and underestimates lateral displacement near interfaces. Step size artifacts due to distortions in electron fluence and angular distributions reduce the cavity dose by up to 39%. Accurate cavity response can be achieved using severe user-imposed step size restrictions. [They] suggest that improvements in the electron transport algorithm in GEANT4 should address the handling of boundary crossing.