Article
Conserving absorbed energy during accumulation of dose from a 4D geometry
Search Medline for
Authors
Published: | September 24, 2009 |
---|
Outline
Text
Purpose: The aim of this work is to account for changes in patient geometry by accumulating all dose to a single fixed geometry, I f , while preserving deposited energy in all moving geometries.
Materials: We consider a single instance I m, from the set of moving geometries where I m is related to I f by deformable registration. It is assumed that the mass is preserved between I m and I f . The energy absorbed by I m should then also be preserved which implies that the dose should not be the transferred quantity since a voxel in I f may consist of several I m voxels with different composition. Dose varies smoother than absorbed energy and may be a preferred subject for an interpolation scheme. The accumulated dose is by choice selected to be presented in I f , therefore the mass in I f should be preserved.
To fulfil these requirements for the dose in the fixed geometry, D
f
, an energy conservation interpolation method (ECI) is proposed according to
D
f
(r
f
) = D
m
(r
m
) ρ
m
(r
m
) / ρ
f
(r
f
), (1)
where D
m
(r
m
) and ρ
m
(r
m
) are interpolated quantities in I
m
and the position r
f
is related to r
m
by the registration. The energy in I
m
is approximately preserved and the interpolation is done in the smoother dose domain.
A post energy transfer method (PETM) is also proposed where each voxel in I m is divided into eight subvoxels and the energy absorbed by each subvoxel is transferred to I f using the registration. D f (r f ) is calculated as the sum of transferred energy divided by the mass in I f . The aim of PETM is to more accurately conserve the absorbed energy in I m into I f .
The ECI and PETM methods are applied to a lung case treated with a 6x6 cm2 proton field with a SOBP of 4 cm wed. I m has an artificial displacement of the target volume of ~9 mm compared to I f . The dose is calculated with Geant4 (Agostinelli, et al. Nucl Instrum Meth. 2003;A 506) in a 2x2x3 mm grid.
The transfer of dose from I m to I f is done using ECI and PETM and is compared to an energy transfer method (Siebers, et al. Med Phys. 2008;35) (ETM), and direct dose mapping (Rosu, et al. Med Phys. 2005;32) (DM) in Im.
The ETM can be considered to be accurate but is only applicable to Monte Carlo dose calculation and requires the less commonly existing deformation field from I m to I f . DM is a simple interpolation scheme but is sensitive to redistribution of mass between voxels.
Results: Figure 1 [Fig. 1] shows the CTV DVH in I f for the calculation of the same plan in I m where the dose is transferred to I f using ECI, PETM, DM and ETM. For the data set considered in this work, the ECI and PETM method shows a qualitatively better agreement with the ETM than the DM.
Conclusions: We present two new schemes for transferring a dose distribution from a moving to a fixed geometry. Both methods better preserve the energy absorbed by the moving geometry.