[Gate-users] Dose Calculation interface problem?

[Kirchhof, Simon]

Dear Theodora, Dear GATE community,


I stumbled across your email today and I wanted to add some thoughts about GATE's dose actor:


GATE's dose actor does not seem to communicate with the Geant4 navigator, i.e. it does not limit the geometrical step size. Hence, for photons in homogeneous volumes the step size of the simulation can be (much) bigger than the resolution of the dose actor. Depending on the position used for scoring, i.e. "pre-step-position", "post-step-position" or something in between, the assignment of an energy deposition to a voxel can be off, for the mentioned case.


Do the irregularities you observe occur for all HitTypes, i.e. "pre, post, middle, random"?


What you might observe is: You add a volume, you add a boundary, the stepping changes due to the (virtual) boundary. The stepping positions are distributed differently and that might change the result of GATE's dose actor.


Any thoughts from the Dose Actor developers? Do you agree/disagree?


Best regards,

Simon


________________________________
Von: Gate-users <gate-users-bounces at lists.opengatecollaboration.org> im Auftrag von Theodora Kostou <theokost at upatras.gr>
Gesendet: Freitag, 11. August 2017 12:04
An: gate-users at lists.opengatecollaboration.org
Betreff: [Gate-users] Dose Calculation interface problem?


Dear GATE users,

I am trying to use GATE to calculate the effect of a nanoparticle on the depth-dose curve in radiation therapy. As I was getting incorrect results, I scaled up the problem in order to eliminate possible nano-effects. However, even at the mm level, the problem still persists.

It seems to me that there is a problem with geometrical interfaces. Any help would be appreciated.

Here is the setup:

The phantom is a box of water, 20x20x20 mm3 in size. It is irradiated by a 2x2 mm2 square source emitting monoenergetic 140.5 keV photons, placed 2 mm outside the phantom (The small sizes of the phantom and the source are necessary in order to achieve good statistics when a nanoparticle is inserted in the phantom). Dose is scored in a 200x200x200 matrix. Each voxel is 0.1x0.1x0.1 mm3.

Then, I add a 2mm thick layer of water, placed 5 mm below the surface of the phantom, i.e inside the phantom. The depth-dose curve now exhibits two sharp peaks at the location of the interfaces of the water layer and the water phantom (at 4 and 6 mm).

The simulations were run on the VIP platform using GATE v6.2. To narrow down the causes I turned off all physical processes, except the Compton effect.

Since physically nothing has changed - we are still dealing with a cube of water - I suspect that the cause of the problem is the mere introduction of a new geometrical volume.

Has anyone ever encountered this problem? What are the solutions?

Thank you,
Theodora

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opengatecollaboration.org/pipermail/gate-users/attachments/20170926/bea16b0d/attachment.html>