rangeShifter.sourceRashiDistance

[Yu-Sim]

How is the rangeShifter.sourceRashiDistance defined: to the proximal RaShi edge (closer to the source) or to the distal edge?

[JenHardt]

Hello, so the rangeShifter.sourceRashiDistance is set in the function matRad/matRad/steering/matRad_STFGeneratorParticleIMPT line 181, and it should be defined to the proximal edge so closer to the source

[Yu-Sim]

Hi Jenn, thanks. As I'm still a bit confused: If "rangeShifter.sourceRashiDistance" is defined to the proximal edge, is this consistent with the following code (+ rsWidth) from "matRad/doseCalc/+DoseEngines/matRad_ParticleMCsquareEngine.m"?

[wahln]

I think this should be consistent, because in MCsquare we need the distance from isocenter to proximal edge (while matRad always thinks about distance from source).

I might be wrong about how MCsquare sets up the range-shifter, though, and if we misinterpret their definition this would be a bug.

[Yu-Sim]

Hi Niklas, I didn't mean to focus on MC, what I mean is that the calculation in Row 713 makes only sense (correct me if I'm wrong) if "rangeShifter.sourceRashiDistance" is defined to the distal! RaShi edge and this is the opposite of the answer I got from Jenn. So again, my question was weather the "sourceRashiDistance" in matRad is defined to the distal (I guess so) or proximal edge?

[wahln]

I think there is a bug in the line converting the distance. It should read:
isoToRaShi = stf(i).SAD - rangeShifter.sourceRashiDistance - rsWidth;
@JenHardt or @remocristoforetti or @SimonaFa: given my limited time at the moment can you confirm this?
Mabye also validate by comparing an MCsquare and a TOPAS setup.

[JenHardt]

That depends does on which reference point MCsquare uses to define its Geometry, TOPAS for example uses the center of the component. Which also then unfortunately mean that i think in the TOPAS setup, the distance to the range shifter is defined relative to the centre of the range shifter so neither to the proximal or distal edge. But we will have a look at this more closely

[remocristoforetti]

Hello,
thank you for rising this issue as it gave us the chance to review some related features as well.
The sourceRashiDistance is a coordinate point that refers to the position of the Range Shifter (RS) in the matRad coordinate system (the origin is at the virtual source position, the isocenter is at stf(i).SAD from the origin).
This parameter is used by the Pencil Beam analytical algorithm to estimate the additional beam widening introduced by the RS according to (Szymanowski et al., 2002) through matRad_calcSigmaRashi. If I'm not wrong, the way this is implemented assumes that the sourceRashiDistance corresponds to the beginning of the RS slab, i.e the edge proximal to the nozzle to propagate the beam width to the patient surface.
If what Niklas says is correct, and MCsquare needs the distance between isocenter to the proximal (closer to isocenter) edge for the component then we need isoToRaShi = stf(i).SAD - rangeShifter.sourceRashiDistance - rsWidth as the position to feed to MCsquare.
So far I was not able to find an explanation of how MCsquare intends the value to be used.
If my interpretation of the geometry is correct, then we need to update the Topas interface as well, as Topas positions the RS with respect to its center and the nozzle and we thus need to increase this distance by 0.5*rsWidth.
As of the current status of the MCsquare engine, the interface will not be able to run through the parameter file generation in presence of a range shifter because of minor issues in handling of indexes. I fixed this as well and will soon update the dev branch to include this bug fixes. Sorry for the inconvenience and thanks for spotting this inconsistency.

Hope this helps,
Best :)
Remo