[Gate-users] TPSPencilBeam source description

[jbaran]

Dear Gate Community,

I would like to continue the thread and ask one more question regarding 
the TPS plan based simulations. I used clitkDicomRTPlan2Gate program to 
convert TPS plan to Gate format and get the 'Segmentation fault (core 
dumped)' but the file has been produced. When I tried to run the 
simulation I get an error as follows:

[Beam-0] source description file specifies energy spread in MeV
[Beam-0] (This overrides whatever you configured for the 
'setSigmaEnergyInMeVFlag' in the configuration of TPSPencilBeam.)
[Beam-0] [TPSPencilBeam] Source description file successfully loaded.
GateSourceTPSPencilBeam.cc (l.238): Something went wrong while parsing 
plan description file "plan_description.txt":
wrong number of values (1) on line 25 of plan_description.txt, expected 
3 value(s) of type d

I checked the plan description file (plan_description.txt) and it looks 
as follows (a few first lines):

#TREATMENT-PLAN-DESCRIPTION
#PlanName

#NumberOfFractions
37
##FractionID
1
##NumberOfFields
10
###FieldsID
1
###FieldsID
2
###FieldsID
3
###FieldsID
4
###FieldsID
5
###FieldsID
6
###FieldsID
7
###FieldsID
8                   <----------- line no. 25
###FieldsID
9
###FieldsID
10
#TotalMetersetWeightOfAllFields
483

I am wondering if anyone of you have similar problem and might have an 
idea how to solve that. I use GATE 8.2.

All the best,
Jakub

On 2019-08-21 15:06, David Boersma wrote:
> Hi Jakub,
> 
> Yes, sorry, this is a relatively new (~1 year old) addition to the
> "source properties" in GateSourceTPSPencilBeam.
> 
> We ("GateRTion" crowd) are even thinking about changing it yet again.
> 😊
> 
> If the treatment plans that you work with are specifying the spot
> weights in "number of ions" (option "setSpotIntensityAsNbIons" true),
> then in principle you do not need to worry about this, the monitor
> calibration will be ignored (add a dummy polynomial, maybe you can
> even omit it).
> 
> If it does use MU as a spot weight unit, then during commsissioning of
> your TPS you or your medical physicist colleagues probably had to
> provide tables/curves of N/MU as function of beam energy. A polynomial
> fit to those data would be good first try for a good N/MU calibration
> in the source properties file. Maybe another possibility: fit them
> together with the energy polynomials, if you happen to have a set of
> IDD curves with absolute calibration.
> 
> In older versions of Gate (8.0 and older) the N/MU calibration was
> hardcoded in GateSourceTPSPencilBeam. I made a polynomial fit to that
> calibration function and inserted it in the source properties file of
> the "dosimetry/dosimetry/protontherapy" example:
> 
> https://github.com/OpenGATE/GateContrib/blob/master/dosimetry/dosimetry/protontherapy/data/Source-Properties.txt
> 
> 
> A few years ago, when I was working in Uppsala, I compared this
> function with the measured calibration curve of the Skandion clinic.
> Amazingly, they were completely compatible within the accuracy of the
> measurement. If the treatment machines for which you are going to
> perform simulations are similar to those at Skandion, then I bet that
> you can be optimistic about the similarity of your curve with that
> "standard calibration curve". 😊
> 
> About the planned change to this functionality: right now, only the
> shape of the N/MU curve matters, not the overall normalization factor.
> The spot weights (be it in N or in MU) are effectively renormalized to
> the number of requested primaries: if you request 10 times more
> primaries, then the DoseActor will record a 10 times higher dose. In
> order to compare with TPS dose, you need to rescale the distributions
> with NTPS/NMC. It has been suggested that the dose actor should apply
> this rescaling factor directly (this is technically not so hard, using
> the particle weight). This probably wouldn't affect the source
> properties file, but it would put stronger requirements on the quality
> of the fit that you provide (not only the shape, but the normalization
> will matter too). For the implementation, we need to take care that
> this will work correctly with variable number of primaries (for the
> SaveData functionality of the dose actor, e.g. with "everyNseconds",
> or in combination with the "stop on script" actor).
> 
> HTH,
> 
> David
> 
> -------------------------
> 
> VON: Gate-users <gate-users-bounces at lists.opengatecollaboration.org>
> im Auftrag von jbaran <jbaran at ifj.edu.pl>
> GESENDET: Mittwoch, 21. August 2019 13:18:44
> AN: Gate Users
> BETREFF: [Gate-users] TPSPencilBeam source description
> 
> Dear Gate Community,
> 
> According to the TPSPencilBeam source description
> (http://wiki.opengatecollaboration.org/index.php/Users_Guide:Source#TPS_Pencil_Beam_source)
> 
> I have to provide 9 polynomials. I dealt with all of them. However I
> have a problem to describe the beam monitor calibration in number of
> particles per monitor unit (N/MU). Unfortunately, there is no
> description of that in that example:
> https://github.com/OpenGATE/GateContrib/blob/master/dosimetry/Radiotherapy/example5/data/Source-Properties.txt
> 
> .
> 
> Is anyone have similar problem and could provide the proper way of
> doing
> that.
> 
> All the best,
> Jakub
> 
> _______________________________________________
> Gate-users mailing list
> Gate-users at lists.opengatecollaboration.org
> http://lists.opengatecollaboration.org/mailman/listinfo/gate-users
> _______________________________________________
> Gate-users mailing list
> Gate-users at lists.opengatecollaboration.org
> http://lists.opengatecollaboration.org/mailman/listinfo/gate-users