Difference: FREDOptimizerStudies (4 vs. 5)

Revision 52020-10-12 - AlessioSarti

 			 META TOPICPARENT 
			 name="SBAMSoftwareProjects" 
		
 FRED Optimizer tips & helps 


 
  FRED Optimizer tips & helps
   
   Installing and running 
  The input file
   The new workflow
 
   Checking the output 
  Creating the DVHs and analysing the output 
  Other / older tools
   Browsing the results using itksnap
 
 
   Prostate studies 
  The first case with seven fields
   Comments
 
 

  
 Installing and running 

The Optimizer lives under git.
- META TOPICPARENT
+ name="SBAMSoftwareProjects"
-<
<
+Download using
->
>
+Download using
-<
<
+<--?xml version="1.0" encoding="UTF-8"?-->
 git clone http://XXXuser@arpg-serv.ing2.uniroma1.it:8181/git/Optimizer.git 

where you need to replace XXXuser with the account under the git server (if you don't know if you have an account, ask a.sarti) and compile it as it is explained inside README.txt

To run the optimizer one has to launch it using the syntax
-<
<
 ./optiRoot #inputfile
-<
<
+A template for the input file that uses 7 different fields can be found in the folder inpFiles.
->
>
+A template for the input file that uses 7 different fields can be found in the folder inpFiles.
  The input file 

The input files contains several fields and lines to setup the optimizer: 

 The dij for each field can be specified using dij:
  ptv: defines what is the PTV and the goal dose, the weight
  oar: for the organs at risk one can set the max dose, the weight and the DMF.
  NoT: for the normal tissues (CT - PTV - OARs) one can set the max dose, the weight, the DMF and the threshold applied to consider the voxels as NoT
  outFlag: flags the output files
 
As an example:
-<
<
 dij: SingleFieldMatrix/70MeV/Dij_merge1.bin
dij: SingleFieldMatrix/70MeV/Dij_merge2.bin
dij: SingleFieldMatrix/70MeV/Dij_merge3.bin
dij: SingleFieldMatrix/70MeV/Dij_merge4.bin
dij: SingleFieldMatrix/70MeV/Dij_merge5.bin
dij: SingleFieldMatrix/70MeV/Dij_merge6.bin
dij: SingleFieldMatrix/70MeV/Dij_merge7.bin
ptv: PTVTrue_NEW_itk.mhd Dgoal 2
oar: vescica_NEW_itk.mhd maxDose 2.2 DMF 1. weight 20 
oar: retto_NEW_itk.mhd maxDose 2.2 DMF 1. weight 20 
oar: bulbopenieno_NEW_itk.mhd maxDose 2.2 DMF 1. weight 20 
oar: femoredx_NEW_itk.mhd maxDose 1. DMF 1. weight 20 
oar: femoresin_NEW_itk.mhd maxDose 1. DMF 1. weight 20 
oar: ano_NEW_itk.mhd maxDose 2.2 DMF 1. weight 20 
NoT: maxDose 1.2 DMF 1. weight 10 thre 5.
-<
<
+outFlag: 70MeV_VeReBp22Gy_NT12_DMF1_OAR20
->
>
+outFlag: 70MeV_VeReBp22Gy_NT12_DMF1_OAR20
-<
<
+The use of the weights herebefore sets NoT @ 10%, OARs @ 20% and PTV @ 100%.
 The new workflow
-<
<
+- legge il DMF_NT (default = 1)
- carica i PTV con relativa plannedDose ed eventuale DMF (default = 1)
- carica gli OAR con relativa maxDose ed eventuale DMF (default = 1)
- crea la mappa “dense” del DMF (per ogni voxel della CT)
- carica le varie Dij (ad esempio una per ogni field) applicando il DMF
->
>
-<
<
+- output: scrive la mappa con i voxel appartenenti alle varie ROI => rois.mhd
- output: scrive la mappa densa con il DMF delle varie ROI => alldmf.mhd
->
>
+The is also the possibility to choose the initial particle numbers (the initial 'guess') using the directive fluences.
->
>
+Typical input=
...
fluences: choice
...
With the following options for choice
1) "ones" [default] : set all PBs to unitary fluence
2) "random" : fill PBs fluences with random values in [1:10] range
3) path : load initial fluences from given file (format same as writePBs() function)
-<
<
+- lancia l’ottimizzatore
->
>
+The use of the weights herebefore sets NoT @ 10%, OARs @ 20% and PTV @ 100%.
->
>
+ The new workflow 
 
 PTVs are loaded, setting plannedDose, DMF (default = 1) and weight (default 100%)
  OARS are loaded, setting maxDose, DMF, weights. If a voxel is flagged as PTV, it is not overwritten.
  loads the NT as ALL - PTV - OARS applying a threshold that can be tuned in the NoT field.
  DMF is populated and applied when computing the dose/loading the Dij
 
After the configuration: 
 As a first thing the optimizer writes the voxel maps in the different ROIs: rois_outflag.mhd.
  Then the dens map of the DMF is written as well: alldmf_outflag.mhd
  The optimizer is launched
  Reconstruct and writes the optimized dose map: optiDose_outflag.mhd
 
To investigate the output, please look at the following section.
-<
<
+- output: ricostruisce e scrive la mappa di dose ottimizzata (usando il DMF) => optiDose.mhd
  Checking the output 
 Creating the DVHs and analysing the output 

In order to generate the DVH a script is provided by git.

To run it:
-<
<
 ./testDVH/mhd_DVH.py -p -v -roi PTVTrue_NEW_itk.mhd vescica_NEW_itk.mhd retto_NEW_itk.mhd bulbopenieno_NEW_itk.mhd 
ano_NEW_itk.mhd femoredx_NEW_itk.mhd femoresin_NEW_itk.mhd -dose optiDose_70MeV_0.8Op_PTV_2GyVesRet_1Gyall.mhd 
--norm-volume -Dgoal 200

The available options can be prompted using
-<
<
 ./testDVH/mhd_DVH.py -h

The output is.
->
>
+To check the dose using itksnap, you can on lisa use the installed binaries.
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/nvidia/nsight-systems/2019.3.7/Host-x86_64/ 
export DYLD_LIBRARY_PATH=$DYLD_LIBRARY_PATH:/opt/nvidia/nsight-systems/2019.3.7/Host-x86_64/
 

To run the software:
/home_arpg/software/itksnap-3.8.0-20190612-Linux-gcc64-qt4/bin/itksnap  

First of all you can add the dicom, then the optiDose mhd to make a superposition.
 --   Alessio Sarti  - 2020-10-06 

 Comments 


 
  (Edit - Preview)
 
<--/commentPlugin-->
->
>

View topic | History: r7 < r6 < r5 < r4 | More topic actions...