RE: [fluka-discuss]: Possible Flair error and stopping power calibration/corrfact/voxel geometry

From: Vasilis Vlachoudis <Vasilis.Vlachoudis_at_cern.ch>
Date: Wed, 2 Nov 2016 08:24:30 +0000

Thank you Lars,

the mismatch on the CORRFACTOR was due to a typo in the last release of flair.
I've corrected it on the svn repository, and I am currently working on introducing
an additional column in the table to have HUmin and HUmax, to display more
clearly the range.
Once I have this functionality operational I will push an update.

Vasilis

________________________________
From: Lars Fredrik Fjæra [lfj049_at_student.uib.no]
Sent: Monday, October 17, 2016 17:22
To: Franz Siegfried Englbrecht; Vasilis Vlachoudis
Cc: fluka-discuss_at_fluka.org
Subject: Re: [fluka-discuss]: Possible Flair error and stopping power calibration/corrfact/voxel geometry

Dear all,

Thank you all for your answers.
You have cleared up a lot for me.
Franz, being able to read the .vxl files have been something I've been wondering about, so that is really helpful!

I also found the CORRFACT cards in the output right after I submitted my mail. I should really get better at looking into the output files. I also there noticed that the correction factors to the stopping power were applied having, however, negative values. So providing the CdEdx_rel_min and CdEdx_rel_max using negative values makes perfect sense.

I was actually rather referring to the unit specification on the far left in the voxel tab and not the materials. If you look at the red boxes outlined in the attached image, there was previously in Flair (image from the manual) unit specified using < with a line beneath (<=), and now in the current versions of Flair, this line has been removed (<). I presume that this line should still be in Flair considering the HUs are given in less or equal. Or am I misunderstanding?

Now that we have established that the HU-range is taken as <=, it brings me to a follow-up question, if you don't mind.
If I extend my previous example:

< Unit Material Crho_min Crho_max
900 HU<900 0.9803 1.0192
1000 HU<1000 0.9810 1.0168
1100 HU<1100 0.9853 1.0144

The HU<1000 material is defined as having a density of 1.5873 g/cm3. So this density will correspond to HU=950, and should be scaled according to the correction factors (0.9810 and 1.0168) for material HU<1000.
And the since my HU=1000 is given the material HU<1000, I would then believe that the correction factor to the density should be 1.0168, i.e. an up-scaling from the density of 1.5873 g/cm3 to arrive at 1.6140 g/cm3.
However, when I look at the CORRFACT in the output file the following is given:

***** Next control card ***** CORRFACT -1.000 0.9853 0.000 6.000 0.000 0.000

and

6 VOXEL002 60 HU<1000 OFF 0.00000E+00 9.99852E+04
                        ( 60 HU<1000 OFF )

So the region number 6 (6.000 in the CORRFACT) is my HU=1000 (material HU<1000) region. And as you can see, the correction factor applied to this region is 0.9853, equal to the lower correction factor of material HU<1100. Material HU<1100 have a density of 1.6387 g/cm3 so I will arrive at a final density of 1.6146 g/cm3 (which is basically the same as the previously mentioned up-scaling to 1.6146 g/cm3). Even though the final value of the density will practically be the same for both cases, is it not therefore an incon
sistency, considering that the materials are given by Hounsfield units as <=, while apparently the correction factors are given by Hounsfield units as < ?

Thank you very much again!



---
Lars Fredrik Fjæra
On Mon, Oct 17, 2016 at 11:34 AM, Franz Siegfried Englbrecht <Franz.Englbrecht_at_physik.uni-muenchen.de<mailto:Franz.Englbrecht_at_physik.uni-muenchen.de>> wrote:
Dear Lars, dear FLUKA Users,
I want to adress only the question of 'So where are these CORRFACT cards?'
You can, on a linux machine, adress the ASCII content of the binary .vlx file by typing 'strings FILE.vlx'. There you will find the automatically generated cards.
Best,
Franz
Quoting Lars Fredrik Fjæra <lfj049_at_student.uib.no<mailto:lfj049_at_student.uib.no>>:
Dear FLUKA users,
I have two problems I want to address.
First, is there an error in the Flair Dicom/Voxel tab? The unit range on
the far left specifies Hounsfield units "less than" (<), however, when I
was using artificial CT-scans, I noticed that Flair applies materials
according to units of "less or equal" (<=). As an example, I have the
following defined:
< Unit          Material          Crho_min          Crho_max
1000           HU<1000        0.9810<tel:0.9810>               1.0168
1100           HU<1100         0.9852<tel:0.9852>               1.0144
In my CT scan I have only two Hounsfield units, HU=1000 and HU=-1000. Using
the HU=1000 as an example, it should, according to Flair, be assigned the
material called HU<1100 considering that 999<HU1000<1100. However, the HU
of 1000 gets assigned the material HU<1000. This then means that the unit
specification in Flair should be (=<). I have also noticed that this was
the case in previous versions in Flair referring to for example the image
in the manual. I have also attached images of my voxel/geometry setup to
further illustrate the problem. The same also happens for my HU=-1000. It
gets assigned the material HU<-1000 specified with the unit <-1000, where
it really should be assigned the material HU<-1010.
Next problem. I really miss a proper explanation/documentation of the
stopping power calibration in the voxel tab. Performing the density
calibration (Crho_min and Crho_max) is straight forward and works
perfectly. I am however struggling with dE/dx calibration. First off, it is
specified that the dE/dx correction are relative corrections. Relative to
what? According to the manual, "dE/dx is applied relative to the correction
of density (negative WHAT(1) in CORRFACT)". However, I was under the
impression that if you perform the calibration in the .vxl file itself, you
don't need any separate CORRFACT cards? The manual also states:
"CdEdx_rel_min:  relative correction factor on dE/dx for the minimum unit
in the range         CdEdx_rel_max:  relative correction factor on dE/dx
for the maximum unit in the range". So then I wonder, is the correction
applied to the density corrections? I.e. If I have rho = 1, and Crho_max =
1.5 and CdEdx_rel_max = 1.5, the effective density for the stopping power
will be 2.25?
Further, from the FLUKA course slides about medical application, there is a
sentence that states "FLAIR automatically appends the CORRFACT cards
calculated taking into account the calibration curves provided by the user
at the end of the .vxl file." So where are these CORRFACT cards?
And to add to all this, I have been playing around with the  CdEdx_rel_min
and  CdEdx_rel_max, just trying to change them to see if I can obtain some
difference in the simulated ranges for Bragg peaks. However, I can apply
whatever value to CdEdx_rel_min and  CdEdx_rel_max, and still obtain
exactly the same curves. I've even tried to obtain an error in my
simulations by setting the CdEdx_rel_min and  CdEdx_rel_max beyond the
range [2/3,3/2], however the simulation runs fine. I did the same thing for
the Crho_min and max, where I successfully obtained an error. So it seems
that FLUKA doesn't even take into account my input to CdEdx_rel_min and
 CdEdx_rel_max. On the other hand, I am perfectly able to apply corrections
using regular CORRFACT cards. Can someone please explain thoroughly how to
properly use CdEdx_rel_min and  CdEdx_rel_max?
Thank you!
---
*Lars Fredrik Fjæra*
__________________________________________________________________________
You can manage unsubscription from this mailing list at https://www.fluka.org/fluka.php?id=acc_info
Received on Wed Nov 02 2016 - 11:47:30 CET

This archive was generated by hypermail 2.3.0 : Wed Nov 02 2016 - 11:47:40 CET