[Gate-users] Using IAEA phase spaces

[germano.russo at i-seecomputing.com]

Dear all,

I recently tried to use as source an IAEA phase space, specifically the
one named "VarianClinaciX_6MV_20x20_aboveMLC_w1.IAEAphsp" (downloadable
from https://www-nds.iaea.org/phsp/photon1/), via the "/gate/source/[Source
name]/addPhaseSpaceFile" macro command.

The simulation (Gate 7.0) run smoothly, however I got meaningless results.
Digging into the related Gate source code, I believe I found a few bugs
which I report to you asking for confirmation.

   1) In the function 'GateSourcePhaseSpace::OpenIAEAFile()' the call
'pIAEARecordType->initialize()' is placed after
'pIAEAReader->getRecordContents(pIAEARecordType)'. These two lines shall be
swapped, otherwise the record contents information extracted form the IAEA
header and stored in the the object pointed by pIAEARecordType is cleared
by the initialisation operation.

   2) The "GateIAEA*" source files appear to be based on an old C++
implementation for handling phase spaces distributed by the IAEA. Some
things seem to be outdated, for instance in the "GateIAEAConfig.h" the
definition 'typedef long IAEA_I32' must be replaced with 'typedef int
IAEA_I32' in order to conform to the current "*.IAEAphsp" file format.

   3) In the function 'GateSourcePhaseSpace::GenerateIAEAVertex()' the
'pIAEARecordType->w' variable seems to be misused in the expression 'dz *=
pIAEARecordType->w'. The 'pIAEARecordType->w' variable contains the Z
direction cosine of the particle, not just its sign. So it should rather be
'dz = pIAEARecordType->w'. Since the 'pIAEARecordType->w' variable is
slightly smaller than 1, this mistake causes small errors in the
normalisation of the momentum (and hence of the kinetic energy).

After having applied these correction, I was able to obtain a reasonable
output dose distribution.
Could you please check the above and give me a feedback? Thanks!

Kind regards,

Germano Russo
R&D I-See (Internet - Simulation Evaluation Envision) company