971 / 1096
S955
ESTRO 36
_______________________________________________________________________________________________
EP-1736 Radiation and lasers isocenters coincidence
with ArcCheck phantom
F. Tato de las Cuevas
1
, J. Yuste Lopez
1
1
Hosp. Univ. de Canarias, Medical Physics Dept., Santa
Cruz de Tenerife, Spain
Purpose or Objective
One tool of Machine QA module of ArcCheck phantom (
AC
)
software checks Radiation and Lasers Isocenters
Coincidence (
RLIC
). The purpose of this work is to
evaluate the precision and accuracy of this software tool,
comparing it to the same test made with EPID (Electronic
Portal Imaging Device).
Material and Methods
The LINAC is an Elekta Synergy with Agility MLC and 6 MV
energy.
The RLIC with ArcCheck phantom (
AC
) are obtained
following the instructions of the software manual. The
measurements are done in continuous gantry movement
and for discrete gantry angles. Measurements are made at
9 º collimator angle for a 1x25 cm field. A series of
measurements were made also in 99º to see the MLC
effect, as Agility head has not backup jaws. The AC
displacements from laser isocenter in two directions are
made in order to check software sensitivity.
RLIC are made with EPID, positioning a Bearing Ball (
BB
)
in the lasers isocenter of a 5x5 cm field and acquiring
Images from 0º to 360º gantry angles in 45 º steps. The
radiation center of the squared field and the center of the
BB are calculated with a MATLAB in-house software. BB
center is calculated with sub-pixel accuracy in each
direction, 3 profiles are obtained and fitted to Gaussian
curves, and the mean maximum of the 3 curves is
calculated. Radiation field center is obtained calculating
the 50% pixel value of a vertical and horizontal profile.
The difference between BB center and radiation field
center are computed for each gantry angle for in-plane
and cross-plane directions. The RLIC for EPID
measurements are computed using these values.
Results
The RLIC results obtained with AC for each gantry are
compared with EPID in the first figure. The mean distance
over all gantry angles, for AC (for 9 and 99 º collimator
degrees) and EPID are: 0.3, 0.6 and, 0.7 mm, respectively.
The AC results are just distance (because this phantom is
not capable of give deviation in in-plane direction for each
gantry angle). The results for AC for 9º are higher than for
99º because of the irregular MLC radiation field limit
exposed for 9º to the AC diodes. The RLIC for EPID are
given in in-plane and cross-plane directions, the distance
for each gantry angle is calculated from both directions
and show a bigger mean value than for AC, because of
being calculated in just one direction in this phantom.
Table 1 shows RLIC results for AC and EPID. AC data is
given by phantom software. It can be noticed that the
coincidence for both isocenters is lower for the EPID, this
can be explained because AC just take into account one
direction in each gantry angle
.
The sensitivity of AC for RLIC is fairly good taking into
account the uncertainties of measurement 1 mm between
laser positions.
Conclusion
ArcCheck software is capable of give a fairly accurate
measurement of the laser and radiation isocenters
coincidence, taking into account to add about 0.5 mm
displacement in X and Z directions.
EP-1737 Efficient troubleshooting of accelerator faults
using the TrueBeam Log Viewer software application
L.H. Praestegaard
1
1
Aarhus University Hospital, Department of Oncology,
Aarhus, Denmark
Purpose or Objective
In case of an accelerator fault, the identification of the
root cause often takes much longer time than the
correction of the fault itself (for example replacement of
a component or a calibration). Accordingly, the uptime of
an accelerator very much depends on an efficient
troubleshooting process. In addition, an overview of
existing faults is essential for an efficient planning of
service tasks.
Material and Methods
The TrueBeam Log Viewer application is an in-house
software application developed in C#. For the Varian
TrueBeam system all faults occurring during beam on
produces an event. With the application all events, for
user-selected accelerators and data interval, can be listed
including fault details, treatment plan details, mechanical
axes, imaging parameters, and imaging arm positions at
the time of the event (see figure 1). For each event the
TrueBeam system produces a number of node records,
each with detailed information about system parameters
versus time just before the fault occurred (for example all
MLC positions and motor currents at 500 time steps of 10
ms). All node records are readily available in the
application from the list of events. In addition, the
application can generate an event alarm, including the
corresponding event data, each time an event occurs for
user-selected accelerators.