Evaluation of the truebeam machine performance check (MPC) geometric checks for daily IGRT geometric accuracy quality assurance

Abstract Machine Performance Check (MPC) is an automated and integrated image‐based tool for verification of beam and geometric performance of the TrueBeam linac. The aims of the study were to evaluate the performance of the MPC geometric tests relevant to OBI/CBCT IGRT geometric accuracy. This included evaluation of the MPC isocenter and couch tests. Evaluation was performed by comparing MPC to QA tests performed routinely in the department over a 4‐month period. The MPC isocenter tests were compared against an in‐house developed Winston–Lutz test and the couch compared against routine mechanical QA type procedures. In all cases the results from the routine QA procedure was presented in a form directly comparable to MPC to allow a like‐to‐like comparison. The sensitivity of MPC was also tested by deliberately miscalibrating the appropriate linac parameter. The MPC isocenter size and MPC kV imager offset were found to agree with Winston–Lutz to within 0.2 mm and 0.22 mm, respectively. The MPC couch tests agreed with routine QA to within 0.12 mm and 0.15°. The MPC isocenter size and kV imager offset parameters were found to be affected by a change in beam focal spot position with the kV imager offset more sensitive. The MPC couch tests were all unaffected by an offset in the couch calibration but the three axes that utilized two point calibrations were sensitive to a miscalibration of the size in the span of the calibration. All MPC tests were unaffected by a deliberate misalignment of the MPC phantom and roll of the order of one degree.

integrated into the linac. The most common form of this imageguided radiotherapy (IGRT) is with kiloVoltage (kV) imaging systems aligned to the linac isocenter. Such systems can be used as either planar x rays [On-Board-Imager (OBI)] or as Cone-Beam Computed Tomography (CBCT). 1 The correct functioning and accuracy of these systems are paramount for the geometrically accurate delivery of the treatment 2-6 and the required accuracy is increasing as treatment margins are reduced to spare healthy tissue.
Daily Quality Assurance (QA) testing of Linear Accelerator (linac) IGRT functionality is recommended in the AAPM Task Group report 179, 7  It is the aim of this study to compare the MPC geometric checks that are relevant to the OBI/CBCT system against standard QA tests to provide an evaluation of MPC as an IGRT geometric QA device.
The study was performed over a longer period (4 months) than the Clivio study and provides an assessment of the MPC stability and sensitivity to drift of the linac systems being tested. Sensitivity is further examined by the use of deliberate changes to both offset and span of the couch calibrations and for the isocenter tests to an offset in the beam focal spot position. The study attempts to provide standard QA results in a form that is directly comparable to the equivalent MPC test. This study evaluates different MPC checks than those evaluated by Barnes and Greer 10 and hence the two bodies of work complement each other toward evaluating the whole of MPC as a linac QA device.

| METHODS
All measurements in this study were performed on a single Varian TrueBeam STx (software version 2.0) linac fitted with an aS1200 EPID and six degree of freedom couch.

2.A.1 | MPC Geometric checks
The MPC geometric tests utilize a series of kV and 6 MV beam images of the IsoCal phantom situated in a specific bracket on the IGRT couch top to assess: treatment/radiation isocenter size, MV and kV imager center pixel offsets from projected radiation isocenter, accuracy of collimator and gantry angles, accuracy of jaw and MLC leaf positions and accuracy of couch positioning including pitch and roll. All measurements are highly automated and the user is simply required to setup the IsoCal phantom and bracket onto the treatment couch at position H2 and to beam-on for each required energy. For the geometric tests the system makes all required motions automatically and beams on when all is in position. Images are automatically analyzed at the TrueBeam console and results are presented with a nonuser definable pass/fail criteria applied. Functionality for presenting trends in results is also available in the package. The relevant MPC checks to the Varian OBI/CBCT IGRT systems are the isocenter size, kV imager offset, and the couch tests.

2.A.2 | Winston-Lutz
The radiation isocenter position and size has been traditionally tested using the Winston-Lutz test. 11 For routine QA the department uses a variant of the Winston-Lutz test developed by Rowshanfarzad et al., 2011, 12 whereby the field is defined by a stereotactic cone and MV images are taken of a ball bearing that has been prepositioned at the imaging isocenter. Imaging is performed using the EPID in cine acquisition mode while a conformal gantry arc, collimator. or couch rotation is performed. An in-house developed MATLAB script (The Mathworks Inc., Natick, MA, USA) is used to calculate the position of the ball bearing compared to the center of the field for each EPID cine frame to allow calculation of the mean and maximum displacement of the ball bearing and hence imaging isocenter from the radiation field center defined by the stereotactic cone.

2.B.1 | Repeatability
Short-term repeatability of the MPC geometric tests was evaluated by taking five successive measurements and calculating the standard deviation.

2.B.2 | Isocenter
The radiation isocenter centroid is the ideal intersection point of the beams central axis over a full gantry rotation. The central beam axis BARNES AND GREER | 201 in MPC is defined by the center of rotation of the MLC. This is measured using EPID images of open MLC defined fields at eight representative gantry angles 45°apart. At each gantry angle two images are taken with 180 degree-opposed collimator angles to determine the beam central axis independent of MLC positional accuracy. In MPC, the size of the radiation isocenter spheroid is defined as the maximum distance of the beam's central axis from the idealized isocenter centroid. 13 The MPC isocenter size parameter reported is a single value, which means that no information is available on isocenter shape or in which direction the isocenter has deviated the most from the centroid. The MPC isocenter size tolerance is set at AE 0.5 mm. Besides isocenter size, MPC also reports the kV and MV imager offset. These parameters represent the maximum distance of the imager center from the projection of the radiation isocenter centroid. These parameters are included to provide a measure of the correctness of the IsoCal calibration, which is important for aligning the radiation and imaging isocenters and for CBCT image quality.

Isocenter size
The in-house Winston-Lutz analysis program reports both the maximum and mean measured deviation of the radiation field center of the cone from the ball bearing. The initial setup of the ball bearing using cone-beam CT places it at the estimated centroid of imaging isocenter. Results are presented in the plane of the EPID (scaled to isocenter distance) for both the panel inplane and "crossplane" directions. In this method, the mean deviation parameter represents the distance between the centroids of imaging and radiation isocenters and the maximum deviation represents the greatest distance between any point within the radiation isocenter and the centroid of imaging isocenter. By calculating the difference between measured maximum deviation and mean deviation and then calculating the vector magnitude from the inplane and crossplane components, the maximum size of the radiation isocenter spheroid is determined and is then directly compared to the MPC isocenter size parameter.

kV imager offset
For accurate IGRT the imaging system and radiation isocenters must coincide. 7 This is achieved on Varian linacs using the IsoCal calibration procedure, 14

2.B.3 | Couch
In the authors department couch readouts are tested by comparing the readout against an external measure at a few representative points. MPC does not report on the absolute couch positioning but rather on the measured distance traveled between two points. This is the most clinically important aspect of couch motion for ensuring that couch shifts based upon IGRT imaging are accurate. To allow a meaningful comparison between the MPC couch travel and the departmental couch tests, the departmental couch test results are presented in terms of difference between the two most extreme measurement points. This measured range is then scaled to the range over which the MPC travel range is measured. The difference of this measured range from expected is then compared to MPC.

Sensitivity to miscalibration
In an experiment to test the sensitivity of the MPC couch results the couch was deliberately miscalibrated in each axis using the standard calibration procedures. This was done in two ways. Firstly, a systematic offset was introduced into the calibration. This was done to all couch axes individually using offsets of the magnitude similar to the MPC tolerance. Secondly, the span of the calibration was made successively both smaller and larger. The magnitude of the miscalibration was calculated to cause error at about MPC tolerance.
The altered span miscalibration was performed only for couch lateral, longitudinal, and vertical. This was not possible for couch pitch, roll, and rotation because these utilized single point calibration procedures. In all cases the measured change in MPC was compared against expected from the miscalibration.

2.B.4 | Sensitivity of MPC to Phantom tilt
In an experiment to test the sensitivity of the MPC couch pitch and roll test to discrepancies in the phantom setup, the roll and pitch of the phantom were successively deliberately adjusted by approximately 1°. With the couch pitch and roll set to zero, the MPC bracket and phantom were attached to the couch top as per usual.
The digital spirit level was used on the top surface of the phantom to measure the phantom pitch. The spirit level was then placed on the phantom handle as an initial measure of roll. MPC was then performed. The roll of the phantom was then adjusted by wedging paper sheets between the phantom and bracket until a change in roll was measured on the spirit level of one degree. MPC was repeated with the paper wedges in situ. The wedges were then removed and placed under the phantom to induce a change in pitch of one degree on the spirit level and MPC was again repeated. Any changes in MPC parameters between the three acquisitions were recorded.

3.A | Repeatability
The results of Table 1 show how repeatable each of the MPC geometric tests were across five successive measurements. The repeatability results of Table 1 show that for all MPC tests the methods are repeatable to within 0.05 mm or 0.04 degrees for all parameters at one standard deviation.

3.B.1 | Isocenter size
The results of Fig. 1 show that the MPC measured isocenter size ranged between 0.29 mm and 0.37 mm over the period. No drift in the results was detected so a calculation of the mean and standard deviation was performed and found to be 0.34 AE 0.02 mm (1 SD).

3.B.3 | Sensitivity to focal spot position change
The results of Fig. 3 appear to show that the MPC isocenter size and kV imager offsets were relatively stable and constant before the focal spot adjustment. After the adjustment, the isocenter size appears unchanged, however, a systematic shift in the results is apparent for the kV imager offset results. The mean and standard deviation values for the isocenter size and kV imager offsets are presented both before and after the focal spot adjustment in Table 2 and the mean values were tested for statistical agreement using the t-test. The t-test shows that neither parameter was statistically

3.C | Couch
The results of Table 3 show that the MPC couch lateral mean measurement agrees within three standard deviations of the mechanical QA measurements. MPC couch vertical and rotation agree within two standard deviations and couch longitudinal, pitch, and roll within one standard deviation. The results of Table 4 show the MPC couch measurements after changes in the calibration span for the lateral, longitudinal, and vertical axes. The results show agreement between MPC and the expected value to within 0.04 mm.

3.C.1 | Sensitivity to miscalibration
F I G 2 . MPC kV imager offset and kV imager offset and in-house Winston-Lutz distance between imaging and radiation isocenter.  Table 1.

4.A | Repeatability
The repeatability results of Table 1 are well inside the tolerances for all tests indicating that the tolerances are meaningful in that recorded fails are distinguishable from day to day variation.

4.B | Isocenter
The statistical disagreement using the t-test between MPC isocenter The results of Table 2 and of the t-test show that the kV imager offset parameter is sensitive to changes in beam focal spot position.
This is expected as the effect of altering the beam focal spot is to shift the beam laterally and hence the projection of the radiation isocenter will shift relative to the imager center. Since the kV imager offset is primarily a measure of the correctness of the IsoCal calibration the results suggest that the IsoCal calibration should be performed after any focal spot position beam steering.

4.C | Couch
The insensitivity of MPC to offsets introduced into the couch cali-

4.D | Sensitivity to phantom tilt
The lack of sensitivity of the MPC geometric tests to varying the phantom pitch and roll by one degree indicates that none of the tests are reliant on the accurate pitch and roll of the phantom. The MPC couch pitch and roll tests as well as the gantry-relative tests are based upon the relative changes across multiple images of the phantom. As such, the pitch and roll of the phantom cancels out and does not affect the measurement.

| CONCLUSION
For accurate IGRT the radiation isocenter size, coincidence of radiation isocenter with imaging isocenter and accuracy of couch shifts must all be accurately quantified. The MPC checks are adjudged to be accurate for radiation isocenter size and for couch shift accuracy.
The kV imager offset parameter does not provide a direct measure of radiation to kV isocenter coincidence, but acts as a surrogate.
However, if a fail in kV imager offset is recorded then redoing the IsoCal calibration is indicated. The IsoCal calibration should then improve alignment between the radiation and kV isocenter spheroids. For a daily test of isocenter alignment, the MPC kV imager offset should suffice and could be assured with a less frequent Winston-Lutz or Isocal verification measurement.