iCheckNet - TrueBeam Machine QA

About iCheckNet

iCheckNet was developed in-house by Jochen Reiterer and Michael Papauschek (supervision of .NET coding).

Background Information

The basis for machine QA on one hand is the ONORM S5290 series of Austrian standards.

Since the last update (ONORM S5290 has been renamed to S5290-1) and the publication of two additional parts (S5290-2: intensity-modulated radiation therapy and S5290-3: electronic image receptors), the time needed for some checks on the linac has increased considerably.

On the other hand, we follow certain Varian recommendations regarding safety checks, imager recalibrations, IsoCal etc., which of course are not covered by published QA standards. Typical default expiration period for Flood Field, Pixel Correction Maps, CBCT Air Norm and Crescent, Optical Stereo Camera, Isocenter and kV Collimator is 30 days.

Daily Check

With the QUICKCHECKwebline, dose output, symmetry, beam quality factor etc. is determined for all 11 energies (5 photons and 6 electrons). The measurement results are not stored in iCheckNet, but in QUICKCHECK 1.5 software. In iCheckNet, only a checkbox ("DosisCheck QCWebline") indicates that the measurement has been done.

The beam-on hours, as displayed on the modulator cabinet, are also recorded. This is needed for an estimation of the weekly workload. We assume an average dose rate of 360 Gy/hr (6 Gy/min) for all beam deliveries. The beam-on time from Monday to Monday is multiplied by 360. This gives the dose per week delivered on the TrueBeam. The average over one year must not be over 3000 Gy (work load limit per machine).

Various warning lamps, intercom, backup MU counter are also checked every morning. A separate document describes what needs to be done to complete the "Imager-/Safety-Tests". It includes acquisition of kV-MV-image pairs, matching, online correction, dry run and check of safety features like collision protection and others (e.g., opening the maze door during beam-on). A picket fence pattern is also imaged and evaluated visually. This has once led to the discovery that the collimator rotation can be off by as much as 0.8° on the TrueBeam.

The "Catalyst-Test" is a check of the surface scanning system CatalystHD.

Weekly Check - Perfect Pitch

The functionality of Varian's 6DoF-Couchtop is checked weekly. It is a simple end-to-end test with the QUASAR Penta-Guide phantom.

The cube has been modified with two rubber feet. When put on a level surface, it will stand slightly tilted. The tilting is smaller than 3° in the inplane and the crossplane direction, so that it can be corrected with the couch (max pitch and roll is 3°).

The "PerfectPitch" plan is loaded on the TrueBeam, and the phantom is freely positioned on the couch, close to isocenter. It may be rotated and shifted in all three directions (see laser on phantom):

Here is a photo which shows one of the rubber feet:

This tilting, plus the rotation of the couch shall be corrected after CBCT imaging and matching.

The CBCT is acquired next,

followed by an auto-match. CBCT plus match takes about a minute:

This is the match result:

Once more the deltas that need to be applied:

After applying the six deltas,

the cube is level and centered. This will be verified by visual inspection in a minute. But first, a treatment field is delivered with MV image, and centering is checked:

Now we enter the room and look at the lasers:

Correction has been correctly applied.

One can see that the couchtop does the compensation:

Monthly Check

Although imager recalibrations are not "machine QA" in the strict sense (tests of functional performance characteristics are required by a standard), we perform these recalibrations in the context of the QA checks, and document both in the same report.

Due to the default expiration period of 30 days for most imaging calibrations (see above), a significant part of the time needed for the monthly check goes into PVA calibration.

Isocenter Calibration ("IsoCal") in TrueBeam 2.0 has a longer expiration period than the Verification of Isocenter Calibration ("IsoCal Verification"). But once a year the Calibration has to be repeated, even if the last Verification was fine.

This is the TrueBeam screen after Isocenter Calibration,

and after IsoCal verification:

The lengthy programme is the reason why we usually divide the monthly test into several afternoons for it. Note the bottom line of the data entry windows, which contains a "Completed" checkbox. Since iCheckNet alerts the user with a yellow exclamation mark on the main screen that a certain check will soon become due (for the monthly check, this happens after two weeks of each month), the problem had to be solved that as soon as a new check was created (but not completed!), the exclamation mark did not appear any more. The checkbox avoids this. It has to be checked manually, but this tells the software that all tasks of the check were done, and the monthly check is complete. This is only relevant for checks that are sometimes performed "in portions", due to linac accessibility.

The monthly check has two side documents: one is the EPIQA RapidArc Commissioning Report, the other is the VeriSoft report for the leafspeed test.

Quarterly Check

The HU values on tab 1 are taken from the report of the semi-automatic AutoQA evaluation. An instruction document describes the procedure how to export a CBCT scan from ARIA, import it in AutoQA software, and evaluate the scan.

The AutoQA report also covers low contrast resolution, uniformity and slice thickness. Uniformity and slice thickness are required by ONORM S5290-3, and are copied to iCheckNet.

High contrast resolution and the size of the reconstructed phantom are also required, but have to be determined manually. We do this in Offline Review.

The quarterly check has two side documents, the AutoQA reports for the HEAD and PELVIS scans.

Half-Yearly Check

For the Half-Yearly Check, the machine is down for two full days. The reason is mainly the large MP3 water phantom, which is used for scanning depth dose curves and absolute dosimetry.

Profiles were measured in water previously, but we now measure them only with the STARCHECKmaxi. If it turns out that beam angle servos have to be adjusted, this is done with STARCHECKmaxi anyway (using BeamAdjust software). Therefore, we see no reason to measure profiles in water.

For depth dose scanning, we use a Farmer chamber (photons) and a Roos plane-parallel chamber (electrons). The reason is the following. Although absolute dose is measured in 10 cm depth (at SSD 100 cm), we still want to maintain the 100 MU = 1 Gy output calibration at dmax depth. This means that we need a depth dose curve of a 10 x 10 cm field, to link the two depths. We use a Farmer chamber chamber, because for this type saturation correction parameters were determined by Bruggmoser. This is convenient for FFF energies.

Beam characterisation on the other hand was done with small volume chambers, to get a consistent data set even for small field sizes. This is no problem here, because field size is always 10 x 10 cm during absolute dosimetry.

This also explains why the reference depth dose curve for the half-yearly check is not from the Eclipse beam characterisation data set, but from an early absolute dose measurement on the TrueBeam which was used to calibrate the monitor chambers.

Recently, the evaluation protocol in PTW MultiCheck was changed, because now the PTW software understands the difference between flattened and unflattened beams. It automatically uses the FFF - Varian protocol, which implements the methodology of the Bellinzona group for FFF beams.

The Civco Iso-Align phantom is used for checking light field/radiation field coincidence at Gantry 90°/270°, together with the EPID:

PTW's EPID QC phantom is used to cover the MV image quality tests as required by S5290-3 (high and low contrast resolution of MV images):

The evaluation is done with PTW's "epidSoft" software.

For each of the six acquisition modes (6, 10 and 15 MV "high res" and "integrated"), a PDF is created (see 6X highres for example).

The PDFs are not printed, to save paper. A test where the check interval (6 months) is much longer than the recalibration period (1 month for the kV and MV imaging systems) is rather questionable. But the test is required by S5290-3.

The kV high and low contrast resolution is determined visually.

No evaluation software comparable to epidSoft exists for the kV images.

Literature

ÖNORM S 5290-1: Medical electron accelerators - Part 1: Constancy tests of functional performance characteristics. Austrian Standards Institute, 2011-10-15.

ÖNORM S 5290-2: Medical electron accelerators - Part 2: Constancy tests of functional performance characteristics in intensity-modulated radiation therapy. Austrian Standards Institute, 2012-07-15.

ÖNORM S 5290-3: Medical electron accelerators - Part 3: Constancy testing of functional performance characteristics of electronic image receptors. Austrian Standards Institute, 2013-06-01.

Since our first report in 2005, our home-made Physics QA software "iCheck" has grown considerably.

(Main screen of iCheckNet. Checks which will be due soon are marked by the software.)

It is now called "iCheckNet" and includes (from left to right on the main screen):

Daily, weekly, monthly, quarterly, half-yearly and yearly checks on the TrueBeam,
daily, weekly and monthly checks on the LAP laser system which is installed on the CT (our Virtual Simulator or VSim),
statistics of the weekly radiation workload on the CT ("Total mAmin"),
reports of first-day-treatments ("Daily Reports"),
quaterly checks of medical image displays ("MonitorCheck"),
a one-page overview of all Checks (excluding daily and weekly) with their day of completion ("CheckView"),
a log of daily checks of the defibrillator ("DefiCheck"),
an application for checking in and out radiation dosimetry badges using a barcode reader ("PersonenDosimeterListe"),
a field match calculation program ("FMC", work in progress),
Dose - Electron - MonitorUnit Calculation ("DeMC"), which is the main application for forward calculation of MUs (photons and electrons), secondary plan checks (calculation of dose from MU), and entering the results of Portal Dosimetry plan QA measurements (EPIQA and Varian Portal Dosimetry),
monthly checks of the Eclipse TPS ("Planungssystem MonatsCheck"),
calculation sheets and documentation of absolute dose measurements ("AbsDosimetry Linac"), and finally
documentation of all VMAT and IMRT plan QA measurements ("Portal Dosimetry QA").

Currently, the user interface uses a mixture of German and English, but this is constantly changing. The single items will become clear when looking at the detailed screenshots. This page covers TrueBeam Machine QA only (leftmost column on the iCheckNet main screen).

TrueBeam - Daily Check

The daily check on the TrueBeam is typically done by a therapist. Every morning around 6:00 a.m., several tests and measurements are performed:

Reports can be printed for the morning check itself, and for the beam hours (for "copy to clipboard" and further processing in Excel). This is done typically only once a year, shortly before the ministry of health visits us for the regulatory review.

TrueBeam - Weekly Check

Weekly dose measurements have to be performed with a verified measurement system. We use the Unidos/Farmer combination for that. Unidos is controlled by iCheckNet.

iCheckNet first searches the COM ports for an Unidos device and connects for it. Then Unidos is controlled via the measurement window:

This is the tab for the dose measurements after they have been completed:

The second tab is for the check of isocenter display (light field crosshairs and backpointer). It has four subtabs for the four major Gantry angles. Since the deactivation of the backpointer lasers some years ago (nobody missed them!), we only check the deviation of the light field cross hair at Gantry 0°, 90° and 270°:

The third tab covers emergency-off buttons, a visual check of electron applicators and their inserts, the accessory mount, and various periodic rebootings (as recommended by Varian). It also includes a test of the 6-DoF PerfectPitch couch, where a cube is imaged, matched, and online-corrected:

The final report of the weekly check has two pages. It is printed on a duplex printer and signed. Only one sheet of paper is produced per week.

TrueBeam - Monthly Check

The window for data entry has five tabs. Tab one is about the size of light field (checked with ruler paper), the ODI display (checked with front pointers) and the vertical couch readout:

The second tab covers stability during rotational therapy and the light field of irregular MLC shapes. For the former we do more than required, by performing a full RapidArc commissioning per month (evaluation with EPIQA), cycling through the energies. For the latter we use a Beam's-Eye-View printout of two irregular MLC fields and check the accuracy of the leaf positions via light field:

Tab three contains visual evaluation of the picket fence image acquired in the morning, and a test of leaf speed constancy, which is described in ONORM S5290-2:

The integrated images acquired in the leafspeed test are compared in VeriSoft, and a report is created.

The fourth tab contains items from ONORM S5290-3, such as accuracy of isocenter display with radiation, and noise on kV and MV images:

Varian's imaging phantom is used to acquire four pairs of orthogonal kV-MV-image pairs. The CAX shift on the images is documented. If it is too large, usually the lasers have to be adjusted.

The most time-consuming part of the check is on the last tab. It contains calibration of all imaging modes (kV, MV, CBCT), plus the IsoCal check/recalibration:

As with the weekly check, the report of the monthly check, which is printed and signed, fits on a single sheet of paper.

TrueBeam - Quarterly Check

On the first tab, beam symmetry results from the morning check on the same day are filled in. The remaining check is about CBCT imaging. The Catphan 504 is imaged with both HEAD and PELVIS norms, and HU values are recorded:

On the second tab, the characteristics of geometrical reconstruction are assessed. Size of the reconstructed Catphan phantom, high contrast resolution, uniformity and slice thickness are checked:

The printout of the quarterly check has two pages. Side documents are printed for HEAD and PELVIS norms, which contain more evaluations.

TrueBeam - Half-Yearly Check

In iCheckNet, the check has eight tabs. The first tab contains the results from the water phantom scans:

The second tab contains measured inplane and crossplane field size and symmetry values, for the 10 x 10 cm and 40 x 40 cm photon field (measured with STARCHECKmaxi):

Tab three covers Electron beam symmetry and Enhanced Dynamic Wedge factors. The EDW factors are measured with the EPID and evaluated in EPIQA (side document: EPIQA wedge factor report):

In tab four, the coincidence of light field and radiation field is checked. With the Gantry at 0°, this is done using STARCHECKmaxi and the FieldCheck test object at the two distances of 100 and 150 cm. The Civco Iso-Align phantom is used for Gantry 90°/270°, together with the EPID.

A dosimetric measurement of two irregular MLC fields (measurement with 2D-Array, evaluation with VeriSoft) is also on this tab:

Tab five contains an unusual application of the QUICKCHECKwebline: here it is not used for measuring dose, but simply to check the light field at Gantry 90°/270°.

The second part is a field abutment test. Four quadrants of 10 x 10 cm add up to a 20 x 20 cm field. The test is first performed with the jaws and once more with the MLC (2D-Array measurement):

The accuracy of digital readouts (gantry, collimator rotation, couch rotation) is documented on tab six:

Tab seven contains a picketfence test performed at the four major gantry angles (see plan parameters), and an interrupted DMLC treatment. Both the picketfence and the interrupt are imaged and evaluated visually in Portal Dosimetry:

The last tab covers MV and kV image quality:

The printout of the half-yearly check has six pages.

TrueBeam - Yearly Check

Machine performance parameters which are not expected to change rapidly are checked at fairly lengthy intervals. Tab one is about the proportionality and reproducibility of the dose monitor chambers (photons):

Tab two continues with the same for the highest electron energy.

Then the dependence of measured dose on gantry angle (static) is checked:

Tab three repeats the test for rotational therapies at MaxMU/° and MinMU/°.

The second topic is the stability of dose output during a full working day:

Dependence of photon beam symmetry on gantry angle is checked with the EPID. Electron beams are checked with the QUICKCHECK, which is attached to a gantry holder: