Commissioning and QA

In Vivo Dosimetry

In vivo dosimetry is a method of quality assurance (QA) used to verify the dose delivered to the patient. These measurements can reveal clinically relevant differences (or preferably, similarities) between the planned and delivered dose and potentially satisfy legal requirements (according to free responses provided). In a 2020 survey, about half—53.6%—of clinicians noted performing in vivo dosimetry measurements routinely.

The most common treatment types for routine in vivo dosimetry were static photon fields and electron fields (each with 26.1% of respondents selecting this option). Other treatment types with frequent in vivo dosimetry use are Total Body Irradiation (11.6%), dynamic MLC fields (9.0%), dynamic arc plans, Total Skin Electrons (TSE), and Pacemakers (7.2% each) and SBRT, SRS, and brachytherapy with 2.9%, 1.4%, and 1.4% respectively.

Commissioning a Varian Linac

Commissioning is typically performed two ways, either by the in-house physics staff or outsourcing to a contractor or specialist group. According to a 2019 survey, 51.2% of clinics used in-house physicists to commission their linac, 41.9% used a contractor service, and 6.9% used a combination of the two.

Data acquisition is an essential step in the commissioning process. Physicists use the data to model treatment planning algorithms, which directly impact treatment delivery accuracy. In a separate survey regarding data acquisition, 30.4% of departments used data collected by a contractor to create their models, 35.7% used in-house physicist data, and 26.8% relied on preconfigured golden beam data. The remaining 9.8% compared their measured data to the provided golden beam data and used the vendor data if it was within tolerance. Of those who measured their data, 89.7% performed various manipulations to create smooth and symmetrical profiles before importing them into the treatment planning system.

Data Collection For Commissioning

Commissioning: 3D Tank -vs- 1D Tank with array

Beam data gathered at commissioning forms the basis for beam modeling. Errors made during commissioning will present systematically, diffusing downstream to affect all treatments performed with said machine. The need for high-quality data increases as field sizes decrease and plan complexity increases. The World Health Organization (WHO) estimated that 24% of reported “adverse events” in radiation oncology stemmed from the commissioning stage. Therefore, it is imperative that commissioning measurements are completed accurately and verified rigorously.

In a recent 2020 survey, 67.2% of respondents indicated the 3D scanning tank was the preferred method for collecting beam data for commissioning purposes. Indeed, the 3D tank allows for a more extensive data set over a broader range of depths and field sizes. However, due to variability in QA standards, equipment, and funding, 1D tanks are a potential alternative for commissioning. With that said, one respondent noted that utilizing a 1D scanner with a 2D array “would require excellent calibrations and multiple exposures at multiple depths to represent everything a 3D tank would do for you.”

Those that utilize 1D tanks coupled with a 2D array justify their choices through time savings, lower cost, and because pre-configured models are typically of high quality, requiring spot-checking only. One user noted that “Varian machines have a reliable history with beam stability. The test plans verify your algorithms calculate accurately, and the 1D and profiler scans help establish baselines for constancy checks”.

Testing for Linac Head Leakage

One of IAEA’s recommended safety checks during acceptance is testing the linac head for leakage. These tests are performed by closing collimator jaws and covering the treatment head with film or mounted detectors to locate any hotspots. Despite agency recommendations, only 43.9% of respondents chose to test for head leakage when accepting a new linac.

Varian’s Machine Performance Check (MPC)

Available on the Varian TrueBeam, the Machine Performance Check (MPC) is a tool that automates a sequence of kV and MV images at predefined gantry, collimator, and couch angles to perform mechanical checks in a matter of minutes. While it doesn’t execute the full range of recommended monthly mechanical tests, it does provide the ability to track and trend the results. Its ease of use and the practicality have convinced 87.5% of clinicians to adopt it. Among users, 63.0% performed the checks daily, 6.5% weekly, and 17.4% monthly. While the mechanical data is useful, less than half (44.7%) of physicists accepted it as a full replacement for routine QA tests, opting to use it as an adjunct to current practices. Those who chose to replace some standard QA tests recommended comparing with previous QA and commissioning results before doing so to ensure validity.

Linac Replacement

Replacing a linac is a significant undertaking. It is time-consuming and expensive. But unlike some consumer technology in which planned obsolescence is built-in, well-maintained linacs can serve departments for many years. Depreciation schedules often operate on a 10-year schedule financially, but according to a survey of both physicists and linac engineers, 79.1% lasted longer than 12 years. After a decade, the linacs are likely less advanced than newer technology, but the sentiment that linacs can run upwards of 20 years was popular in comments.

Linac Machine Years of Service Before Replacement

MLC Parameters

In the past couple of decades, IMRT use has become widespread due to its ability to deliver highly conformal doses. IMRT uses multi-leaf collimators (MLCs) to shape the fields into small beamlets to achieve the desired dose distribution. As a result of the added complexity, treatment accuracy is highly influenced by TPS modeling. Improper modeling of MLCs from flawed commissioning data causes discrepancies in physical linac output and TPS model, which is a primary reason for failed gamma analysis in IMRT plans. Currently, there is no reference data from regulatory agencies on MLC parameters. However, Glenn et al. sourced 2818 models from 642 institutions to help institutions spot potentially suspicious values. The survey data exhibits similarilarities to Glenn’s data regarding the average dosimetric leaf gap (DLG) and MLC transmission values for the Millennium 120 MLC. The 2013 survey’s average DLG value was 1.6mm, while Glenn’s was 1.8 mm. For MLC transmission, the most common transmission factor was 0.015 in both sets of data.

Dosimetric Leaf Gap — Millennium 120 MLC 6MV

MLC Transmission — Millennium 120 MLC 6MV