The sensitivity of gamma index analysis to detect multileaf collimator (MLC) positioning errors using Varian TrueBeam EPID and ArcCHECK for patient-specific prostate volumetric-modulated arc therapy (VMAT) quality assurance

Abstract

BackgroundDue to the increased degree of modulation and complexity of volumetric-modulated arc therapy (VMAT) plans, it is necessary to have a pre-treatment patient-specific quality assurance (QA) programme. The gamma index is commonly used to quantitatively compare two dose distributions. In this study we investigated the sensitivity of single- and multi-gamma criteria techniques to detect multileaf collimator (MLC) positioning errors using the Varian TrueBeam Electronic Portal Imaging DeviceTM (EPID) dosimetry and the ArcCHECKTM device.Materials and methodsAll active MLC positions of seven intact prostate patients VMAT plans were randomly changed with a mean value of 0.25, 0.5, 1 and 2 mm and a standard deviation of 0.1 mm on 25, 50, 75 and 100% of the control points. The change in gamma passing rates of six gamma criteria of 3%/3 mm, 3%/2 mm, 3%/1 mm, 2%/2 mm, 2%/1 mm and 1%/1 mm were analysed individually (single-gamma criterion) and as a group (multi-gamma criteria) as a function of the simulated errors. We used the improved and global gamma calculation algorithms with a low dose threshold of 10% in the EPID and ArcCHECK software, respectively. The changes in the planning target volume dose distributions and the organs at risk due to the MLC positioning errors were also studied.ResultsWhen 25, 50, 75 and 100% of the control points were modified by the introduction of the simulated errors, the smallest detectable errors with the EPID were 2, 1, 0.5 and 0.5 mm, respectively, using the multi-gamma criteria technique. Similarly for the single-gamma criteria technique errors as small as 2, 1, 1 and 1 mm applied to 25, 50, 75 and 100% of the control points, respectively, were detectable using a 2%/2 mm criterion. However, the smallest detectable errors with the ArcCHECK when using the multi-gamma criteria technique were 2, 2 and 1 mm when MLC errors were applied on 50, 75 and 100% of the control points. When only 25% of the control points were affected the ArcCHECK were unable to detect any of the errors applied. No noticeable difference was observed in the sensitivity using the single- or the multi-gamma criteria techniques with the ArcCHECK.ConclusionThe Varian TrueBeam EPID dosimetry shows a higher sensitivity in detecting MLC positioning errors compared with the ArcCHECK regardless of using the single- or the multi-gamma criteria techniques. Higher sensitivity was observed using the multi-gamma criteria technique compared with the single-criterion technique when using the EPID.

Detecting MLC modeling errors using radiomics‐based machine learning in patient‐specific QA with an EPID for intensity‐modulated radiation therapy

“…We conducted the present study to investigate whether the radiomics‐based method combined with machine learning is effective in detecting errors in MLC modeling, that is, the MLC TF, and DLG errors and an MLC positional error distinguished from the error‐free condition, by evaluating fluence maps of clinical IMRT plans measured with an EPID. The spatial resolution of the EPID is superior to that of detector arrays, and the EPID shows high sensitivity in the tuning or quality control of MLC modeling parameters and in detecting MLC positional errors in patient‐specific IMRT QA 5,27‐32 . In this study, we used calculated and measured fluence difference maps instead of gamma maps, since it was shown that simple subtracted maps were more sensitive to MLC positional errors than gamma maps 26 …”

Predicting gamma passing rates for portal dosimetry‐based IMRT QA using machine learning

“…In this study, we attempt to overcome these limitations by using multiple tree‐based ML methods to predict IMRT QA results generated by portal dosimetry. Portal dosimetry provides several advantages for IMRT QA: it is simple to use; it does not require a phantom and it is rigidly attached to the gantry, thus eliminating potential setup errors; and its high resolution and per‐beam analysis capability supports high sensitivity in detecting, for example, MLC positioning errors and thus enables 2%/2 mm gamma analysis . Several machine learning methods have been implemented that use the LINAC type, beam characteristics and plan complexity metrics as input features to accurately predict IMRT gamma passing rates.…”

Assessment of the Sun Nuclear ArcCHECK to detect errors in 6MV FFF VMAT delivery of brain SABR using ROC analysis

“…This is in line with a previous study, which was based on the treatment plans of seven prostate patients who received VMAT and which found that the ArcCHECK was able to detect MLC leaf position errors of 1 mm under similar circumstances when a 2%/2 mm criterion was used. 21 The same study found the ArcCHECK to be inferior to the electronic portal imaging device (EPID) also tested in terms of MLC leaf position error detectability. The superiority of EPID systems with regards to the detectability of MLC leaf position errors was confirmed by a study which introduced systematic MLC leaf position errors into IMRT and VMAT treatment plans and which used an EPID with a spatial resolution of 0.392 mm.…”

“…13,14 ROC curves are particularly useful for evaluating detector performance because they are independent of biases in the decision threshold which determines whether a plan passes or fails the QA procedure. 15 Examples of ROC-based error detectabiliy studies include research by Carlone et al, 15 McKenzie et al, 16 Bojechko & Ford, 17 Nithiyanantham et al, 18 Liang et al, 19 Sjölin & Edmund, 20 Maraghechi et al, 21 and Scarlet. 22 However, combining the findings even of studies investigating the same detector can prove difficult because of limitations such as a small dataset, no differentiation between different treatment sites or delivery techniques, or some sources of error not having been studied.…”