On the use of trajectory log files for machine & patient specific QA

ACPSEM position paper: pre-treatment patient specific plan checks and quality assurance in radiation oncology

The Australasian College of Physical Scientists and Engineers in Medicine (ACPSEM) has not previously made recommendations outlining the requirements for physics plan checks in Australia and New Zealand. A recent workforce modelling exercise, undertaken by the ACPSEM, revealed that the workload of a clinical radiation oncology medical physicist can comprise of up to 50% patient specific quality assurance activities. Therefore, in 2022 the ACPSEM Radiation Oncology Specialty Group (ROSG) set up a working group to address this issue. This position paper authored by ROSG endorses the recommendations of the American Association of Physicists in Medicine (AAPM) Task Group 218, 219 and 275 reports with some contextualisation for the Australia and New Zealand settings. A few recommendations from other sources are also endorsed to complete the position.

Investigation of a 3D dosimetry system utilizing radio-chromic gel and telecentric optical-CT readout

BACKGROUND

Sophisticated modern radiation therapy treatments require comprehensive validation in 3D.

PURPOSE

Investigation and characterization of a novel 3D dosimetry system consisting of ClearView radiochromic gel dosimeters (commercially available from Modus Inc) and an in-house telecentric optical CT scanner DLOS (the Duke Large Field of View Optical-CT Scanner).

METHODS

Spectrophotometry measurements were made on small volumes of ClearView gel irradiated with 6X photon doses up to 40 Gy to determine linearity and temporal stability of dose response. Clinical evaluation of Clearview/DLOS system was conducted in two phases. Phase one involved simple photon and electron benchmark irradiations, delivered to 15 and 10 cm diameter dosimeters, at various energies and doses. Phase 2 investigated application to the verification of two single isocenter multi-target (SIMT) stereotactic radiosurgery (SRS) deliveries. These were patient treatments for two and five brain lesions, respectively, and delivered to 15 cm diameter dosimeters. SIMT treatments were delivered by Varian TrueBeam 6X with doses of 40 Gy. For dose read-out, dosimeters were optically scanned in the DLOS both pre- and post- irradiation (within 24 h). 3D reconstructions (1 mm3 resolution) of the change in linear-optical- attenuation (proportional to dose) was obtained using in-house software and 3D Slicer. Measured and predicted (Eclipse TPS) doses were compared through percent depth-dose (PDD), cross plane and in-plane profiles, and relative 3D gamma analysis (performed at a range of 7%/4 mm down to 2%/2 mm). Regions of known artifacts were excluded from analysis (jar base, neck, and wall). The SIMT SRS deliveries were additionally compared to SciMoca, an independent Monte Carlo second check software.

RESULTS

Linearity of dose response was confirmed with R2 ≥ 0.9986 at both 520 and 630 nm wavelengths and at three post-irradiation time points: 21 h, 6 and 10 days. Dose profiles of all benchmark irradiations, in both 15 and 10 cm dosimeters, show good agreement in useable areas of the gel compared to Eclipse dose calculations, with root mean square errors (RMSE) ≤ 0.0054, and R2 ≥ 0.9808. Gamma pass rates for the 15 cm dosimeter benchmark irradiations were ≥ 94% at 2%/2 mm (central axis), ≥ 90% at 3%/3 mm (left lateral), ≥ 90% at 2%/2 mm (electron), and ≥ 94% at 3%/2 mm (stacking field). Similar high passing rates were observed for benchmark irradiations to the smaller 10 cm diameter dosimeters. Very high Gamma pass rates were found for SIMT SRS deliveries, with 99.82% and 97.80% at 3%/2 mm, for the two and five target plans, respectively.

CONCLUSION

This work presents the first investigation of ClearView dosimeters in combination with a telecentric optical-CT scanner (DLOS). Simple benchmark irradiations demonstrate ClearView/DLOS can accurately recreate and measure relative 3D dose within non-artifact regions (i.e., > 1 cm away from walls). Application to SIMT SRS deliveries demonstrated the viability of the system as a means for comprehensive 3D verification of complex treatment deliveries as well as confirming the treatment planning system dose distribution. The results indicate that DLOS/ClearView system is a highly effective 3D verification tool for SIMT irradiations and can be applied with 3%/2 mm gamma criteria where passing rates of > 95% are to be expected.

Feasibility of 4D VMAT-CT

Objective.Feasibility of three-dimensional (3D) tracking of volumetric modulated arc therapy (VMAT) based on VMAT-computed tomography (VMAT-CT) has been shown previously by our group. However, 3D VMAT-CT is not suitable for treatments that involve significant target movement due to patient breathing. The goal of this study was to reconstruct four-dimensional (4D) VMAT-CT and evaluate the feasibility of tracking based on 4D VMAT-CT.Approach.Synchronized portal images of phantoms and linac log were both sorted into four phases, and VMAT-CT+ was generated in each phase by fusing reconstructed VMAT-CT and planning CT using rigid or deformable registration. Dose was calculated in each phase and was registered to the mean position planning CT for 4D dose reconstruction. Trackings based on 4D VMAT-CT+ and 4D cone beam CT (CBCT) were compared. Potential uncertainties were also evaluated.Main results.Tracking based on 4D VMAT-CT+ was accurate, could detect phantom deformation and/or change of breathing pattern, and was superior to that based on 4D CBCT. The impact of uncertainties on tracking was minimal.Significance.Our study shows it is feasible to accurately track position and dose based on 4D VMAT-CT for patients whose VMAT treatments are subject to respiratory motion. It will significantly increase the confidence of VMAT and is a clinically viable solution to daily patient positioning,in vivodosimetry and treatment monitoring.

Impact of systematic MLC positional uncertainties on the quality of single-isocenter multi-target VMAT-SRS treatment plans

Purpose

To study the impact of systematic MLC leaf positional uncertainties (stemming from mechanical inaccuracies or sub‐optimal MLC modeling) on the quality of intracranial single‐isocenter multi‐target VMAT‐SRS treatment plans. An estimation of appropriate tolerance levels is attempted.

Methods

Five patients, with three to four metastases and at least one target lying in close proximity to organs‐at‐risk (OARs) were included in this study. A single‐isocenter multi‐arc VMAT plan per patient was prepared, which served as the reference for dosimetric impact evaluation. A range of leaf offsets was introduced (±0.03 mm up to ±0.30 mm defined at the MLC plane) to both leaf banks, by varying the leaf offset MLC modeling parameter in Monaco for all the prepared plans, in order to simulate projected leaf offsets of ±0.09 mm up to ±0.94 mm at the isocenter plane, respectively. For all offsets simulated and cases studied, dose distributions were re‐calculated and compared with the corresponding reference ones. An experimental dosimetric procedure using the SRS mapCHECK diode array was also performed to support the simulation study results and investigate its suitability to detect small systematic leaf positional errors.

Results

Projected leaf offsets of ±0.09 mm were well‐tolerated with respect to both target dosimetry and OAR‐sparing. A linear relationship was found between D_95% percentage change and projected leaf offset (slope: 12%/mm). Impact of projected offset on target dosimetry was strongly associated with target volume. In two cases, plans that could be considered potentially clinically unacceptable (i.e., clinical dose constraint violation) were obtained even for projected offsets as small as 0.19 mm. The performed experimental dosimetry check can detect potential small systematic leaf errors.

Conclusions

Plan quality indices and dose–volume metrics are very sensitive to systematic sub‐millimeter leaf positional inaccuracies, projected at the isocenter plane. Acceptable and tolerance levels in systematic MLC uncertainties need to be tailored to VMAT‐SRS spatial and dosimetric accuracy requirements.

Machine learning models to predict the delivered positions of Elekta multileaf collimator leaves for volumetric modulated arc therapy

PURPOSE

Accurate positioning of multileaf collimator (MLC) leaves during volumetric modulated arc therapy (VMAT) is essential for accurate treatment delivery. We developed a linear regression, support vector machine, random forest, extreme gradient boosting (XGBoost), and an artificial neural network (ANN) for predicting the delivered leaf positions for VMAT plans.

METHODS

For this study, 160 MLC log files from 80 VMAT plans were obtained from a single institution treated on 3 Elekta Versa HD linear accelerators. The gravity vector, X1 and X2 jaw positions, leaf gap, leaf position, leaf velocity, and leaf acceleration were extracted and used as model inputs. The models were trained using 70% of the log files and tested on the remaining 30%. Mean absolute error (MAE), root mean square error (RMSE), the coefficient of determination R² , and fitted line plots showing the relationship between delivered and predicted leaf positions were used to evaluate model performance.

RESULTS

The models achieved the following errors: linear regression (MAE = 0.158 mm, RMSE = 0.225 mm), support vector machine (MAE = 0.141 mm, RMSE = 0.199 mm), random forest (MAE = 0.161 mm, RMSE = 0.229 mm), XGBoost (MAE = 0.185 mm, RMSE = 0.273 mm), and ANN (MAE = 0.361 mm, RMSE = 0.521 mm). A significant correlation between a plan's gamma passing rate (GPR) and the prediction errors of linear regression, support vector machine, and random forest is seen (p < 0.045).

CONCLUSIONS

We examined various models to predict the delivered MLC positions for VMAT plans treated with Elekta linacs. Linear regression, support vector machine, random forest, and XGBoost achieved lower errors than ANN. Models that can accurately predict the individual leaf positions during treatment can help identify leaves that are deviating from the planned position, which can improve a plan's GPR.