Consequences of the spectral response of an a-Si EPID and implications for dosimetric calibration

Deep learning-based tools to distinguish plan-specific from generic deviations in EPID-based in vivo dosimetry

BACKGROUND

Dose distributions calculated with electronic portal imaging device (EPID)-based in vivo dosimetry (EIVD) differ from planned dose distributions due to generic and plan-specific deviations. Generic deviations are characteristic to a class of plans. Examples include limitations in EIVD dose reconstruction, inaccuracies in treatment planning system (TPS) calculations and systematic machine deviations. Plan-specific deviations have an unpredictable character. Examples include discrepancies between the patient model used for dose calculation and the patient position or anatomy during delivery, random machine deviations, and data transfer, human or software errors. During the inspection work performed with traditional γ-evaluation statistical methods: (i) generic deviations raise alerts that need to be inspected but that rarely lead to action as their root cause is usually understood and (ii) the detection of relevant plan-specific deviations may be hindered by the presence of generic deviations.

PURPOSE

To investigate whether deep learning-based tools can help in identifying γ-alerts raised by generic deviations and in improving the detectability of plan-specific deviations.

METHODS

A 3D U-Net was trained as an autoencoder to reconstruct underlying patterns of generic deviations in γ-distributions. The network was trained for four treatment disease sites differently affected by generic deviations: volumetric modulated arc therapy (VMAT) lung (no known deviations), VMAT prostate (TPS inaccuracies), VMAT head-and-neck (EIVD limitations) and intensity modulated radiation therapy (IMRT) breast (large EIVD limitations). The network was trained with virtual non-transit γ-distributions: 60 train/10 validation for the VMAT sites and 30 train/10 validation for IMRT breast. It was hypothesized that in vivo γ-distributions obtained in the presence of plan-specific deviations would differ from those seen during training. For each disease site, the sensitivity of γ-analysis and the network to detect (synthetically introduced) patient-related deviations was compared by receiver operator characteristic analysis. The investigated deviations were patient positioning errors, weight gain or loss, and tumor volume changes. The clinical relevance was illustrated qualitatively with 793 in vivo clinical cases (141 lung, 136 head-and-neck, 209 prostate and 307 breast).

RESULTS

Error detectability of patient-related deviations was better with the network than with γ-analysis. The average area under the curve values over all sites were 0.86 ± 0.12(1SD) and 0.69 ± 0.25(1SD), respectively. Regarding in vivo clinical results, the percentage of cases differently classified by γ-analysis and the network was 1%, 19%, 18% and 64% for lung, head-and-neck, prostate, and breast, respectively. In head-and-neck and breast cases, 45 γ-only alerts were examined, of which 43 were attributed to EPID dose reconstruction limitations. For prostate, all 15 investigated γ-only alerts were due to known TPS inaccuracies. All 59 investigated network alerts were explained by either patient-related deviations or EPID acquisition incidents. Some patient-related deviations detected by the network were not detected by γ-analysis.

CONCLUSIONS

Deep learning-based tools trained to reconstruct underlying patterns of generic deviations in γ-distributions can be used to (i) automatically identify false positives within the set of γ-alerts and (ii) improve the detection of plan-specific deviations, hence minimizing the likelihood of false negatives. The presented method provides clear additional value to the γ-alert management process for large scale EIVD systems.

A convolutional neural network model for EPID-based non-transit dosimetry

PURPOSE

To develop an alternative computational approach for EPID-based non-transit dosimetry using a convolutional neural network model.

METHOD

A U-net followed by a non-trainable layer named True Dose Modulation recovering the spatialized information was developed. The model was trained on 186 Intensity-Modulated Radiation Therapy Step & Shot beams from 36 treatment plans of different tumor locations to convert grayscale portal images into planar absolute dose distributions. Input data were acquired from an amorphous-Silicon Electronic Portal Image Device and a 6 MV X-ray beam. Ground truths were computed from a conventional kernel-based dose algorithm. The model was trained by a two-step learning process and validated through a five-fold cross-validation procedure with sets of training and validation of 80% and 20%, respectively. A study regarding the dependance of the amount of training data was conducted. The performance of the model was evaluated from a quantitative analysis based the ϒ-index, absolute and relative errors computed between the inferred dose distributions and ground truths for six square and 29 clinical beams from seven treatment plans. These results were also compared to those of an existing portal image-to-dose conversion algorithm.

RESULTS

For the clinical beams, averages of ϒ-index and ϒ-passing rate (2%-2mm > 10% D_max ) of 0.24 (±0.04) and 99.29 (±0.70)% were obtained. For the same metrics and criteria, averages of 0.31 (±0.16) and 98.83 (±2.40)% were obtained with the six square beams. Overall, the developed model performed better than the existing analytical method. The study also showed that sufficient model accuracy can be achieved with the amount of training samples used.

CONCLUSION

A deep learning-based model was developed to convert portal images into absolute dose distributions. The accuracy obtained shows that this method has great potential for EPID-based non-transit dosimetry.

Characterization of Gas Electron Multiplier-based detector for external beam radiation therapy dosimetry

PURPOSE

Electronic portal imaging devices (EPIDs) are commonly installed on modern linear accelerators (LINACs) and are convenient for imaging and, potentially, dosimetry. However, owing to their construction with metal and scintillating layers of high atomic number, they exhibit nonwater-equivalent response and oversensitivity to low-energy photons. Therefore, EPIDs are not ideal for dosimetry purposes. Additionally, nonlinearities due to the combined use of scintillators and photodiodes have been reported. Here, an EPID which employs a variable gain Gas Electron Multiplier (GEM) and direct detection of electrons is introduced. To investigate its dosimetric performance, measurements characterizing the novel EPID are performed and compared with measurements from ionization chambers and conventional EPIDs.

METHODS

Linearity, dose rate dependence, field size dependence, off-axis response, and transmission response were measured for all available energy settings (6, 10, 6 MV Flattening Filter Free (FFF) and 10 MVFFF) using three different detector gain settings. Additionally, an evaluation of the ghosting and image lag of the panel was completed. Reference ionization chamber measurements were performed for the off-axis and transmission response and existing data for conventional EPIDs and ionization chambers from equivalent measurements were used for comparison of the field size dependence. Elsewhere, values from the linac monitoring chambers were used.

RESULTS

In the range from 10 to 1000 Monitor Units (MU), the detector was linear within 1% for all combinations of gain settings and energies. The dose rate dependence was also within 1% for all energies and for two out of three gain settings. Regarding field size dependency, the ratio of ionization chamber and panel values was 0.94 and 0.98 for the conventional EPID and GEMini respectively, at 20 × 20 cm² and 10 MV. For 6 MV, 6 MVFFF, and 10MVFFF these ratios were 0.97, 0.98, and 0.99 for the GEMini, and 0.95, 0.97, and 0.97 for the conventional EPID. Similar performance between the GEMini and conventional EPID is observed for field sizes smaller than 10 × 10 cm² . The transmission response was within 5% for all energies for thicknesses up to 30 cm, compared to 10-20% for a conventional EPID. The off-axis response for shifts up to 16 cm was within 1% and 3% for 6 MV and 10 MV, with and without phantom. The rise and fall of the signal from the detector correspond well to monitor chamber measurements indicating little ghosting and image lag, regardless of gain setting.

CONCLUSION

The GEM EPID exhibits dose rate dependence and linearity within 1%, and negligible ghosting and image lag. In this regard, it performs particularly well using 50 and 250 V of gain, and either could be chosen. For higher sensitivity, 250 V is the recommended base gain setting, although other applications may warrant different gains. For most tests performed in this study, the GEM EPID demonstrates a more water-equivalent response than conventional EPIDs making GEMs a viable technology for dosimetry in radiation therapy.

In vivo dosimetry in external beam photon radiotherapy: Requirements and future directions for research, development, and clinical practice

External beam radiotherapy with photon beams is a highly accurate treatment modality, but requires extensive quality assurance programs to confirm that radiation therapy will be or was administered appropriately. In vivo dosimetry (IVD) is an essential element of modern radiation therapy because it provides the ability to catch treatment delivery errors, assist in treatment adaptation, and record the actual dose delivered to the patient. However, for various reasons, its clinical implementation has been slow and limited. The purpose of this report is to stimulate the wider use of IVD for external beam radiotherapy, and in particular of systems using electronic portal imaging devices (EPIDs). After documenting the current IVD methods, this report provides detailed software, hardware and system requirements for in vivo EPID dosimetry systems in order to help in bridging the current vendor-user gap. The report also outlines directions for further development and research. In vivo EPID dosimetry vendors, in collaboration with users across multiple institutions, are requested to improve the understanding and reduce the uncertainties of the system and to help in the determination of optimal action limits for error detection. Finally, the report recommends that automation of all aspects of IVD is needed to help facilitate clinical adoption, including automation of image acquisition, analysis, result interpretation, and reporting/documentation. With the guidance of this report, it is hoped that widespread clinical use of IVD will be significantly accelerated.

The Influence of Acquisition Mode on the Dosimetric Performance of an Amorphous Silicon Electronic Portal Imaging Device

AIMS

This study investigates the impact of cine acquisition mode on the dosimetric characteristics of a Varian aS500 amorphous silicon electronic portal imaging device (a-Si EPID).

MATERIALS AND METHODS

The performance of an a-Si EPID operated in cine mode was assessed and compared to its performance when operated in an integrated mode and dose measurements using an ionization chamber. This study was conducted at different photon energies and the EPID performance was assessed as function of the delivered dose, dose rate, multileaf collimator speed, field size, phantom thickness, and intensity-modulated radiation therapy fields.

RESULTS

The worst nonlinearity was observed at low monitor unit (MU) settings < 100 MU with the highest dose per frame. The nonlinearity of response at a low MU setting was attributed due to the loss of four cine images during each delivery. The EPID response with changing dose rate for 10 MU delivered had similar results to its performance in an integrated mode and ionization chamber. Despite the nonlinearity of response with low MU delivered, EPID performance operated in cine and integrated acquisition modes had comparable responses within 2%.

CONCLUSIONS

For EPID dosimetry application using cine mode, this study recommends the calibration of the EPID images to be undertaken at a large MU. There were no additional corrections that were required when the EPID operated in cine acquisition mode as compared to calibration in integrated mode.

A portal dosimetry dose prediction method based on collapsed cone algorithm using the clinical beam model

PURPOSE

Amorphous silicon electronical portal imaging devices (EPIDs) are widely used for dosimetric measurements in Radiation Therapy. The purpose of this work was to determine if a portal dose prediction method can be utilized for dose map calculations based on the linear accelerator model within a commercial treatment planning system (Pinnacle³ v8.0 m).

METHODS

The method was developed for a 6 MV photon beam on the Varian Clinac 21-EX, at a nominal dose rate of 400 MU/min. The Varian aS1000 EPID was unmounted from the linear accelerator and scanned to acquire CT images of the EPID. The CT images were imported into Pinnacle³ and were used as a quality assurance phantom to calculate dose on the EPID setup at a source to detector distance of 105 cm. The best match of the dose distributions was obtained considering the image plane located at 106 cm from the source to detector plane. The EPID was calibrated according to the manufacturer procedure and corrections were made for output factors. Arm-backscattering effect, based on profile correction curves, has been introduced. Five low-modulated and three high-modulated clinical planned treatments were predicted and measured with the method presented here and with MatriXX (IBA Dosimetry, Schwarzenbruck, Germany).

RESULTS

A portal dose prediction method based on Pinnacle³ was developed without modifying the commissioned parameters of the model in use in the clinic. CT images of the EPID were acquired and used as a quality assurance phantom. The CT images indicated a mean density of 1.16 g/cm³ for the sensitive area of the EPID. Output factor measured with the EPID were lower for small fields and larger for larger fields (beyond 10 × 10 cm² ). Arm-backscatter correction showed a better agreement at the target side of the EPID. Analysis of Gamma index comparison (3%, 3 mm) indicated a minimum of 97.4% pass rate for low modulated and 98.3% for high modulated treatments. Pass rates were similar for MatriXX measurements.

CONCLUSIONS

The method developed here can be easily implemented into clinic, as neither additional modeling of the clinical energy nor an independent image prediction algorithm are necessary. The main advantage of this method is that portal dose prediction is calculated with the same algorithm and beam model used for patient dose distribution calculation. This method was independently validated with an ionization chamber matrix.

Evaluation of the TrueBeam machine performance check (MPC) beam constancy checks for flattened and flattening filter-free (FFF) photon beams

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 MPC beam performance tests against current daily quality assurance (QA) methods, to compare MPC performance against more accurate monthly QA tests and to test the sensitivity of MPC to changes in beam performance. The MPC beam constancy checks test the beam output, uniformity, and beam center against the user defined baseline. MPC was run daily over a period of 5 months (n = 115) in parallel with the Daily QA3 device. Additionally, IC Profiler, in‐house EPID tests, and ion chamber measurements were performed biweekly and results presented in a form directly comparable to MPC. The sensitivity of MPC was investigated using controlled adjustments of output, beam angle, and beam position steering. Over the period, MPC output agreed with ion chamber to within 0.6%. For an output adjustment of 1.2%, MPC was found to agree with ion chamber to within 0.17%. MPC beam center was found to agree with the in‐house EPID method within 0.1 mm. A focal spot position adjustment of 0.4 mm (at isocenter) was measured with MPC beam center to within 0.01 mm. An average systematic offset of 0.5% was measured in the MPC uniformity and agreement of MPC uniformity with symmetry measurements was found to be within 0.9% for all beams. MPC uniformity detected a change in beam symmetry of 1.5% to within 0.3% and 0.9% of IC Profiler for flattened and FFF beams, respectively.

Use of electronic portal imaging devices for electron treatment verification

This study aims to help broaden the use of electronic portal imaging devices (EPIDs) for pre-treatment patient positioning verification, from photon-beam radiotherapy to photon- and electron-beam radiotherapy, by proposing and testing a method for acquiring clinically-useful EPID images of patient anatomy using electron beams, with a view to enabling and encouraging further research in this area. EPID images used in this study were acquired using all available beams from a linac configured to deliver electron beams with nominal energies of 6, 9, 12, 16 and 20 MeV, as well as photon beams with nominal energies of 6 and 10 MV. A widely-available heterogeneous, approximately-humanoid, thorax phantom was used, to provide an indication of the contrast and noise produced when imaging different types of tissue with comparatively realistic thicknesses. The acquired images were automatically calibrated, corrected for the effects of variations in the sensitivity of individual photodiodes, using a flood field image. For electron beam imaging, flood field EPID calibration images were acquired with and without the placement of blocks of water-equivalent plastic (with thicknesses approximately equal to the practical range of electrons in the plastic) placed upstream of the EPID, to filter out the primary electron beam, leaving only the bremsstrahlung photon signal. While the electron beam images acquired using a standard (unfiltered) flood field calibration were observed to be noisy and difficult to interpret, the electron beam images acquired using the filtered flood field calibration showed tissues and bony anatomy with levels of contrast and noise that were similar to the contrast and noise levels seen in the clinically acceptable photon beam EPID images. The best electron beam imaging results (highest contrast, signal-to-noise and contrast-to-noise ratios) were achieved when the images were acquired using the higher energy electron beams (16 and 20 MeV) when the EPID was calibrated using an intermediate (12 MeV) electron beam energy. These results demonstrate the feasibility of acquiring clinically-useful EPID images of patient anatomy using electron beams and suggest important avenues for future investigation, thus enabling and encouraging further research in this area. There is manifest potential for the EPID imaging method proposed in this work to lead to the clinical use of electron beam imaging for geometric verification of electron treatments in the future.

Portal dosimetry in wedged beams

Portal dosimetry using electronic portal imaging devices (EPIDs) is often applied to verify high‐energy photon beam treatments. Due to the change in photon energy spectrum, the resulting dose values are, however, not very accurate in the case of wedged beams if the pixel‐to‐dose conversion for the situation without wedge is used. A possible solution would be to consider a wedged beam as another photon beam quality requiring separate beam modeling of the dose calculation algorithm. The aim of this study was to investigate a more practical solution: to make aSi EPID‐based dosimetry models also applicable for wedged beams without an extra commissioning effort of the parameters of the model. For this purpose two energy‐dependent wedge multiplication factors have been introduced to be applied for portal images taken with and without a patient/phantom in the beam. These wedge multiplication factors were derived from EPID and ionization chamber measurements at the EPID level for wedged and nonwedged beams, both with and without a polystyrene slab phantom in the beam. This method was verified for an EPID dosimetry model used for wedged beams at three photon beam energies (6, 10, and 18 MV) by comparing dose values reconstructed in a phantom with data provided by a treatment planning system (TPS), as a function of field size, depth, and off‐axis distance. Generally good agreement, within 2%, was observed for depths between dose maximum and 15 cm. Applying the new model to EPID dose measurements performed during ten breast cancer patient treatments with wedged 6 MV photon beams showed that the average isocenter underdosage of 5.3% was reduced to 0.4%. Gamma‐evaluation (global 3%/3 mm) of these in vivo data showed an increase in percentage of points with γ≤1 from 60.2% to 87.4%, while γmean reduced from 1.01 to 0.55. It can be concluded that, for wedged beams, the multiplication of EPID pixel values with an energy‐dependent correction factor provides good agreement between dose values determined by an EPID and a TPS, indicating the usefulness of such a practical solution.

PACS numbers: 87.55.km, 87.55.kd, 87.55.Qr, 87.56a.ng

Two-dimensional in vivo dose verification using portal imaging and correlation ratios

The electronic portal imaging device (EPID) has the potential to be used for in vivo dosimetry during radiation therapy as an additional dose delivery check. In this study we have extended a method developed by A. Piermattei and colleagues in 2006 that made use of EPID transit images (acquired during treatment) to calculate dose in the isocenter point. The extension allows calculation of two-dimensional dose maps of the entire radiation field at the depth of isocenter. We quantified the variability of the ratio of EPID signal to dose in the isocenter plane in Solid Water phantoms of various thicknesses and with various field sizes, and designed a field edge dose calculation correction. To validate the method, we designed three realistic conventional radiation therapy treatment plans on a thorax and head anthropomorphic phantom (whole brain, brain primary, lung tumor). Using CT data, EPID transit images, EPID signal-to-dose correlation, and our edge correction, we calculated dose in the isocenter plane and compared it with the treatment planning system's prediction. Gamma evaluation (3%, 3 mm) showed good agreement (Pγ<1 ≥ 96.5%) for all fields of the whole brain and brain primary plans. In the presence of lung, however, our algorithm overestimated dose by 7%-9%. This 2D EPID-based in vivo dosimetry method can be used for posttreatment dose verification, thereby improving the safety and quality of patient treatments. With future work, it may be extended to measure dose in real time and to prevent harmful delivery errors.

2D lag and signal nonlinearity correction in an amorphous silicon EPID and their impact on pretreatment dosimetric verification

PURPOSE

This investigation provides measurements of signal lag and nonlinearity separately for the Varian aS500 electronic portal imaging device (EPID), and an algorithm to correct for these effects in 2D; their potential impact on intensity modulated radiation therapy (IMRT) verification is also investigated. The authors quantify lag, as a function of both delivered monitor units (MU) and time, by using a range of MUs delivered at a clinically used rate of 400 MU∕min. Explicit cumulative lag curves are thus determined for a range of MUs and times between the end of irradiation and the end of image acquisition. Signal nonlinearity is also investigated as a function of total MUs delivered. The family of cumulative lag curves and signal nonlinearity are then used to determine their effects on dynamic multileaf collimator (MLC) (IMRT) deliveries, and to correct for theses effects in 2D.

METHODS

Images acquired with an aS500 EPID and Varis Portal-Vision software were used to quantify detector lag and signal-nonlinearity. For the signal lag investigation, Portal-Vision's service monitor was used to acquire EPID images at a rate of 8 frames/s. The images were acquired during irradiation and 66 s thereafter, by inhibiting the M-holdoff-In signal of the Linac for a range of 4.5-198.5 MUs. Relative cumulative lag was calculated by integrating the EPID signal for a time after beam-off, and normalizing this to the integrated EPID signal accumulated during radiation. Signal nonlinearity was studied by acquiring 10 × 10 cm(2) open-field EPID images in "integrated image" mode for a range of 2-500 MUs, and normalized to the 100 MU case. All data were incorporated into in-house written software to create a 2D correction map for these effects, using the field's MLC file and a field-specific calculated 2D "time-map," which keeps track of the time elapsed from the last fluence delivered at each given point in the image to the end of the beam delivery.

RESULTS

Relative cumulative lag curves reveal that the lag alone can deviate the EPID's perceived dose by as large as 6% (1 MU delivery, 60 s postirradiation). For signal nonlinearity relative to 100 MU, EPID signals per MU of 0.84 and 1.01 were observed for 2 and 500 MUs, respectively. Correction maps were applied to a 1 cm sweeping-window 14 × 14 cm(2) field and clinical head-and-neck IMRT field. A mean correction of 1.028 was implemented in the head-and-neck field, which significantly reduced lag-related asymmetries in the EPID images, and restored linearity to the EPID imager's dose response. Corrections made to the sweeping-field showed good agreement with the treatment planning system-predicted field, yielding an average percent difference of 0.05% ± 0.91%, compared to the -1.32% ± 1.02% before corrections, or 1.75% ± 1.04% when only a signal nonlinearity correction is made.

CONCLUSIONS

Lag and signal-nonlinearity have been quantified for an aS500 EPID imager, and an effective 2D correction method has been developed which effectively removes nonlinearity and lag effects. Both of these effects were shown to negatively impact IMRT verifications. Especially fields that involve prolonged irradiation and small overall MUs should be corrected for in 2D.

Monte Carlo-based adaptive EPID dose kernel accounting for different field size responses of imagers

The aim of this study is to present an efficient method to generate imager-specific Monte Carlo (MC)-based dose kernels for amorphous silicon-based electronic portal image device dose prediction and determine the effective backscattering thicknesses for such imagers. EPID field size-dependent responses were measured for five matched Varian accelerators from three institutions with 6 MV beams at the source to detector distance (SDD) of 105 cm. For two imagers, measurements were made with and without the imager mounted on the robotic supporting arm. Monoenergetic energy deposition kernels with 0-2.5 cm of water backscattering thicknesses were simultaneously computed by MC to a high precision. For each imager, the backscattering thickness required to match measured field size responses was determined. The monoenergetic kernel method was validated by comparing measured and predicted field size responses at 150 cm SDD, 10 x 10 cm2 multileaf collimator (MLC) sliding window fields created with 5, 10, 20, and 50 mm gaps, and a head-and-neck (H&N) intensity modulated radiation therapy (IMRT) patient field. Field size responses for the five different imagers deviated by up to 1.3%. When imagers were removed from the robotic arms, response deviations were reduced to 0.2%. All imager field size responses were captured by using between 1.0 and 1.6 cm backscatter. The predicted field size responses by the imager-specific kernels matched measurements for all involved imagers with the maximal deviation of 0.34%. The maximal deviation between the predicted and measured field size responses at 150 cm SDD is 0.39%. The maximal deviation between the predicted and measured MLC sliding window fields is 0.39%. For the patient field, gamma analysis yielded that 99.0% of the pixels have gamma < 1 by the 2%, 2 mm criteria with a 3% dose threshold. Tunable imager-specific kernels can be generated rapidly and accurately in a single MC simulation. The resultant kernels are imager position independent and are able to predict fields with varied incident energy spectra and a H&N IMRT patient field. The proposed adaptive EPID dose kernel method provides the necessary infrastructure to build reliable and accurate portal dosimetry systems.

Measurement-based Monte Carlo dose calculation system for IMRT pretreatment and on-line transit dose verifications

The aim of this study was to develop a dose simulation system based on portal dosimetry measurements and the BEAM Monte Carlo code for intensity-modulated (IM) radiotherapy dose verification. This measurement-based Monte Carlo (MBMC) system can perform, within one systematic calculation, both pretreatment and on-line transit dose verifications. BEAMnrc and DOSXYZnrc 2006 were used to simulate radiation transport from the treatment head, through the patient, to the plane of the aS500 electronic portal imaging device (EPID). In order to represent the nonuniform fluence distribution of an IM field within the MBMC simulation, an EPID-measured efficiency map was used to redistribute particle weightings of the simulated phase space distribution of an open field at a plane above a patient/phantom. This efficiency map was obtained by dividing the measured energy fluence distribution of an IM field to that of an open field at the EPID plane. The simulated dose distribution at the midplane of a homogeneous polystyrene phantom was compared to the corresponding distribution obtained from the Eclipse treatment planning system (TPS) for pretreatment verification. It also generated a simulated transit dose distribution to serve as the on-line verification reference for comparison to that measured by the EPID. Two head-and-neck (NPC1 and NPC2) and one prostate cancer fields were tested in this study. To validate the accuracy of the MBMC system, film dosimetry was performed and served as the dosimetry reference. Excellent agreement between the film dosimetry and the MBMC simulation was obtained for pretreatment verification. For all three cases tested, gamma evaluation with 3%/3 mm criteria showed a high pass percentage (> 99.7%) within the area in which the dose was greater than 30% of the maximum dose. In contrast to the TPS, the MBMC system was able to preserve multileaf collimator delivery effects such as the tongue-and-groove effect and interleaf leakage. In the NPC1 field, the TPS showed 16.5% overdose due to the tongue-and-groove effect and 14.6% overdose due to improper leaf stepping. Similarly, in the NPC2 field, the TPS showed 14.1% overdose due to the tongue-and-groove effect and 8.9% overdose due to improper leaf stepping. In the prostate cancer field, the TPS showed 6.8% overdose due to improper leaf stepping. No tongue-and-groove effect was observed for this field. For transit dose verification, agreements among the EPID measurement, the film dosimetry, and the MBMC system were also excellent with a minimum gamma pass percentage of 99.6%.

An empirical model of electronic portal imager response implemented within a commercial treatment planning system for verification of intensity-modulated radiation therapy fields

Quality assurance (QA) of an intensity‐modulated radiation therapy (IMRT) plan is more complex than that of a conventional plan. To improve the efficiency of QA, electronic portal imaging devices (EPIDs) can be used. The major objective of the present work was to use a commercial treatment planning system to model EPID response for the purpose of pre‐treatment IMRT dose verification.

Images were acquired with an amorphous silicon flat panel portal imager (aS500: Varian Medical Systems, Palo Alto, CA) directly irradiated with a 6‐MV photon beam from a Clinac 21EX linear accelerator (Varian Medical Systems). Portal images were acquired for a variety of rectangular fields, from which profiles and relative output factors were extracted. A dedicated machine model was created using the physics tools of the Pinnacle³ (Philips Medical Systems, Madison, WI) treatment planning system to model the data. Starting with the known photon spectrum and assuming an effective depth of 7 cm, machine model parameters were adjusted to best fit measured profile and output factors. The machine parameters of a second model, which assumed a 0.8 MeV monoenergetic photon spectrum and an effective depth in water of 3 cm, were also optimized. The second EPID machine model was used to calculate planar dose maps of simple geometric IMRT fields as well as a 9‐field IMRT plan developed for clinical trials credentialing purposes.

The choice of energy and depth for an EPID machine model influenced the best achievable fit of the optimized machine model to the measured data. When both energy and depth were reduced by a significant amount, a better overall fit was achieved. In either case, the secondary source size and strength could be adjusted to give reasonable agreement with measured data. The gamma evaluation method was used to compare planar dose maps calculated using the second EPID machine model with the EPID images of small IMRT fields. In each case, more than 95% of points fell within 3% of the maximum dose or 3 mm distance to agreement. These results are slightly poorer than those obtained using an ion chamber array, which confirms agreement to within 2% of the maximum dose or 2 mm distance to agreement for all points within these fields.

PACS numbers: 87.55Qr, 87.56.Fc

A literature review of electronic portal imaging for radiotherapy dosimetry

Electronic portal imaging devices (EPIDs) have been the preferred tools for verification of patient positioning for radiotherapy in recent decades. Since EPID images contain dose information, many groups have investigated their use for radiotherapy dose measurement. With the introduction of the amorphous-silicon EPIDs, the interest in EPID dosimetry has been accelerated because of the favourable characteristics such as fast image acquisition, high resolution, digital format, and potential for in vivo measurements and 3D dose verification. As a result, the number of publications dealing with EPID dosimetry has increased considerably over the past approximately 15 years. The purpose of this paper was to review the information provided in these publications. Information available in the literature included dosimetric characteristics and calibration procedures of various types of EPIDs, strategies to use EPIDs for dose verification, clinical approaches to EPID dosimetry, ranging from point dose to full 3D dose distribution verification, and current clinical experience. Quality control of a linear accelerator, pre-treatment dose verification and in vivo dosimetry using EPIDs are now routinely used in a growing number of clinics. The use of EPIDs for dosimetry purposes has matured and is now a reliable and accurate dose verification method that can be used in a large number of situations. Methods to integrate 3D in vivo dosimetry and image-guided radiotherapy (IGRT) procedures, such as the use of kV or MV cone-beam CT, are under development. It has been shown that EPID dosimetry can play an integral role in the total chain of verification procedures that are implemented in a radiotherapy department. It provides a safety net for simple to advanced treatments, as well as a full account of the dose delivered. Despite these favourable characteristics and the vast range of publications on the subject, there is still a lack of commercially available solutions for EPID dosimetry. As strategies evolve and commercial products become available, EPID dosimetry has the potential to become an accurate and efficient means of large-scale patient-specific IMRT dose verification for any radiotherapy department.

The impact of MLC transmitted radiation on EPID dosimetry for dynamic MLC beams

The purpose of this study was to experimentally quantify the change in response of an amorphous silicon (a-Si) electronic portal imaging device (EPID) to dynamic multileaf collimator (dMLC) beams with varying MLC-transmitted dose components and incorporate the response into a commercial treatment planning system (TPS) EPID prediction model. A combination of uniform intensity dMLC beams and static beams were designed to quantify the effect of MLC transmission on EPID response at the central axis of 10 x 10 cm2 beams, at off-axis positions using wide dMLC beam profiles, and at different field sizes. The EPID response to MLC transmitted radiation was 0.79 +/- 0.02 of the response to open beam radiation at the central axis of a 10 x 10 cm2 field. The EPID response to MLC transmitted radiation was further reduced relative to the open beam response with off-axis distance. The EPID response was more sensitive to field size changes for MLC transmitted radiation compared to open beam radiation by a factor of up to 1.17 at large field sizes. The results were used to create EPID response correction factors as a function of the fraction of MLC transmitted radiation, off-axis distance, and field size. Software was developed to apply the correction factors to each pixel in the TPS predicted EPID image. The corrected images agreed more closely with the measured EPID images in areas of intensity modulated fields with a large fraction of MLC transmission and, as a result the accuracy of portal dosimetry with a-Si EPIDs can be improved. Further investigation into the detector response function and the radiation source model are required to achieve improvements in accuracy for the general case.

Testing the GlaaS algorithm for dose measurements on low- and high-energy photon beams using an amorphous silicon portal imager

The GLAaS algorithm for pretreatment intensity modulation radiation therapy absolute dose verification based on the use of amorphous silicon detectors, as described in Nicolini et al. [G. Nicolini, A. Fogliata, E. Vanetti, A. Clivio, and L. Cozzi, Med. Phys. 33, 2839-2851 (2006)], was tested under a variety of experimental conditions to investigate its robustness, the possibility of using it in different clinics and its performance. GLAaS was therefore tested on a low-energy Varian Clinac (6 MV) equipped with an amorphous silicon Portal Vision PV-aS500 with electronic readout IAS2 and on a high-energy Clinac (6 and 15 MV) equipped with a PV-aS1000 and IAS3 electronics. Tests were performed for three calibration conditions: A: adding buildup on the top of the cassette such that SDD-SSD = d(max) and comparing measurements with corresponding doses computed at d(max), B: without adding any buildup on the top of the cassette and considering only the intrinsic water-equivalent thickness of the electronic portal imaging devices device (0.8 cm), and C: without adding any buildup on the top of the cassette but comparing measurements against doses computed at d(max). This procedure is similar to that usually applied when in vivo dosimetry is performed with solid state diodes without sufficient buildup material. Quantitatively, the gamma index (gamma), as described by Low et al. [D. A. Low, W. B. Harms, S. Mutic, and J. A. Purdy, Med. Phys. 25, 656-660 (1998)], was assessed. The gamma index was computed for a distance to agreement (DTA) of 3 mm. The dose difference deltaD was considered as 2%, 3%, and 4%. As a measure of the quality of results, the fraction of field area with gamma larger than 1 (%FA) was scored. Results over a set of 50 test samples (including fields from head and neck, breast, prostate, anal canal, and brain cases) and from the long-term routine usage, demonstrated the robustness and stability of GLAaS. In general, the mean values of %FA remain below 3% for deltaD equal or larger than 3%, while they are slightly larger for deltaD = 2% with %FA in the range from 3% to 8%. Since its introduction in routine practice, 1453 fields have been verified with GLAaS at the authors' institute (6 MV beam). Using a DTA of 3 mm and a deltaD of 4% the authors obtained %FA = 0.9 +/- 1.1 for the entire data set while, stratifying according to the dose calculation algorithm, they observed: %FA = 0.7 +/- 0.9 for fields computed with the analytical anisotropic algorithm and %FA = 2.4 +/- 1.3 for pencil-beam based fields with a statistically significant difference between the two groups. If data are stratified according to field splitting, they observed %FA = 0.8 +/- 1.0 for split fields and 1.0 +/- 1.2 for nonsplit fields without any significant difference.

Experimental investigation of the response of an amorphous silicon EPID to intensity modulated radiotherapy beams

The aim of this work was to experimentally determine the difference in response of an amorphous silicon (a-Si) electronic portal imaging device (EPID) to the open and multileaf collimator (MLC) transmitted beam components of intensity modulated radiation therapy (IMRT) beams. EPID dose response curves were measured for open and MLC transmitted (MLCtr) 10 x 10 cm2 beams at central axis and with off axis distance using a shifting field technique. The EPID signal was obtained by replacing the flood-field correction with a pixel sensitivity variation matrix correction. This signal, which includes energy-dependent response, was then compared to ion-chamber measurements. An EPID calibration method to remove the effect of beam energy variations on EPID response was developed for IMRT beams. This method uses the component of open and MLCtr fluence to an EPID pixel calculated from the MLC delivery file and applies separate radially dependent calibration factors for each component. The calibration procedure does not correct for scatter differences between ion chamber in water measurements and EPID response; these must be accounted for separately with a kernel-based approach or similar method. The EPID response at central axis for the open beam was found to be 1.28 +/- 0.03 of the response for the MLCtr beam, with the ratio increasing to 1.39 at 12.5 cm off axis. The EPID response to MLCtr radiation did not change with off-axis distance. Filtering the beam with copper plates to reduce the beam energy difference between open and MLCtr beams was investigated; however, these were not effective at reducing EPID response differences. The change in EPID response for uniform sliding window IMRT beams with MLCtr dose components from 0.3% to 69% was predicted to within 2.3% using the separate EPID response calibration factors for each dose component. A clinical IMRT image calibrated with this method differed by nearly 30% in high MLCtr regions from an image calibrated with an open beam calibration factor only. Accounting for the difference in EPID response to open and MLCtr radiation should improve IMRT dosimetry with a-Si EPIDs.

Off-axis dose response characteristics of an amorphous silicon electronic portal imaging device

Amorphous silicon (a-Si) electronic portal imaging devices (EPIDs) have typically been calibrated to dose at central axis (CAX). Division of acquired images by the flood-field (FF) image that corrects for pixel sensitivity variation as well as open field energy-dependent off-axis response variation should result in a flat EPID response over the entire matrix for the same field size. While the beam profile can be reintroduced to the image by an additional correction matrix, the CAX EPID response to dose calibration factor is assumed to apply to all pixels in the detector. The aim of this work was to investigate the dose response of the Varian aS500 amorphous silicon detector across the entire detector area. First it was established that the EPID response across the panel became stable (within approximately 0.2%) for MU settings greater than approximately 200 MU. The EPID was then FF calibrated with a high MU setting of approximately 400 for all subsequent experiments. Whole detector images with varying MU settings from 2-500 were then acquired for two dose rates (300 and 600 MU/min) for 6 MV photons for two EPIDs. The FF corrected EPID response was approximately flat or uniform across the detector for greater than 100 MU delivered (within 0.5%). However, the off-axis EPID response was greater than the CAX response for small MU irradiations, giving a raised EPID profile. Up to 5% increase in response at 20 cm off-axis compared to CAX was found for very small MU settings for one EPID, while it was within 2% for the second (newer) EPID. Off-axis response nonuniformities attributed to detector damage were also found for the older EPID. Similar results were obtained with the EPID at 18 MV energy and operating in asynchronous mode (acquisition not synchronized with beam pulses), however the profiles were flatter and more irregular for the small MU irradiations. By moving the detector laterally and repeating the experiments, the increase in response off-axis was found to depend on the pixel position relative to the beam CAX. When the beam was heavily filtered by a phantom the off-axis response variation was reduced markedly to within 0.5% for all MU settings. Independent measurements of off-axis point doses with ion chamber did not show any change in off-axis factor with MUs. Measurements of beam quality (TMR20-10) for MU settings of 2, 5, and 100 at central axis and at 15 cm off-axis could not explain the effect. The response change is unlikely to be significant for clinical IMRT verification with this imaging/acclerator system where MUs are of the order of 100-300, provided the detector does not exhibit radiation damage artifacts.

Verification of helical tomotherapy delivery using autoassociative kernel regressiona)

Quality assurance (QA) is a topic of major concern in the field of intensity modulated radiation therapy (IMRT). The standard of practice for IMRT is to perform QA testing for individual patients to verify that the dose distribution will be delivered to the patient. The purpose of this study was to develop a new technique that could eventually be used to automatically evaluate helical tomotherapy treatments during delivery using exit detector data. This technique uses an autoassociative kernel regression (AAKR) model to detect errors in tomotherapy delivery. AAKR is a novel nonparametric model that is known to predict a group of correct sensor values when supplied a group of sensor values that is usually corrupted or contains faults such as machine failure. This modeling scheme is especially suited for the problem of monitoring the fluence values found in the exit detector data because it is able to learn the complex detector data relationships. This scheme still applies when detector data are summed over many frames with a low temporal resolution and a variable beam attenuation resulting from patient movement. Delivery sequences from three archived patients (prostate, lung, and head and neck) were used in this study. Each delivery sequence was modified by reducing the opening time for random individual multileaf collimator (MLC) leaves by random amounts. The errof and error-free treatments were delivered with different phantoms in the path of the beam. Multiple autoassociative kernel regression (AAKR) models were developed and tested by the investigators using combinations of the stored exit detector data sets from each delivery. The models proved robust and were able to predict the correct or error-free values for a projection, which had a single MLC leaf decrease its opening time by less than 10 msec. The model also was able to determine machine output errors. The average uncertainty value for the unfaulted projections ranged from 0.4% to 1.8% of the detector signal. The low model uncertainty indicates that the AAKR model is extremely accurate in its predictions and also suggests that the model may be able to detect errors that cause the fluence to change by less than 2%. However, additional evaluation of the AAKR technique is needed to determine the minimum detectable error threshold from the compressed helical tomotherapy detector data. Further research also needs to explore applying this technique to electronic portal imaging detector data.

Amorphous silicon EPID calibration for dosimetric applications: comparison of a method based on Monte Carlo prediction of response with existing techniques

For EPID dosimetry, the calibration should ensure that all pixels have a similar response to a given irradiation. A calibration method (MC), using an analytical fit of a Monte Carlo simulated flood field EPID image to correct for the flood field image pixel intensity shape, was proposed. It was compared with the standard flood field calibration (FF), with the use of a water slab placed in the beam to flatten the flood field (WS) and with a multiple field calibration where the EPID was irradiated with a fixed 10x10 field for 16 different positions (MF). The EPID was used in its normal configuration (clinical setup) and with an additional 3 mm copper slab (modified setup). Beam asymmetry measured with a diode array was taken into account in MC and WS methods. For both setups, the MC method provided pixel sensitivity values within 3% of those obtained with the MF and WS methods (mean difference<1%, standard deviation<2%). The difference of pixel sensitivity between MC and FF methods was up to 12.2% (clinical setup) and 11.8% (modified setup). MC calibration provided images of open fields (5x5 to 20x20 cm2) and IMRT fields to within 3% of that obtained with WS and MF calibrations while differences with images calibrated with the FF method for fields larger than 10x10 cm2 were up to 8%. MC, WS and MF methods all provided a major improvement on the FF method. Advantages and drawbacks of each method were reviewed.

Investigation of an amorphous silicon EPID for measurement and quality assurance of enhanced dynamic wedge

This work assessed the dosimetric performance of an amorphous silicon electronic portal imaging device (EPID) for measurement and quality assurance of enhanced dynamic wedge (EDW) profile and wedge factor. EPID measurements of EDW profiles were corrected for pixel sensitivity variation and spectral over-response relative to ion chamber and compared to ion-chamber and diode-array measurements. The dependence of EPID measurements on wedge direction and source to EPID distance was assessed. The long-term stability was investigated by weekly measurement of EDW profiles and wedge factors over a seven month interval. An empirical correction method was developed to improve EPID profile agreement with diode-array measurements. The EPID profiles differed from conventional measurements by up to 5%. Backscatter from the EPID housing was also found to affect measurements by up to 4%, resulting in changes in EPID measured profile with the direction of the moving jaw and source to EPID distance. EPID profile measurements varied by a maximum of 0.3% (1 SD) within the umbra, and wedge factors varied by 0.3% (1 SD) over the seven month interval. The correction function improved agreement between EPID and diode array to within 2% for all wedge angles and energies. Due to the ease of use and reproducibility of the EPID-measured EDW profiles the device is highly suited to regular measurement of EDW.

Monte Carlo modelling of a-Si EPID response: The effect of spectral variations with field size and position

This study focused on predicting the electronic portal imaging device (EPID) image of intensity modulated radiation treatment (IMRT) fields in the absence of attenuation material in the beam with Monte Carlo methods. As IMRT treatments consist of a series of segments of various sizes that are not always delivered on the central axis, large spectral variations may be observed between the segments. The effect of these spectral variations on the EPID response was studied with fields of various sizes and off-axis positions. A detailed description of the EPID was implemented in a Monte Carlo model. The EPID model was validated by comparing the EPID output factors for field sizes between 1 x 1 and 26 x 26 cm2 at the isocenter. The Monte Carlo simulations agreed with the measurements to within 1.5%. The Monte Carlo model succeeded in predicting the EPID response at the center of the fields of various sizes and offsets to within 1% of the measurements. Large variations (up to 29%) of the EPID response were observed between the various offsets. The EPID response increased with field size and with field offset for most cases. The Monte Carlo model was then used to predict the image of a simple test IMRT field delivered on the beam axis and with an offset. A variation of EPID response up to 28% was found between the on- and off-axis delivery. Finally, two clinical IMRT fields were simulated and compared to the measurements. For all IMRT fields, simulations and measurements agreed within 3%-0.2 cm for 98% of the pixels. The spectral variations were quantified by extracting from the spectra at the center of the fields the total photon yield (Ytotal), the photon yield below 1 MeV (Ylow), and the percentage of photons below 1 MeV (Plow). For the studied cases, a correlation was shown between the EPID response variation and Ytotal, Ylow, and Plow.

Using fluence separation to account for energy spectra dependence in computing dosimetric a-Si EPID images for IMRT fields

This study develops a method to improve the dosimetric accuracy of computed images for an amorphous silicon flat-panel imager. Radially dependent kernels derived from Monte Carlo simulations are convolved with the treatment-planning system's energy fluence. Multileaf collimator (MLC) beam hardening is accounted for by having separate kernels for open and blocked portions of MLC fields. Field-size-dependent output factors are used to account for the field-size dependence of scatter within the imager. Gamma analysis was used to evaluate open and sliding window test fields and intensity modulated patient fields. For each tested field, at least 99.6% of the points had gamma < 1 with a 3%, 3-mm criteria. With a 2%, 2-mm criteria, between 81% and 100% of points had gamma < 1. Patient intensity modulated test fields had 94%-100% of the points with gamma < 1 with a 2%, 2-mm criteria for all six fields tested. This study demonstrates that including the dependencies of kernel and fluence on radius and beam hardening in the convolution improves its accuracy compared with the use of radial and beam-hardening independent kernels; it also demonstrates that the resultant accuracy of the convolution method is sufficient for pretreatment, intensity modulated patient field verification.

Pretreatment verification of IMRT absolute dose distributions using a commercial a-Si EPID

A commercial amorphous silicon electronic portal imaging device (EPID) has been studied to investigate its potential in the field of pretreatment verifications of step and shoot, intensity modulated radiation therapy (IMRT), 6 MV photon beams. The EPID was calibrated to measure absolute exit dose in a water-equivalent phantom at patient level, following an experimental approach, which does not require sophisticated calculation algorithms. The procedure presented was specifically intended to replace the time-consuming in-phantom film dosimetry. The dosimetric response was characterized on the central axis in terms of stability, linearity, and pulse repetition frequency dependence. The a-Si EPID demonstrated a good linearity with dose (within 2% from 1 monitor unit), which represent a prerequisite for the application in IMRT. A series of measurements, in which phantom thickness, air gap between the phantom and the EPID, field size and position of measurement of dose in the phantom (entrance or exit) varied, was performed to find the optimal calibration conditions, for which the field size dependence is minimized. In these conditions (20 cm phantom thickness, 56 cm air gap, exit dose measured at the isocenter), the introduction of a filter for the low-energy scattered radiation allowed us to define a universal calibration factor, independent of field size. The off-axis extension of the dose calibration was performed by applying a radial correction for the beam profile, distorted due to the standard flood field calibration of the device. For the acquisition of IMRT fields, it was necessary to employ home-made software and a specific procedure. This method was applied for the measurement of the dose distributions for 15 clinical IMRT fields. The agreement between the dose distributions, quantified by the gamma index, was found, on average, in 97.6% and 98.3% of the analyzed points for EPID versus TPS and for EPID versus FILM, respectively, thus suggesting a great potential of this EPID for IMRT dosimetric applications.

Monte Carlo and Lambertian light guide models of the light output from scintillation crystals at megavoltage energies

A new model of the light output from single-crystal scintillators in megavoltage energy x-ray beams has been developed, based on the concept of a Lambertian light guide model (LLG). This was evaluated in comparison with a Monte Carlo (MC) model of optical photon transport, previously developed and reported in the literature, which was used as a gold standard. The LLG model was developed to enable optimization of scintillator detector design. In both models the dose deposition and light propagation were decoupled, the scintillators were cuboids, split into a series of cells as a function of depth, with Lambertian side and entrance faces, and a specular exit face. The signal in a sensor placed 1 and 1000 mm beyond the exit face was calculated. Cesium iodide (CSI) crystals of 1.5 and 3 mm square cross section and 1, 5, and 10 mm depth were modeled. Both models were also used to determine detector signal and optical gain factor as a function of CsI scintillator thickness, from 2 to 10 mm. Results showed a variation in light output with position of dose deposition of a factor of up to approximately 5, for long, thin scintillators (such as 10 X 1.5 x 1.5 mm3). For short, fat scintillators (such as 1 X 3 X 3 mm3) the light output was more uniform with depth. MC and LLG generally agreed to within 5%. Results for a sensor distance of 1 mm showed an increase in light output the closer the light originates to the exit face, while a distance of 1000 mm showed a decrease in light output the closer the light originates to the exit face. For a sensor distance of 1 mm, the ratio of signal for a 10 mm scintillator to that for a 2 mm scintillator was 1.98, whereas for the 1000 mm distance the ratio was 3.00. The ratio of quantum efficiency (QE) between 10 and 2 mm thicknesses was 4.62. We conclude that these models may be used for detector optimization, with the light guide model suitable for parametric study.