Conjoint Professor Peter Greer

Purpose: Reducing margins during treatment planning to decrease dose to healthy organs surrounding the prostate can risk inadequate treatment of subclinical disease. This study aimed to investigate whether lack of dose to subclinical disease is associated with increased disease progression by using high-quality prostate radiation therapy clinical trial data to identify anatomically localized regions where dose variation is associated with prostate-specific antigen progression (PSAP). Methods and Materials: Planned dose distributions for 683 patients of the Trans-Tasman Radiation Oncology Group 03.04 Randomized Androgen Deprivation and Radiotherapy (RADAR) trial were deformably registered onto a single exemplar computed tomography data set. These were divided into high-risk and intermediate-risk subgroups for analysis. Three independent voxel-based statistical tests, using permutation testing, Cox regression modeling, and least absolute shrinkage selection operator feature selection, were applied to identify regions where dose variation was associated with PSAP. Results from the intermediate-risk RADAR subgroup were externally validated by registering dose distributions from the RT01 (n = 388) and Conventional or Hypofractionated High Dose Intensity Modulated Radiotherapy for Prostate Cancer Trial (CHHiP) (n = 253) trials onto the same exemplar and repeating the tests on each of these data sets. Results: Voxel-based Cox regression revealed regions where reduced dose was correlated with increased prostate-specific androgen progression. Reduced dose in regions associated with coverage at the posterior prostate, in the immediate periphery of the posterior prostate, and in regions corresponding to the posterior oblique beams or posterior lateral beam boundary, was associated with increased PSAP for RADAR and RT01 patients, but not for CHHiP patients. Reduced dose to the seminal vesicle region was also associated with increased PSAP for RADAR intermediate-risk patients. Conclusions: Ensuring adequate dose coverage at the posterior prostate and immediately surrounding posterior region (including the seminal vesicles), where aggressive cancer spread may be occurring, may improve tumor control. It is recommended that particular care be taken when defining margins at the prostate posterior, acknowledging the trade-off between quality of life due to rectal dose and the preferences of clinicians and patients.

Introduction: In the management of early-stage breast cancer using radiation therapy, computed tomography (CT) simulation is used to identify the breast conservation surgery (BCS) seroma as a proxy for the tumour bed. The delineation or contouring of the seroma is generally a task performed by a radiation oncologist (RO). With increasing patient numbers and other demands placed on ROs, the scope of practice for radiation therapists (RTs) is continually expanding, and the need for skills transfer from one profession to another has been investigated in recent years. This study aims to compare the BCS seroma volumes contoured by RTs with those contoured by ROs to add evidence in support of expanding the RTs' role in the treatment planning process in the management of early-stage breast cancer. Methods: A study was undertaken using the CT-simulation (CT-sim) data sets of patients with early-stage breast cancer treated in 2013. The CT-sim data sets had BCS seromas contoured by 1 of 5 ROs as part of routine clinical management. This study involved 4 RTs who each used the patient information to identify and contour breast seromas on 50 deidentified CT-sim data sets. Metrics used to compare RT versus RO contours included volume size, overlap between volumes, and geographical distance from the centre of volumes. Results: There were 50 CT-sim data sets with 1 RO contour and 4 RT contours analysed. The contour volumes of the 4 RTs and the ROs were assessed. Although there were 50 CT-sim data sets presented to each RT, analysis was carried out on 45, 43, 46, and 45 CT-sim data sets. There were no comparisons made where contours were not delineated. The contour volumes of the 4 RTs and the ROs were assessed with an interclass correlation coefficient, with a result of excellent reliability (0.975, 95% [0.963, 0.985]). The DICE similarity coefficient was used to compare the overlap of each RT contour with the RO contour; the results were favourable with mean (95% CI) DSCs 0.685, 0.640, 0.678, and 0.681, respectively. Comparing the RT and RO geographical centre of the seroma volumes, good to excellent reliability between the RTs and ROs was demonstrated (95% CI mean RO vs RT distances (mm): 3.75, 4.99, 7.71, and 3.39). There was no statistically significant difference between the distances (P = 0.65). Conclusion: BCS seromas contoured by RTs compared well with those contoured by an RO. This research has provided further evidence to support RTs in assuming additional contouring responsibilities in radiation therapy planning for patients with early-stage breast cancer.

Introduction: Support and investment in increasing a research-active culture in clinical practice needs to be translated at the department and hospital levels as well as regional, state and national levels. We aimed to improve the research culture of our department, to enable more clinical staff to become more research competent and research active. Methods: We describe and discuss the appointment of a Director of Research and a Research Coordinator into our already-research-active department and the interactions at the research¿clinical interface. By identifying barriers and instituting enablers which ameliorate their effect, we explore how a clinical department can utilize the resources already available with the goal of developing a more confident and competent clinician-researcher culture as measured by a range of research metrics. Results: We observed an improved research culture within our department. Our department's improved research culture was reflected by increased numbers of peer-reviewed publications (of 30%), research students/supervisions (of 60%) and engagement of external speakers. We also observed double the number of first-authored peer-reviewed articles and a growth in conference presentations, posters and speaker invitations/awards. In the majority of the research performance metrics tracked, there was a steady improvement noted over the four years monitored. Conclusions: By responding to the barriers of staff (such as time, expertise and ideas) with structural and personal enablers, as well as funded resources, it is possible to develop research capacity and confidence in a clinical setting.

Purpose: An ideal commissioning and quality assurance (QA) program for Volumetric Modulated Arc Therapy (VMAT) delivery systems should assess the performance of each individual dynamic component as a function of gantry angle. Procedures within such a program should also be time-efficient, independent of the delivery system and be sensitive to all types of errors. The purpose of this work is to develop a system for automated time-resolved commissioning and QA of VMAT control systems which meets these criteria. Methods: The procedures developed within this work rely solely on images obtained, using an electronic portal imaging device (EPID) without the presence of a phantom. During the delivery of specially designed VMAT test plans, EPID frames were acquired at 9.5 Hz, using a frame grabber. The set of test plans was developed to individually assess the performance of the dose delivery and multileaf collimator (MLC) control systems under varying levels of delivery complexities. An in-house software tool was developed to automatically extract features from the EPID images and evaluate the following characteristics as a function of gantry angle: dose delivery accuracy, dose rate constancy, beam profile constancy, gantry speed constancy, dynamic MLC positioning accuracy, MLC speed and acceleration constancy, and synchronization between gantry angle, MLC positioning and dose rate. Machine log files were also acquired during each delivery and subsequently compared to information extracted from EPID image frames. Results: The largest difference between measured and planned dose at any gantry angle was 0.8% which correlated with rapid changes in dose rate and gantry speed. For all other test plans, the dose delivered was within 0.25% of the planned dose for all gantry angles. Profile constancy was not found to vary with gantry angle for tests where gantry speed and dose rate were constant, however, for tests with varying dose rate and gantry speed, segments with lower dose rate and higher gantry speed exhibited less profile stability. MLC positional accuracy was not observed to be dependent on the degree of interdigitation. MLC speed was measured for each individual leaf and slower leaf speeds were shown to be compensated for by lower dose rates. The test procedures were found to be sensitive to 1 mm systematic MLC errors, 1 mm random MLC errors, 0.4 mm MLC gap errors and synchronization errors between the MLC, dose rate and gantry angle controls systems of 1. In general, parameters measured by both EPID and log files agreed with the plan, however, a greater average departure from the plan was evidenced by the EPID measurements. Conclusion: QA test plans and analysis methods have been developed to assess the performance of each dynamic component of VMAT deliveries individually and as a function of gantry angle. This methodology relies solely on time-resolved EPID imaging without the presence of a phantom and has been shown to be sensitive to a range of delivery errors. The procedures developed in this work are both comprehensive and time-efficient and can be used for streamlined commissioning and QA of VMAT delivery systems.

Purpose Respiratory variation can increase the variability of tumor position and volume, accounting for larger treatment margins and longer treatment times. Audiovisual biofeedback as a breath-hold technique could be used to improve the reproducibility of lung tumor positions at inhalation and exhalation for the radiation therapy of mobile lung tumors. This study aimed to assess the impact of audiovisual biofeedback breath-hold (AVBH) on interfraction lung tumor position reproducibility and volume consistency for respiratory-gated lung cancer radiation therapy. Methods Lung tumor position and volume were investigated in 9 patients with lung cancer who underwent a breath-hold training session with AVBH before 2 magnetic resonance imaging (MRI) sessions. During the first MRI session (before treatment), inhalation and exhalation breath-hold 3-dimensional MRI scans with conventional breath-hold (CBH) using audio instructions alone and AVBH were acquired. The second MRI session (midtreatment) was repeated within 6 weeks after the first session. Gross tumor volumes (GTVs) were contoured on each dataset. CBH and AVBH were compared in terms of tumor position reproducibility as assessed by GTV centroid position and position range (defined as the distance of GTV centroid position between inhalation and exhalation) and tumor volume consistency as assessed by GTV between inhalation and exhalation. Results Compared with CBH, AVBH improved the reproducibility of interfraction GTV centroid position by 46% (P = .009) from 8.8 mm to 4.8 mm and GTV position range by 69% (P = .052) from 7.4 mm to 2.3 mm. Compared with CBH, AVBH also improved the consistency of intrafraction GTVs by 70% (P = .023) from 7.8 cm to 2.5 cm . Conclusions This study demonstrated that audiovisual biofeedback can be used to improve the reproducibility and consistency of breath-hold lung tumor position and volume, respectively. These results may provide a pathway to achieve more accurate lung cancer radiation treatment in addition to improving various medical imaging and treatments by using breath-hold procedures. 3 3

Background and Purpose: Electronic portal imaging devices (EPIDs) can be used to reconstruct dose inside a virtual phantom. This work aims to study the feasibility of using this method for remote dosimetry auditing of clinical trials. Materials and Methods: Six centres participated in an intensity modulated radiotherapy (IMRT) pilot study of this new audit approach. Each centre produced a head and neck (HN) and post-prostatectomy (PP) trial plan and transferred the plans to virtual phantoms to calculate a reference dose distribution. They acquired in-air images of the treatment fields along with calibration images using their EPID. These data were sent to the central site where the images were converted to 2D field-by-field doses in a flat virtual water phantom and to 3D combined field doses in a cylindrical virtual phantom for comparison with corresponding reference dose distributions. Additional test images were used to assess the accuracy of the method when using different EPIDs. Results: Field-by-field 2D analysis yielded mean gamma pass-rates of 99.6% (±0.3%) and 99.6% (±0.6%) for HN and PP plans respectively (3%/3 mm, doses greater than 10% global max). 3D combined field analysis gave mean pass-rates of 97.9% (±2.6%) and 97.9% (±1.8%) for the HN and PP plans. Dosimetry tests revealed some field size limitations of the EPIDs. Conclusions: The remote auditing methodology using EPIDs is feasible and potentially an inexpensive method.

Purpose: The dynamic keyhole is a newMRimage reconstruction method for thoracic and abdominal MR imaging. To date, this method has not been investigated with cancer patient magnetic resonance imaging (MRI) data. The goal of this study was to assess the dynamic keyhole method for the task of lung tumor localization using cine-MR images reconstructed in the presence of respiratory motion. Methods: The dynamic keyhole method utilizes a previously acquired a library of peripheral k-space datasets at similar displacement and phase (where phase is simply used to determine whether the breathing is inhale to exhale or exhale to inhale) respiratory bins in conjunction with central k-space datasets (keyhole) acquired. External respiratory signals drive the process of sorting, matching, and combining the two k-space streams for each respiratory bin, thereby achieving faster image acquisition without substantial motion artifacts. This study was the first that investigates the impact of k-space undersampling on lung tumor motion and area assessment across clinically available techniques (zero-filling and conventional keyhole). In this study, the dynamic keyhole, conventional keyhole and zero-filling methods were compared to full k-space dataset acquisition by quantifying (1) the keyhole size required for central k-space datasets for constant image quality across sixty four cine-MRI datasets from nine lung cancer patients, (2) the intensity difference between the original and reconstructed images in a constant keyhole size, and (3) the accuracy of tumor motion and area directly measured by tumor autocontouring. Results: For constant image quality, the dynamic keyhole method, conventional keyhole, and zerofilling methods required 22%, 34%, and 49% of the keyhole size (P < 0.0001), respectively, compared to the full k-space image acquisition method. Compared to the conventional keyhole and zero-filling reconstructed images with the keyhole size utilized in the dynamic keyhole method, an average intensity difference of the dynamic keyhole reconstructed images (P < 0.0001) was minimal, and resulted in the accuracy of tumor motion within 99.6% (P < 0.0001) and the accuracy of tumor area within 98.0% (P < 0.0001) for lung tumor monitoring applications. Conclusions: This study demonstrates that the dynamic keyhole method is a promising technique for clinical applications such as image-guided radiation therapy requiring the MR monitoring of thoracic tumors. Based on the results from this study, the dynamic keyhole method could increase the imaging frequency by up to a factor of five compared with full k-space methods for real-time lung tumor MRI.

In-room cine-MRI guidance can provide non-invasive target localization during radiotherapy treatment. However, in order to cope with finite imaging frequency and system latencies between target localization and dose delivery, tumour motion prediction is required. This work proposes a framework for motion prediction dedicated to cine-MRI guidance, aiming at quantifying the geometric uncertainties introduced by this process for both tumour tracking and beam gating. The tumour position, identified through scale invariant features detected in cine-MRI slices, is estimated at high-frequency (25 Hz) using three independent predictors, one for each anatomical coordinate. Linear extrapolation, auto-regressive and support vector machine algorithms are compared against systems that use no prediction or surrogate-based motion estimation. Geometric uncertainties are reported as a function of image acquisition period and system latency. Average results show that the tracking error RMS can be decreased down to a [0.2; 1.2] mm range, for acquisition periods between 250 and 750 ms and system latencies between 50 and 300 ms. Except for the linear extrapolator, tracking and gating prediction errors were, on average, lower than those measured for surrogate-based motion estimation. This finding suggests that cine-MRI guidance, combined with appropriate prediction algorithms, could relevantly decrease geometric uncertainties in motion compensated treatments.

Active shape models (ASMs) have proved successful in automatic segmentation by using shape and appearance priors in a number of areas such as prostate segmentation, where accurate contouring is important in treatment planning for prostate cancer. The ASM approach however, is heavily reliant on a good initialisation for achieving high segmentation quality. This initialisation often requires algorithms with high computational complexity, such as three dimensional (3D) image registration. In this work, we present a fast, self-initialised ASM approach that simultaneously fits multiple objects hierarchically controlled by spatially weighted shape learning. Prominent objects are targeted initially and spatial weights are progressively adjusted so that the next (more difficult, less visible) object is simultaneously initialised using a series of weighted shape models. The scheme was validated and compared to a multi-atlas approach on 3D magnetic resonance (MR) images of 38 cancer patients and had the same (mean, median, inter-rater) Dice's similarity coefficients of (0.79, 0.81, 0.85), while having no registration error and a computational time of 12-15 min, nearly an order of magnitude faster than the multi-atlas approach.

Purpose: The feasibility of radiation therapy treatment planning using substitute computed tomography (sCT) generated from magnetic resonance images (MRIs) has been demonstrated by a number of research groups. One challenge with an MRI-alone workflow is the accurate identification of intraprostatic gold fiducial markers, which are frequently used for prostate localization prior to each dose delivery fraction. This paper investigates a template-matching approach for the detection of these seeds in MRI. Methods: Two different gradient echo T1 and T2* weighted MRI sequences were acquired from fifteen prostate cancer patients and evaluated for seed detection. For training, seed templates from manual contours were selected in a spectral clustering manifold learning framework. This aids in clustering "similar" gold fiducial markers together. The marker with the minimum distance to a cluster centroid was selected as the representative template of that cluster during training. During testing, Gaussian mixture modeling followed by a Markovian model was used in automatic detection of the probable candidates. The probable candidates were rigidly registered to the templates identified from spectral clustering, and a similarity metric is computed for ranking and detection. Results: A fiducial detection accuracy of 95% was obtained compared to manual observations. Expert radiation therapist observers were able to correctly identify all three implanted seeds on 11 of the 15 scans (the proposed method correctly identified all seeds on 10 of the 15). Conclusions: An novel automatic framework for gold fiducial marker detection in MRI is proposed and evaluated with detection accuracies comparable to manual detection. When radiation therapists are unable to determine the seed location in MRI, they refer back to the planning CT (only available in the existing clinical framework); similarly, an automatic quality control is built into the automatic software to ensure that all gold seeds are either correctly detected or a warning is raised for further manual intervention.

Purpose: Radiation treatments are trending toward delivering higher doses per fraction under stereotactic radiosurgery and hypofractionated treatment regimens. There is a need for accurate 3D in vivo patient dose verification using electronic portal imaging device (EPID) measurements. This work presents a model-based technique to compute full three-dimensional patient dose reconstructed from on-treatment EPID portal images (i.e., transmission images). Methods: EPID dose is converted to incident fluence entering the patient using a series of steps which include converting measured EPID dose to fluence at the detector plane and then back-projecting the primary source component of the EPID fluence upstream of the patient. Incident fluence is then recombined with predicted extra-focal fluence and used to calculate 3D patient dose via a collapsed-cone convolution method. This method is implemented in an iterative manner, although in practice it provides accurate results in a single iteration. The robustness of the dose reconstruction technique is demonstrated with several simple slab phantom and nine anthropomorphic phantom cases. Prostate, head and neck, and lung treatments are all included as well as a range of delivery techniques including VMAT and dynamic intensity modulated radiation therapy (IMRT). Results: Results indicate that the patient dose reconstruction algorithm compares well with treatment planning system computed doses for controlled test situations. For simple phantom and square field tests, agreement was excellent with a 2%/2 mm 3D chi pass rate .98.9%. On anthropomorphic phantoms, the 2%/2 mm 3D chi pass rates ranged from 79.9% to 99.9% in the planning target volume (PTV) region and 96.5% to 100% in the low dose region (>20% of prescription, excluding PTV and skin build-up region). Conclusions: An algorithm to reconstruct delivered patient 3D doses from EPID exit dosimetry measurements was presented. The method was applied to phantom and patient data sets, as well as for dynamic IMRT and VMAT delivery techniques. Results indicate that the EPID dose reconstruction algorithm presented in this work is suitable for clinical implementation.

Purpose To validate automatic substitute computed tomography CT (sCT) scans generated from standard T2-weighted (T2w) magnetic resonance (MR) pelvic scans for MR-Sim prostate treatment planning. Patients and Methods A Siemens Skyra 3T MR imaging (MRI) scanner with laser bridge, flat couch, and pelvic coil mounts was used to scan 39 patients scheduled for external beam radiation therapy for localized prostate cancer. For sCT generation a whole-pelvis MRI scan (1.6 mm 3-dimensional isotropic T2w SPACE [Sampling Perfection with Application optimized Contrasts using different flip angle Evolution] sequence) was acquired. Three additional small field of view scans were acquired: T2w, T2*w, and T1w flip angle 80° for gold fiducials. Patients received a routine planning CT scan. Manual contouring of the prostate, rectum, bladder, and bones was performed independently on the CT and MR scans. Three experienced observers contoured each organ on MRI, allowing interobserver quantification. To generate a training database, each patient CT scan was coregistered to their whole-pelvis T2w using symmetric rigid registration and structure-guided deformable registration. A new multi-atlas local weighted voting method was used to generate automatic contours and sCT results. Results The mean error in Hounsfield units between the sCT and corresponding patient CT (within the body contour) was 0.6 ± 14.7 (mean ± 1 SD), with a mean absolute error of 40.5 ± 8.2 Hounsfield units. Automatic contouring results were very close to the expert interobserver level (Dice similarity coefficient): prostate 0.80 ± 0.08, bladder 0.86 ± 0.12, rectum 0.84 ± 0.06, bones 0.91 ± 0.03, and body 1.00 ± 0.003. The change in monitor units between the sCT-based plans relative to the gold standard CT plan for the same dose prescription was found to be 0.3% ± 0.8%. The 3-dimensional ¿ pass rate was 1.00 ± 0.00 (2 mm/2%). Conclusions The MR-Sim setup and automatic sCT generation methods using standard MR sequences generates realistic contours and electron densities for prostate cancer radiation therapy dose planning and digitally reconstructed radiograph generation.

The quantification of tumor motion in sites affected by respiratory motion is of primary importance to improve treatment accuracy. To account for motion, different studies analyzed the translational component only, without focusing on the rotational component, which was quantified in a few studies on the prostate with implanted markers. The aim of our study was to propose a tool able to quantify lung tumor rotation without the use of internal markers, thus providing accurate motion detection close to critical structures such as the heart or liver. Specifically, we propose the use of an automatic feature extraction method in combination with the acquisition of fast orthogonal cine MRI images of nine lung patients. As a preliminary test, we evaluated the performance of the feature extraction method by applying it on regions of interest around (i) the diaphragm and (ii) the tumor and comparing the estimated motion with that obtained by (i) the extraction of the diaphragm profile and (ii) the segmentation of the tumor, respectively. The results confirmed the capability of the proposed method in quantifying tumor motion. Then, a point-based rigid registration was applied to the extracted tumor features between all frames to account for rotation. The median lung rotation values were -0.6 ± 2.3° and -1.5 ± 2.7° in the sagittal and coronal planes respectively, confirming the need to account for tumor rotation along with translation to improve radiotherapy treatment.

Background: CT-MR registration is a critical component of many radiation oncology protocols. In prostate external beam radiation therapy, it allows the propagation of MR-derived contours to reference CT images at the planning stage, and it enables dose mapping during dosimetry studies. The use of carefully registered CT-MR atlases allows the estimation of patient specific electron density maps from MRI scans, enabling MRI-alone radiation therapy planning and treatment adaptation. In all cases, the precision and accuracy achieved by registration influences the quality of the entire process.Problem: Most current registration algorithms do not robustly generalize and lack inverse-consistency, increasing the risk of human error and acting as a source of bias in studies where information is propagated in a particular direction, e.g. CT to MR or vice versa. In MRI-based treatment planning where both CT and MR scans serve as spatial references, inverse-consistency is critical, if under-acknowledged.Purpose: A robust, inverse-consistent, rigid/affine registration algorithm that is well suited to CT-MR alignment in prostate radiation therapy is presented.Method: The presented method is based on a robust block-matching optimization process that utilises a half-way space definition to maintain inverse-consistency. Inverse-consistency substantially reduces the influence of the order of input images, simplifying analysis, and increasing robustness. An open source implementation is available online at http://aehrc.github.io/Mirorr/.Results: Experimental results on a challenging 35 CT-MR pelvis dataset demonstrate that the proposed method is more accurate than other popular registration packages and is at least as accurate as the state of the art, while being more robust and having an order of magnitude higher inverse-consistency than competing approaches.Conclusion: The presented results demonstrate that the proposed registration algorithm is readily applicable to prostate radiation therapy planning.

The purpose of this study was to investigate performance of the couch and coil mounts designed for MR-simulation prostate scanning using data from ten volunteers. Volunteers were scanned using the standard MR scanning protocol with the MR coil directly strapped on the external body and the volunteer lying on the original scanner table. They also were scanned using a MR-simulation table top and pelvic coil mounts. MR images from both setups were compared in terms of body contour variation and image quality effects within particular organs of interest. Six-field conformal plans were generated on the two images with assigned bulk density for dose calculation. With the MR-simulation devices, the anterior skin deformation was reduced by up to 1.7 cm. The hard tabletop minimizes the posterior body deformation which can be up to 2.3 cm on the standard table, depending on the weight of volunteer. The image signal-to-noise ratio reduced by 14% and 25% on large field of view (FOV) and small FOV images, respectively, after using the coil mount; the prostate volume contoured on two images showed difference of 1.05 ± 0.66 cm³. The external body deformation caused a mean dose reduction of 0.6 ± 0.3 Gy, while the coverage reduced by 22% ± 13% and 27% ± 6% in V98 and V100, respectively. A dedicated MR simulation setup for prostate radiotherapy is essential to ensure the agreement between planning anatomy and treatment anatomy. The image signal was reduced after applying the coil mount, but no significant effect was found on prostate contouring.

To clinically implement MRI simulation or MRI-alone treatment planning requires comprehensive end-to-end testing to ensure an accurate process. The purpose of this study was to design and build a geometric phantom simulating a human male pelvis that is suitable for both CT and MRI scanning and use it to test geometric and dosimetric aspects of MRI simulation including treatment planning and digitally reconstructed radiograph (DRR) generation. A liquid filled pelvic shaped phantom with simulated pelvic organs was scanned in a 3T MRI simulator with dedicated radiotherapy couch-top, laser bridge and pelvic coil mounts. A second phantom with the same external shape but with an internal distortion grid was used to quantify the distortion of the MR image. Both phantoms were also CT scanned as the gold-standard for both geometry and dosimetry. Deformable image registration was used to quantify the MR distortion. Dose comparison was made using a seven-field IMRT plan developed on the CT scan with the fluences copied to the MR image and recalculated using bulk electron densities. Without correction the maximum distortion of the MR compared with the CT scan was 7.5 mm across the pelvis, while this was reduced to 2.6 and 1.7 mm by the vendor's 2D and 3D correction algorithms, respectively. Within the locations of the internal organs of interest, the distortion was <1.5 and <1 mm with 2D and 3D correction algorithms, respectively. The dose at the prostate isocentre calculated on CT and MRI images differed by 0.01% (1.1 cGy). Positioning shifts were within 1 mm when setup was performed using MRI generated DRRs compared to setup using CT DRRs. The MRI pelvic phantom allows end-to-end testing of the MRI simulation workflow with comparison to the gold-standard CT based process. MRI simulation was found to be geometrically accurate with organ dimensions, dose distributions and DRR based setup within acceptable limits compared to CT.

© 2015 Australasian College of Physical Scientists and Engineers in Medicine To quantify the dose calculation error and resulting optimization uncertainty caused by performing inverse treatment planning on inaccurate electron density data (pseudo-CT) as needed for adaptive radiotherapy and Magnetic Resonance Imaging (MRI) based treatment planning. Planning Computer Tomography (CT) data from 10 cervix cancer patients was used to generate 4 pseudo-CT data sets. Each pseudo-CT was created based on an available method of assigning electron density to an anatomic image. An inversely modulated radiotherapy (IMRT) plan was developed on each planning CT. The dose calculation error caused by each pseudo-CT data set was quantified by comparing the dose calculated each pseudo-CT data set with that calculated on the original planning CT for the same IMRT plan. The optimization uncertainty introduced by the dose calculation error was quantified by re-optimizing the same optimization parameters on each pseudo-CT data set and comparing against the original planning CT. Dose differences were quantified by assessing the Equivalent Uniform Dose (EUD) for targets and relevant organs at risk. Across all pseudo-CT data sets and all organs, the absolute mean dose calculation error was 0.2 Gy, and was within 2 % of the prescription dose in 98.5 % of cases. Then absolute mean optimisation error was 0.3 Gy EUD, indicating that that inverse optimisation is impacted by the dose calculation error. However, the additional uncertainty introduced to plan optimisation is small compared the sources of variation which already exist. Use of inaccurate electron density data for inverse treatment planning results in a dose calculation error, which in turn introduces additional uncertainty into the plan optimization process. In this study, we showed that both of these effects are clinically acceptable for cervix cancer patients using four different pseudo-CT data sets. Dose calculation and inverse optimization on pseudo-CT is feasible for this patient cohort.

A new tool with the potential to verify and track jaw position during delivery has been developed. The method should be suitable for independent quality assurance for jaw position during jaw tracking dynamic IMRT and VMAT treatments. The jaw detection and tracking algorithm developed consists of five main steps. Firstly, the image is enhanced by removing a normalised predicted EPID image (that does not include the collimator transmission) from each cine EPID image. Then, using a histogram clustering technique a global intensity threshold level was determined. This threshold level was used to classify each pixel of the image as either under the jaws or under the MLC. Additionally, the collimator angle was automatically detected and used to rotate the image to vertical direction. Finally, this rotation allows the jaw positions to be determined using vertical and horizontal projection profiles. Nine IMRT fields (with static jaws) and a single VMAT clinical field (with dynamic jaws) were tested by determining the root mean square difference between planned and detected jaw positions. The test results give a detection accuracy of ±1mm RMS error for static jaw IMRT treatments and ±1.5mm RMS error for the dynamic jaw VMAT treatment. This method is designed for quality assurance and verification in modern radiation therapy; to detect the position of static jaws or verify the position of tracking jaws in more complex treatments. This method uses only information extracted from EPID images and it is therefore independent from the linear accelerator.

Purpose Gantry-mounted megavoltage electronic portal imaging devices (EPIDs) have become ubiquitous on linear accelerators. WatchDog is a novel application of EPIDs, in which the image frames acquired during treatment are used to monitor treatment delivery in real time. We report on the preliminary use of WatchDog in a prospective study of cancer patients undergoing intensity modulated radiation therapy (IMRT) and volumetric modulated arc therapy (VMAT) and identify the challenges of clinical adoption. Methods and Materials At the time of submission, 28 cancer patients (head and neck, pelvis, and prostate) undergoing fractionated external beam radiation therapy (24 IMRT, 4 VMAT) had =1 treatment fraction verified in real time (131 fractions or 881 fields). EPID images acquired continuously during treatment were synchronized and compared with model-generated transit EPID images within a frame time (~0.1 s). A ¿ comparison was performed to cumulative frames to gauge the overall delivery quality, and the resulting pass rates were reported graphically during treatment delivery. Every frame acquired (500-1500 per fraction) was saved for postprocessing and analysis. Results The system reported the mean ± standard deviation in real time ¿ 91.1% ± 11.5% (83.6% ± 13.2%) for cumulative frame ¿ analysis with 4%, 4 mm (3%, 3 mm) criteria, global over the integrated image. Conclusions A real-time EPID-based radiation delivery verification system for IMRT and VMAT has been demonstrated that aims to prevent major mistreatments in radiation therapy.

Methods: A custom-designed pelvic-shaped phantom was scanned by systematically increasing the anterior body-tocoil (BTC) distance from 30 to 90mm. The image quality near the organs of interest was determined in order to characterize the relationship between image quality and BTC distance at the critical organ structures. The half intensity reduction (HIR) was calculated to determine the sensitivity of each organ structure to the BTC distance change.

MRI-alone treatment planning and adaptive MRI-based prostate radiation therapy are two promising techniques that could significantly increase the accuracy of the curative dose delivery processes while reducing the total radiation dose. State-of-the-art methods rely on the registration of a patient MRI with a MR-CT atlas for the estimation of pseudo-CT [5]. This atlas itself is generally created by registering many CT and MRI pairs. Most registration methods are not symmetric, but the order of the images influences the result [8]. The computed transformation is therefore biased, introducing unwanted variability. This work examines how much a symmetric algorithm improves the registration. Methods: A robust symmetric registration algorithm is proposed that simultaneously optimises a half space transform and its inverse. During the registration process, the two input volumetric images are transformed to a common position in space, therefore minimising any computational bias. An asymmetrical implementation of the same algorithm was used for comparison purposes. Results: Whole pelvis MRI and CT scans from 15 prostate patients were registered, as in the creation of MR-CT atlases. In each case, two registrations were performed, with different input image orders, and the transformation error quantified. Mean residuals of 0.63±0.26 mm (translation) and (8.7±7.3) × 10 rad (rotation) were found for the asymmetrical implementation with corresponding values of 0.038±0.039 mm and (1.6 ± 1.3) × 10 rad for the proposed symmetric algorithm, a substantial improvement. Conclusions: The increased registration precision will enhance the generation of pseudo-CT from MRI for atlas based MR planning methods. © Published under licence by IOP Publishing Ltd. -3 -3

Objective: Manual contouring and registration for radiotherapy treatment planning and online adaptation for cervical cancer radiation therapy in computed tomography (CT) and magnetic resonance images (MRI) are often necessary. However manual intervention is time consuming and may suffer from inter or intra-rater variability. In recent years a number of computer-guided automatic or semi-automatic segmentation and registration methods have been proposed. Segmentation and registration in CT and MRI for this purpose is a challenging task due to soft tissue deformation, inter-patient shape and appearance variation and anatomical changes over the course of treatment. The objective of this work is to provide a state-of-the-art review of computer-aided methods developed for adaptive treatment planning and radiation therapy planning for cervical cancer radiation therapy. Methods: Segmentation and registration methods published with the goal of cervical cancer treatment planning and adaptation have been identified from the literature (PubMed and Google Scholar). A comprehensive description of each method is provided. Similarities and differences of these methods are highlighted and the strengths and weaknesses of these methods are discussed. A discussion about choice of an appropriate method for a given modality is provided. Results: In the reviewed papers a Dice similarity coefficient of around 0.85 along with mean absolute surface distance of 2-4. mm for the clinically treated volume were reported for transfer of contours from planning day to the treatment day. Conclusions: Most segmentation and non-rigid registration methods have been primarily designed for adaptive re-planning for the transfer of contours from planning day to the treatment day. The use of shape priors significantly improved segmentation and registration accuracy compared to other models.

Radiotherapy is an effective therapy in the treatment of cervix cancer. However tumor and normal tissue motion and shape deformation of the cervix, the bladder and the rectum over the course of the treatment can limit the efficacy of radiotherapy and safe delivery of the dose. A number of studies have presented the potential benefits of adaptive radiotherapy for cervix cancer with high soft tissue contrast magnetic resonance images. To enable practical implementation of adaptive radiotherapy for the cervix, computer aided segmentation is necessary. Accurate computer aided automatic or semi-automatic segmentation of the cervix is a challenging task due to inter patient shape variation, soft tissue deformation, organ motion, and anatomical changes during the course of the treatment. This article reviews the methods developed for cervix segmentation in magnetic resonance images. The objective of this work is to present different methods for cervix segmentation in the literature highlighting their similarities, differences, strengths and weaknesses. © 2013 Springer-Verlag.

The set-up variation of 11 patients treated supine with radical radiotherapy for carcinoma of the prostate was measured with an electronic portal imaging device to determine the adequacy of set-up techniques and current margins, as well as the need for immobilization. During the treatments 172 images of the anterior fields and 159 images of the left- lateral fields were taken and the errors in treatment placement were measured by template matching. The variation in the superior-inferior direction was small, 1.4-1.6 mm (1 SD), while the medio-lateral variation was 2.8 mm (1 SD). The anterior-posterior variation was largest, 4.6 mm (1 SD) with an offset of 3.3 mm anterior. This anterior offset and large anterior-posterior variation suggests that set-up techniques were not optimal for this direction. The 1 cm margin used was adequate for set-up variation except in a small number of cases, which was mainly due to the anterior trend. Random (treatment-to-treatment) variations were small (1.1-2.3 mm; 1 SD), indicating that immobilization would result in only modest improvement in reproducibility for these supine patients.

Accurate radiation dose calculation is a major challenge in magnetic resonance imaging (MRI)-only radiation therapy (RT) treatment planning as the required electron density map is not readily provided by this modality. In this work, a number of state-of-the-art synthetic-CT (sCT) generation methods, exhibited promising results in the literature, were evaluated based on common quantitative metrics and patients dataset. This includes four atlas-based approaches, specifically median of atlas images (A-Median) [1], atlas-based voxel-wise weighting (A-VW) [2], bone enhanced atlas-based voxel-wise weighting (A-Bone) [3], iterative atlas-based voxel-wise weighting (A-Iter) [4], and a method based on deep learning convolutional neural network (DL-CNN) [5]. Automatic organ delineation was performed for bladder, rectum and bone. Overall, A-VW, A-Bone, A-Iter and A-VWexhibited comparable performance while DL-CNN showed slightly better segmentation performance resulting in Dice metrics of 0.93, 0.90, and 0.93, respectively. The dosimetric evaluation demonstrated that A-Median, A-VW, A-Bone, A-Iter and DL-CNN resulted in comparable mean dose errors within organs at risk and target volumes showing less than 1% dose difference against the CT-based RT planning. The two-dimensional gamma analysis performed at 1%/1 mm criterion demonstrated comparable pass rates of 94.99±5.15%, 94.59±5.65%, 93.68±5.53% and 93.10±5.99% for A-Bone, DL-CNN, A-Median and A-Iter, respectively. Whereas A-VW and water-only resulted in pass rates of 86.91±13.50% and 80.77±12.10%, respectively. DL-CNN and advanced atlas-based approaches showed promising dosimetric and segmentation accuracy (DL-CNN is slightly better) suggesting that these methods are able to resolve the challenge of synthetic-CT generation from MR images with clinically acceptable errors.

Our group have been developing methods for MRI-alone prostate cancer radiation therapy treatment planning. To assist with clinical validation of the workflow we are investigating a cloud platform solution for research purposes. Benefits of cloud computing can include increased scalability, performance and extensibility while reducing total cost of ownership. In this paper we demonstrate the generation of DICOM-RT directories containing an automatic average atlas based electron density image and fast pelvic organ contouring from whole pelvis MR scans. © Published under licence by IOP Publishing Ltd.

Multimodal registration of CT and MR scans is a required step in leading edge adaptive MR-based image guided radiation therapy protocols. Yet, anatomical changes limit the precision of the registration process and therefore that of the whole intervention. In prostate radiation therapy, the difference in bladder and rectum filling can significantly displace both the targeted area and the organs at risk. Here, we describe a method that integrates an image-based similarity criterion with the anatomical information from manual contours to guide the registration process toward an accurate solution. Whole pelvis CT and MR scans of 33 patients have been nonrigidly registered, and the proposed method leads to an average improvement of 0.17 DSC when compared to a baseline nonrigid registrations. The increased accuracy will thus enhance an MR-based prostate radiation therapy protocol. © 2014 Springer International Publishing.

Purpose: Current model electronic portal imaging devices (EPIDs) used in radiotherapy are optimised for imaging but problematic for accurate dosimetry. The aim of this project is to develop a new EPID capable of simultaneous imaging and water equivalent dosimetry. This work reports our first experimental results. Methods: A prototype device based on a segmented plastic scintillator (SegmentedPS) was developed. The prototype device was tested in comparison with three other detector configurations, all utilising the same a-Si photodiode array used in conventional EPIDs. The configurations tested were; i) standard indirect configuration with phosphor/copper (STANDARD), ii) direct configuration with 1.5 cm solid water build-up (DIRECT), iii) 2 cm thick sheet of plastic scintillator (PSsheet), and iv) 1.5 cm thick prototype SegmentedPS. The sensitivity, dose response and image quality was assessed in each configuration. The dose response was assessed in terms of field size response and off-axis response relative to reference dose in water measurements at the equivalent depth. The image quality was assessed using an image quality phantom. Results: The sensitivity of each device relative to the STANDARD configuration was 0.03(DIRECT), 0.63(PSsheet), and 0.35(SegmentedPS). The agreement in field size response relative to dose in water data for square field sizes 4 cm to 15 cm was within 4.8%(STANDARD), 0.9%(DIRECT), 22%(PSsheet), and 1.3%(SegmentedPS). The agreement in off-axis ratios at 15 cm off-axis, relative to dose in water data, was within 23%(STANARD), 4%(DIRECT), 5%(PSsheet), and 4%(SegmentedPS). Image quality parameters (f50/CNR) for each configuration were 0.41/993(STANDARD), 0.30/167(DIRECT), 0.22/125(PSsheet) and 0.23/214(SegmentedPS). Conclusions: First experiments with the prototype SegmentedPS EPID demonstrated the potential for simultaneous imaging and water equivalent dosimetry. Further design optimisation is required to approach the imaging performance of STANDARD a-Si EPIDs, while maintaining water equivalent dose response. Cancer Council NSW Research Project Grant (RG 11-06) Cancer Institute NSW Research Equipment Grant (10/REG/1-20). © 2012, American Association of Physicists in Medicine. All rights reserved.

Purpose: To demonstrate a new method for real-time geometric and dosimetric verification of IMRT and VMAT using synchronization between predicted and measured EPID images. Methods: Predicted EPID images were calculated using a comprehensive physics-based model. Each predicted image represents the integrated signal expected from the delivery between control points. The measured images are acquired in cine mode and compared to the set of predicted images in real-time. The system performs geometric verification prior to dosimetric verification. When the measured image is acquired, the algorithm automatically detects the MLC leaf positions. A comparison between the leaf positions of the measured image and control points in the MLC file is made using the cosine similarity technique. The similarity index(SI) provides geometric MLC verification and synchronization between the measured and predicted images, as a uniform dose-rate cannot be assumed for IMRT or VMAT deliveries. The SI threshold was based on a series of experiments including 21 dynamic-IMRT fields defining pass/fail boundary(5 brain, 8 H&N, and 8 prostate cases).If geometric verification is successful, dosimetic verification is performed with the Gamma comparison(3%,3mm).The system reports the verification Result in real-time. Results: The system was simulated by MATLAB/SIMULINK and detected geometric and dosimetric errors during delivery. Both artificially introduced errors and clinical data were used for testing and analysis of the system performance. For a tested prostate field, the cumulative dose comparisons showed the minimum and maximum number of points with Gamma index<1 as 93.5% and 98.5%, respectively. For individual dose comparisons on the same field, the values were 87% and 97%, respectively. Conclusion: This method includes automatic MLC leaf positioning, synchronization, and dosimetric verification. The pass/fail boundary of geometry was calculated based on the experiments. This system is a useful approach to detect unexpected possible errors occurring in the clinical setting and to prevent patient overdoses during radiotherapy especially in complex deliveries such as arc-IMRT. © 2012, American Association of Physicists in Medicine. All rights reserved.

Purpose: A method that could enable dose calculations to be performed using magnetic resonance (MR) images for conventional treatment planning and adaptive planning using MR-accelerator systems would be to apply bulk electron densities to the MR images. However currently bone must be manually segmented making this impractical. This work develops and tests an atlas-based method to automatically segment bone on pelvic MR images for dose calculations. Methods: An MR whole-pelvic atlas was created using manually delineated scans from 39 patients. Atlas-based pelvic-bone auto-segmentations were then created for 25 patient scans using deformable image registration of the atlas to each patients scan with a leave-one-out atlas approach. These and corresponding expert manual segmentations were compared using the Dice similarity coefficient. Bone was assigned a density of 1.19 g/cm3 and all other tissues a water equivalent density. Treatment plans were generated on the whole-pelvis MR images and doses compared for the manual and auto-segmented bone plans. Results: The average Dice coefficient was 0.83 (standard deviation = 0.05). The average manual bone volume was 834.6 cm3 compared to the atlas based average volume of 840.0 cm3, with a mean difference of 5.4 cm3 (0.64%). The average ICRU point dose calculated on the MR images using the atlas-based bone segmentations was 0.2% lower (standard deviation 0.3%) than the dose calculated using the manual bone segmentations. Conclusions: The atlas based method for auto-segmentation of pelvic bone enables MR-based prostate radiotherapy dose calculations for treatment and adaptive planning. © 2011, American Association of Physicists in Medicine. All rights reserved.

Purpose: To investigate the performance of a modified backscatter-shielded Electronic Portal Imaging Device (BS-EPID) system and to develop a model to convert the BS-EPID images to water-equivalent dose. Methods: Images were acquired with a modified Varian aS-1000 BS-EPID and compared to those from a comparable clinical EPID. The asymmetry of in-plane profiles was examined as a measure of how well the system reduces the effect of the support arm backscatter. A new pixel-sensitivity correction method was assessed by comparing BS-EPID images of the same fields at different detector offset positions. A model was developed to determine the water equivalent dose from images acquired with the BS-EPID. The model converts the BS-EPID image to fluence using a deconvolution kernel, then to dose in water using a convolution kernel. The model was optimized based on experimental measurements and can be applied to construct dose in water at any depth. The validity of the model was tested using gamma analysis to compare 28 two-dimensional dose maps of IMRT fields measured with the BS-EPID at different depths to those predicted by the Eclipse TPS. Results: The BS-EPID profiles gave reduced asymmetry (0.6% compared to 3.3%), showing that the backscatter shielding is effective. BS-EPID images were consistent with different detector offsets within 0.4% on average. The IMRT dose maps measured with the BS-EPID gave good agreement with those predicted by the TPS, with mean values of 94.6%, 91.8%, 94.0% and 95.3% of pixels meeting the gamma criteria of 2%, 2mm at depths of 1.5, 5, 10 and 20 cm respectively. Conclusions: The BS-EPID performs effectively in reducing the effect of the backscatter from the support arm in Varian EPIDs. The BS-EPID dosimetry system and model allows high-resolution water equivalent doses to be measured with the BS-EPID, streamlining IMRT quality assurance. The authors have received support, in the form of equipment loans and technical information, from Varian Medical Systems iLab GmbH. © 2011, American Association of Physicists in Medicine. All rights reserved.

Purpose: Utilization of an aSi EPID to develop an in vivo patient dose verification system for rotational IMRT (rIMRT) delivery requires accurate knowledge of gantry angle as a function of time. Currently the accuracy of the gantry angle stamp in the header of the EPID image is limited to approximately +/-3 degrees. This work investigates several unique methods for a more accurate determination of the gantry angle during rIMRT. Methods: Gantry angles were determined using: (1) an incremental rotary encoder attached to the rotational axis of the gantry, (2) a direct analogue-to-digital measurement of the gantry potentiometer, and (3) through EPID image analyses of an in-house phantom (manufactured at sub-millimeter precision). The phantom consists of a cylindrical acrylic frame with one wire wrapped helically around its surface and one straight wire traversing its central axis. This design creates EPID images with unique and identifiable wire intersection points as a function of gantry orientation. Analysis of the treatment console log files was compared to the above methods. Results: The gantry potentiometer is considered the most accurate gantry angle but is unavailable during treatment. The ClinacLog produced discrepancies of up to ±2 degrees, the DynaLog up to ±1 degrees, and the encoder up to ±0.5 degrees with respect to the potentiometer. Preliminary analysis comparing our phantom-determined gantry angles with the encoder gantry angles showed agreement within ±0.5 degrees of each other for 85% of the data and differed at most by 1.3 degrees from each other. Conclusions: We have developed several techniques to determine gantry angle as a function of time during rIMRT. We have shown a strong agreement in gantry determination by our phantom and encoder. This investigation of gantry angle is critical to develop an accurate in vivo patient dose verification system for rIMRT delivery. © 2011, American Association of Physicists in Medicine. All rights reserved.

Purpose: To develop a comprehensive Monte Carlo (MC) model of an indirect-detection electronic portal imaging device (EPID) that can self-consistently quantify the effect of optical blur on the output signal. Methods: A model of an indirect-detection EPID was developed using the Geant4 MC toolkit. The EPID was modeled as a series of uniform slabs with thicknesses and material properties obtained from published literature. The model also included a slab of solid water backscatter material directly beyond the EPID rear housing. The standard electromagnetic and optical physics Geant4 modules were incorporated into the model to simultaneously simulate both high energy and optical photon transport relevant for indirect-detection EPIDs. A narrow, monoenergetic beam of 1 MeV photons was used to generate a line of radiation normally incident on the EPID surface. The beam width was equal to the pixel pitch of 0.4 mm used for scoring particle hits and energy deposition in the gadolinium oxysulfide scintillator and amorphous silicon photodiode layers. Optical and gamma photons were scored separately in the photodiode layer to measure their relative effects on the output signal. Line spread functions (LSFs) were generated indicating the distribution of hits and energy deposited across the scintillator and photodiode planes. Results: The LSFs for optical photon hits in the photodiode array and energy deposition events in the scintillator had a FWHM of approximately 4.7 mm and 0.82 mm, respectively. This indicates a significant increase in image blurring due to optical photon scatter. Conclusions: Our results indicate that modeling optical photon transport may be important when simulating imager performance for an indirect-detection EPID. Further analysis of calculated LSFs, including determination of the detector modulation transfer function, is required to further quantify imager performance. Cancer Council NSW Research Project Grant RG 11-06 Cancer Institute NSW Research Equipment Grant 10/REG/1-20. © 2011, American Association of Physicists in Medicine. All rights reserved.

Purpose: To create a physical fluence model for portal dose image prediction that will be used to verify patient radiation treatment delivery. Methods and Materials: A physical fluence model was created to predict portal dose images. This model utilized Monte Carlo simulation and linac-specific engineering schematics of the MLCs to create as accurate a model as possible. The fluence model consists of a focal and extra-focal source, determined to be a Gaussian and Gaussian-like function, respectively. The MLC transmission is calculated by attenuating a pre-MLC BEAMnrc spectrum through the leaves; the MLC leaf-tip and tongue-and-groove are modeled using the schematics. Incident energy fluence profiles from BEAMnrc are used to account for the field shape and beam horns. The asymmetric backscatter from the EPID arm is also modeled. The energy fluence is converted to dose using a superposition of EPID-specific dose kernels. Scatter from the patient or phantom is approximated using Monte Carlo calculated scatter fluence kernels. The model is tested on simple slab phantoms for a variety of field sizes, thicknesses and air gaps. It was also tested on one field acquired during a patient's prostate IMRT treatment. Results: Predicted images with phantoms in the beam agreed within 2%, 2 mm in relative comparison with the measured images. After calibration of measured and predicted images to absolute units, the images agreed within 3% and 3 mm. The patient IMRT field predicted image was not ideal when compared to the measured image, mainly because the patient is heterogeneous. Conclusions: The physical fluence model, in conjunction with the patient scatter model, is accurate and could be used to verify patient treatment. The algorithm is versatile and can be applied to a variety of treatment scenarios. © 2011, American Association of Physicists in Medicine. All rights reserved.

Purpose: A cine-EPID based method to separately measure the mechanical motion characteristics of gantry, jaws and multi-leaf collimator as a function of gantry angle during VMAT has been developed. Methods: Irradiations were performed using 6 MV beams of a Varian Trilogy linear accelerator with an aS1000 EPID. Images were acquired using 360 MU irradiations at 600 MU/min in cine acquisition mode at 2 Hz frame-rate. To establish the gantry isocentre, a Winston-Lutz technique was used with a circular collimator rigidly attached to the gantry head. The displacement of the centre of a fixed tungsten ball at isocentre from the field centre was determined on each image with a sub-pixel thresholding technique. EPID coordinates were transformed to room coordinates. Jaw and MLC sag relative to the gantry head were determined from the displacement of a gantry-mounted tungsten ball relative to static jaw and MLC positions. MLC speed constancy was determined by segmenting MLC positions on each image for a constant leaf-speed test pattern. Gantry speed constancy was assessed with an independent liquid-based inclinometer and the linac gantry angle potentiometer. Results: The gantry isocentre was ~ 1 mm amplitude with changes in isocentre occuring over time suggesting frequent measurement is required. Jaw sag was found to be very small ~ 0.2 mm amplitude, with MLC sag ~ 0.6 mm. Average leaf speed was found to be consistent for the MLC leaves however the variation in speed varied systematically across the leaf bank which requires further investigation. Gantry speed was consistent although the inclinometer was found to lag ~ 2 degrees in reading from the potentiometer. Conclusions: The cine-EPID based method quantifies the mechanical motion characteristics of the linac as a function of gantry angle during arc-therapy with improved precision and efficiency of quality assurance over previous methods. © 2011, American Association of Physicists in Medicine. All rights reserved.

This project develops atlas-based deformable image registration methods to map electron densities and automatically segment organs on MRI scans. This will enable dose calculations to be performed using the MRI scan without the requirement for an additional CT scan. The method developed uses atlas-based deformable image registration. An MRI atlas was developed based on whole pelvic MRI scans for 39 patients. The atlas is then registered to an individual patient MRI scan. The registration of the atlas organ contours gives the automatically segmented organs on that patient scan. A CT or electron-density atlas was also developed that corresponds to the MRI atlas. The deformation vectors that register the MRI-atlas to the patient MRI scan are applied to the CT-atlas to produce a pseudo-CT scan for the patient. This can then be used for dose planning and digitally reconstructed radiographs. The feasibility of the entire workflow has been tested for one patient. The rectum, bladder, prostate and bone were automatically segmented on the MRI scan with Dice coefficient results of 0.82, 0.59, 0.62 and 0.81. The patient's plan was applied to the pseudo-CT using a commercial treatment planning system. Dose at the normalisation point was 2.9% lower than on the full density CT plan. This method will improve the workflow of prostate radiotherapy planning and will reduce systematic uncertainties introduced by MRI-CT registration. © 2010, American Association of Physicists in Medicine. All rights reserved.

The accuracy of dosimetry measurements using a-Si EPIDs is affected by their structural characteristics particularly due to the presence of the Gd O S phosphor layer. Our previous measurements have shown that modification of the structure to direct detection configuration by removal of the phosphor layer can improve the imager properties for transit dosimetry applications. In this study a research dedicated Varian a-Si EPID has been changed to direct detection configuration and evaluated for transit dosimetry measurements using 7 prostate and 9 IMRT fields with a 20 cm thick phantom in the beam by comparison to a MatriXX detector array. The EPID images were converted to dose using a calibrated 0.6 cc ionization chamber. Gamma evaluation (3%, 3 mm criteria) of the results for for all points greater than 10% of the maximum dose showed that the fraction of points with a Gamma index less than 1 was at least 94.6% in head and neck fields and 99.3% in prostate fields. The mean Gamma was 0.360 and 0.288 for head and neck and prostate fields, respectively. In conclusion, The EPID results are very close to the reference dosimeter and the direct EPID appears to be a promising device for online patient dosimetry applications for the future. © 2010, American Association of Physicists in Medicine. All rights reserved. 2 2

INTRODUCTION: Amorphous-silicon electronic portal imaging devices (EPIDs) have been established as useful tools for dosimetry. To accurately reconstruct the patient dose delivered during rotational IMRT, one must acquire time-resolved EPID images as a function of gantry-angle. Dose reconstruction accuracy is directly impacted by the accuracy of the geometry of the imaging system, including the gantry-angle readout (i.e. source geometry) and the EPID support-arm sag (i.e. imager geometry). This work investigates these two factors. METHODS: The EPID support-arm sag was investigated through measurements performed on Varian E-arm and R-arm models at two institutes and employing two different analysis methods. One method imaged an isocentric ball-bearing whose position was tracked over all gantry-angles. The second method involved analysing field edges to obtain the field centre location of all images. Gantry-angle accuracy was examined by comparing the gantry-angle indicated at the treatment console readout to the gantry-angle written to the EPID DICOM header. We developed a method of measuring gantry-angle directly from the gantry-angle potentiometer. RESULTS: The E-arm showed maximum displacement of roughly 0.6mm (cross-plane) and 0.8mm (in-plane). R-arm results were significantly worse, estimated at 8.5mm (cross-plane) and 5.0mm (in-plane). Gantry-angle analysis demonstrated approximately 2 degrees of uncertainty in the gantry-angle contained in the EPID image. A direct measurement of the gantry angle potentiometer was demonstrated. CONCLUSIONS: Two main factors affecting patient dose reconstruction using EPID dosimetry have been investigated. EPID support-arm sag can be measured (and corrected). Near real-time gantry-angle measurement can be performed through directly monitoring the potentiometer signal. © 2010, American Association of Physicists in Medicine. All rights reserved.

The backscattered radiation from the support arm of Varian a-Si EPIDs can affect the accuracy of dosimetric measurements using these devices. In this study the effect of insertion of lead sheets between the EPID and the arm has been investigated for the E-type arms. The optimum lead thickness was determined by comparison of the imager response on and off the arm with increasing lead thicknesses and 2 mm of lead was selected as the optimal thickness considering the reasonable extra weight added to the imager. It reduced the arm backscatter from a maximum of about 6% and 3.5% higher than the off-arm signal in 6 MV and 18 MV beams to about 2% for both energies. On-axis EPID response measurements for different field sizes showed a considerable decrease in arm backscatter with the addition of lead. The symmetry improved for the largest field from about 105% and 103% to 101% and 100% using 2 mm lead. Changing the SDD did not affect the backscatter component more than 1%. The addition of lead decreased the contrast-to-noise ratio and resolution by 1.3% and 0.8% for 6 MV and by 0.5% and 0.4% for 18 MV beams. The root mean square deviation for the difference in EPID central pixel position with and without lead during a whole gantry rotation was one pixel at maximum. In conclusion a 2 mm thick lead layer seems sufficient for acceptable dosimetry results with no major degradation to the routine performance of the imager. © 2010, American Association of Physicists in Medicine. All rights reserved.

PURPOSE: Dosimetric properties of an amorphous-silicon electronic portal imaging device (EPID) operated in a real-time acquisition mode were investigated. This mode will be essential for time-resolved dose verification of dynamic-IMRT and arc-IMRT. METHODS: The EPID was used in continuous acquisition mode, where individual sequential image frames are acquired in real-time. Properties studied include dose linearity and reproducibility. Summed continuous acquisition mode results were also compared to dose results using the well-studied integrated acquisition mode, for example treatment deliveries including dynamic-IMRT and single-arc-IMRT. Comparison was made using percentage dose difference of in-field pixels (pixels >10% of maximum signal). Temporally-resolved EPID response was also compared to that of ion-chamber data for selected points in the deliveries. RESULTS: Using continuous acquisition mode, EPID response was not linear with dose, with response approximately corresponding to 1¿1.5 missed images per irradiation. Reproducibility of EPID response improved with increasing MU. Analysis of the example irradiations revealed summed continuous acquisition mode compared well to integrated acquisition mode, within 2% of maximum dose for more than 95% of in-field pixels. Time resolved EPID data compared well to ion chamber data, with dose increases/decreases overlying each other. CONCLUSION: Continuous acquisition mode is suited for time-resolved dosimetry applications including single-arc-IMRT and dynamic IMRT, giving comparable dose results to the integrated acquisition mode. Linearity and reproducibility should be adequate for clinical applications although caution should be used in low MU work. Time-resolved EPID dose information also compared well to time-resolved ion-chamber measurements. © 2009, American Association of Physicists in Medicine. All rights reserved.

Purpose: To create a series of EPID Monte Carlo dose computation kernels which accounts for observed machine-to-machine variations in EPID response. Method and Materials: Field size response of aS500 and aS1000 imagers are measured for several Varian Cl21-series machines that were dosimetrically matched in a water phantom. Deviations in imager response are attributed to differences in back-scattering materials beneath the imaging panels. Mono-energetic convolution kernels with various backscatter thicknesses are simultaneously created by sub-dividing a thick back-scattering slab into multiple sub-slabs and using the EGSnrc LATCH bit to score sub-slab kernel contributions. Energy-binned particle fluence incident upon the detector convolved with the imager-specific kernels are used to compute the EPID image. Imager-specific kernels are determined by matching computed and measured EPID field-size response, using the number of sub-slabs as a free parameter. Final kernels are used for Monte Carlo-based pre-treatment and in-treatment EPID dose computations. Results: The EPID imagers on dosimetrically matched accelerators are found to differ. Most, but not all of the deviations appear to be correlated with the imager mounting arm type. The imager-specific kernels matched the field-size response for each imager within 1%, and resulted in dosimetric agreement between measured and computed images for pre-treatment dosimetric verification of IMRT fields. Conclusion: Dosimetric differences between portal imagers on matched accelerators can be accounted for by using computation kernels with differing amounts of back-scattering materials. Kernels for multiple different back-scattering thickness can be efficiently calculated. Resultant imager-specific kernels may be useful for efficient pre-treatment and in-treatment Monte Carlo-based EPID dose computations. Conflict of Interest: This work was funded in part by Varian Medical Systems. © 2008, American Association of Physicists in Medicine. All rights reserved.

In this work an amorphous silicon electronic portal imaging device (a-Si EPID) dose prediction model based on the fluence model of the Pinnacle treatment planning system Version 7 (Philips Medical Systems, Madison, WI, USA) is developed. A fluence matrix at very high resolution (0.5 mm) is used to incorporate multileaf collimator (MLC) leaf transmission effects in the predicted EPID images. The primary dose deposited in the EPID is calculated from the fluence using experimentally derived radially dependent EPID interaction coefficients for the open and MLC transmitted fluence components. A spatially invariant EPID dose deposition kernel that describes both radiative dose deposition and optical scatter is convolved with the primary dose. The kernel is further optimised to give accurate EPID scatter factor with changing MLC field size. Model predictions were compared to a-Si EPID images corrected for pixel sensitivity variation, support-arm backscatter and calibrated to dose for various static jaw defined and MLC defined fields and a step and shoot intensity modulated radiotherapy (IMRT) field. For the static fields the model predicts EPID off-axis ratio, penumbral shape as well as inter-leaf leakage. For the IMRT field with a Gamma criteria of 2% and 2 mm, 97.2% of points had a Gamma index less than 1. We found that incorporating the difference in EPID response to open and MLC transmission did not improve the accuracy of the prediction for the IMRT field. The developed model incorporates the effects of MLC design on the dose and therefore should improve the verification of IMRT treatments with EPIDs. © 2008, American Association of Physicists in Medicine. All rights reserved.

Purpose: To develop a software tool for intensity modulated radiotherapy (IMRT) and electronic portal imaging (EPI) dosimetry research. The software tool implements correction factors to a commercially available EPI dosimetry model to account for the change in EPI response to multileaf collimator (MLC) transmitted beam as compared to open beam in IMRT fields. Method and Materials: Software was designed to perform the following tasks: i) Read MLC files from IMRT treatment plans and calculate a matrix of open beam and MLC transmission components. ii) Read portal dose image prediction (PDIP) files exported from the Eclipse treatment planning system (Varian Medical Systems, Palo Alto, CA). iii) Interpolate correction factors from look-up tables for each PDIP based on the MLC transmission components of the corresponding MLC file. iv) Calculate and write a corrected PDIP that can be imported back into the planning system. The software tool was developed using the Microsoft Visual Studio.NET framework with the C¿ compiler. The software tool was validated for functionality and accuracy with a series of test IMRT fields. Results: The software tool correctly calculated the open and MLC beam components for different MLC models and collimator rotations. The corrected PDIP pixel values agreed with manual calculations to within 1% in all cases. Artifacts in the corrected PDIP in regions of high dose gradients were avoided when the MLC transmission matrix was sampled with pixel size ¿ PDIP pixel size. Additional functions available with the software tool include the ability to write the open beam matrix to file, perform arithmetic operations on images, display and save image files, and to plot profile comparisons across images and open beam matrices. Conclusion: A software tool was developed and validated for IMRT and EPI dosimetry investigations. The software tool is being developed further for EPI dosimetry using transit IMRT beams. © 2008, American Association of Physicists in Medicine. All rights reserved.