DEVELOPING QUALITY ASSURANCE PROCEDURES FOR GATED VOLUMETRIC MODULATED ARC THERAPY IN STEREOTACTIC ABLATIVE RADIATION THERAPY by Francis Viel B.Eng., Universit? Laval, 2011 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in The Faculty of Graduate and Postdoctoral Studies (Physics) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) March 2014 ? Francis Viel, 2014 ii Abstract The purpose of this project is to develop a quality assurance (QA) procedure for gated volumetric modulated radiation therapy (VMAT) in stereotactic ablative radiation therapy (SABR) for liver cancer treatments and investigate the gating parameters for acceptable plan delivery in terms of the dose to a moving volume and treatment delivery time. 10 patient plans for VMAT SABR liver were created using the Eclipse? treatment planning system (TPS). The plans were then transferred to a CT-scanned (computed tomography) Quasar? phantom (i.e. a water-equivalent, geometrically simplified representation of a patient) and delivered on a TrueBeam? linear accelerator using a 10FFF (flattening filter free) beam and Varian?s real-time position management (RPM) system for respiratory gating. Two kinds of breathing patterns were used: free breathing (FB) and an interrupted (~5 s pause) end of exhale coached breathing (CB) pattern. Ion chamber and Gafchromic? film measurements were acquired for a gated delivery while the phantom moved under the described breathing patterns and a non-gated, stationary phantom delivery. The gate window was set to obtain a range of residual target motion from 2-5 mm. All gated deliveries have been shown to be dosimetrically equivalent to the static deliveries with differences in point dose measurements under 1% and average gamma (2%, 2 mm) agreement above 98.7%. Comparison with the treatment planning system resulted also in good agreement, with differences in point dose measurements under 2.5% and average gamma (3%, 3 mm) agreement of 92%. The use of a CB pattern increases significantly the duty cycle compared with free breathing and allows for shorter treatment times. Gated VMAT treatments have been delivered successfully to a motorized phantom. FB patterns contain considerable variability and it is difficult to achieve acceptable iiiresults even with very small gate windows. However, a CB pattern combined with a sufficiently small gate resulted in acceptable dose distributions that can be delivered in a reasonable amount of time. iv Preface The work presented in this thesis has been conducted at the British Columbia Cancer Agency ? Vancouver Centre and is using recordings of patients? breathing traces. Ethics approval has been granted by the University of British Columbia ? British Columbia Cancer Agency Ethics Board, certificate number H12-02725. v Table of Contents Abstract .................................................................................................................... ii Preface..................................................................................................................... iv Table of Contents..................................................................................................... v List of Tables.......................................................................................................... vii List of Figures ....................................................................................................... viii List of Abbreviations .............................................................................................. ix Acknowledgements ................................................................................................. x Dedication ............................................................................................................... xi Chapter 1. Introduction ........................................................................................... 1 1.1 Radiation therapy ....................................................................................................1 1.1.1 Overview..........................................................................................................1 1.1.2 Linear accelerators ..........................................................................................2 1.1.3 Delivery techniques .........................................................................................4 1.2 Motion management................................................................................................5 1.2.1 Accounting for motion during treatment planning ............................................6 1.2.2 Breath-hold ......................................................................................................7 1.2.3 Gating ..............................................................................................................8 1.2.4 Tracking.........................................................................................................11 1.3 Quality assurance..................................................................................................11 1.3.1 QA of linear accelerators ...............................................................................12 1.3.2 QA for respiratory gating ...............................................................................13 1.4 Thesis objectives...................................................................................................13 1.4.1 Hepatocellular carcinoma ..............................................................................13 1.4.2 Gated volumetric modulated arc therapy.......................................................14 1.4.3 Study objective ..............................................................................................15 Chapter 2. Materials and Methods........................................................................ 16 2.1 Methods.................................................................................................................16 2.2 Moving phantom....................................................................................................17 vi 2.3 Breathing patterns .................................................................................................18 2.4 Simulation and treatment of gated deliveries ........................................................19 2.5 Dosimetry ..............................................................................................................20 2.5.1 Gafchromic film..............................................................................................20 2.5.2 Ion chamber...................................................................................................23 2.5.3 Gamma analysis............................................................................................24 Chapter 3. Results and Discussion ...................................................................... 26 3.1 Film measurements...............................................................................................26 3.1.1 Film to film comparison..................................................................................26 3.1.2 Comparison with the treatment planning system...........................................30 3.2 Ion chamber measurements..................................................................................31 3.3 Treatment time ......................................................................................................32 3.4 Dose rate dependence ..........................................................................................35 Chapter 4. Conclusion........................................................................................... 38 4.1 Summary ...............................................................................................................38 4.2 Future work ...........................................................................................................39 4.2.1 Imaging..........................................................................................................39 4.2.2 Moving platform .............................................................................................40 4.2.3 Sensitivity tests..............................................................................................40 References ............................................................................................................. 43 Appendices ............................................................................................................ 47 Appendix A: Raw data.......................................................................................................47 Appendix B: Procedures ...................................................................................................49 B.1 Gated VMAT QA Procedure..............................................................................49 B.2 Film Analysis Procedure....................................................................................55 vii List of Tables Table 2.1 Summary of the plans? characteristics..................................................... 16 Table 3.1 Gamma analysis results of the gated deliveries compared to the static delivery. .................................................................................................. 28 Table 3.2 Gamma analysis results of the gated and static deliveries compared to the treatment planning system................................................................ 29 Table 3.3 Percentage difference in dose from ion chamber measurements. .......... 31 Table 3.4 Average ion recombination factor Pion during delivery of plan #9. ........... 36 Table 3.5 Dose rate effect and reproducibility of film measurements...................... 37 Table A.1 Gamma analysis percent of agreement per plan. ................................... 47 Table A.2 Treatment times per plan........................................................................ 48 viiiList of Figures Figure 1.1 Schematic of the head of a linear accelerator........................................... 3 Figure 1.2 Schematic of the different target volumes................................................. 6 Figure 1.3 Schematic of the four-dimensional computed tomography (4DCT) and the corresponding maximum intensity projection (MIP). ................................. 7 Figure 1.4 Two gating methods: (a) amplitude gating with a 5 mm gate window on a sinusoidal motion and (b) phase gating between phases 30% and 70%.10 Figure 2.1 Picture of the Quasar phantom. .............................................................. 17 Figure 2.2 Sample breathing patterns with a 2 mm gate window............................. 19 Figure 2.3 Film calibration curve. ............................................................................. 21 Figure 2.4 Schematic top view of the Quasar phantom during film measurements. 22 Figure 2.5 Schematic top view of the Quasar phantom during calibration of the ion chamber.................................................................................................. 24 Figure 3.1 Typical example of a film to film comparison. ......................................... 27 Figure 3.2 Gamma analysis results.......................................................................... 28 Figure 3.3 Gamma analysis results shown with box plots........................................ 29 Figure 3.4 Percentage difference in dose from ion chamber measurements........... 32 Figure 3.5 Average treatment time per plan............................................................. 33 Figure 3.6 Average duty cycle per plan.................................................................... 34 Figure 4.1 Anterior BEV of a static 4 x 8 cm2 field defined by the MLC with some leaves blocking a few millimetres, used as a first sensitivity test. ........... 41 Figure 4.2 BEV of a simple field delivered as an arc with the collimator at 45? for a second sensitivity test. ............................................................................ 41 Figure 4.3 First sensitivity test results. ..................................................................... 42 Figure B.1 Schematic showing how to cut the films. ................................................ 51 Figure B.2 Schematic showing how to orientate the film in the phantom insert. ...... 51 Figure B.3 Schematic showing how to scan the films. ............................................. 53 ix List of Abbreviations 4DCT Four-Dimensional Computed Tomography AAA Anisotropic Analytical Algorithm AAPM American Association of Physicists in Medicine ABC Active Breathing Control BEV Beam?s Eye View CB Coached Breathing CT Computed Tomography CTV Clinical Target Volume FB Free Breathing FFF Flattening Filter Free GTV Gross Tumour Volume HCC Hepatocellular Carcinoma IGRT Image Guided Radiation Therapy IMRT Intensity-Modulated Radiation Therapy ITV Internal Target Volume kV Kilovoltage Linac Linear Accelerator MIP Maximum Intensity Projection MLC Multileaf Collimator MU Monitor Unit MV Megavoltage OBI On-Board Imager PTV Planning Target Volume QA Quality Assurance RPM Real-Time Position Management? SABR Stereotactic Ablative Radiation Therapy TPS Treatment Planning System VMAT Volumetric Modulated Arc Therapy x Acknowledgements I would like to thank my main supervisor, Dr. Richard Lee, for his close supervision, his availability and his helpful answers to my many questions. I am also thankful to my co-supervisors, Dr. Ermias Gete and Dr. Cheryl Duzenli, for their insightful comments on my work. Thank you to Dr. Hugo Bouchard from the University of Montreal for lending me his code for film analysis. Thank you to the entire Medical Physics department at the BC Cancer Agency ? Vancouver Centre for making it an enjoyable place to work and especially to all those who showed me their clinical work. I am grateful to have received funding from the Natural Sciences and Engineering Research Council of Canada and from the Varian Research Collaborations Program. I give my special thanks to all the residents and staff members of Green College for keeping this community lively, welcoming and simply amazing. It has been a great pleasure to meet these extraordinary people and share with them startling ideas and remarkable friendship. My final thanks go to my parents, my siblings and my partner who have always been very supportive and caring. xi Dedication Pour Danielle et Jean 1 Chapter 1. Introduction 1.1 Radiation therapy 1.1.1 Overview Radiation therapy is the use of ionizing radiation for the treatment of malignant (cancerous) and some benign diseases. During cancer treatment, the tumour cells are being irradiated by ionizing radiation such as X-rays which produce damage to their DNA (both directly and indirectly via free radical production) and can lead to cell death. Unfortunately, the same is true for healthy cells surrounding the tumour. In this respect, the main goal of radiation therapy is to enhance cell kill in tumours while minimizing it in healthy tissues. In general, attaining this goal requires conformal dose distributions, precise alignment, patient immobilization and motion management. In conventional external beam radiation therapy, highly conformal dose distributions are obtained using superposition of multiple fields, alignment is done using positioning lasers and tattoos, immobilization uses masks and vacuum cushions and motion is accounted for with an added margin to the target volume during planning. The challenge in motion management arises when trying to account for respiratory motion, which can result in up to 4 cm of displacement in the abdomen1, affecting mostly lung, breast and liver cancer treatment. Respiratory motion management will be discussed in section 1.2. Several delivery techniques have been developed to improve the conformity and precision of dose distributions as well as decrease normal tissue toxicity. These include intensity-modulated radiation therapy (IMRT) and volumetric modulated arc therapy (VMAT) for increased conformity and image guided radiation therapy (IGRT) for increased precision in target localization. Stereotactic ablative radiation therapy (SABR) can achieve very high dose deposition in tumours with a sharp dose fall-off 2 to minimize normal tissue toxicity. These techniques will be discussed in more details after a brief review of the equipment used for external radiation therapy. Some fundamental concepts of radiation therapy, mainly radiobiology2, physical interactions of x-rays with matter3 and radiation dosimetry4, will not be described in this thesis as they are well documented in the literature. The readers who would appreciate more information on these subjects are encouraged to lookup these references2-4. 1.1.2 Linear accelerators In external beam radiation therapy, linear accelerators (linac) are commonly used to produce X-rays and deliver the treatment5. Typical linacs operate using a waveguide to accelerate electrons to relativistic speeds. The high energy electrons (6 ? 18 MV, i.e. megavolts) are directed to hit a target of high atomic number (Z) material to create X-rays by bremsstrahlung and characteristic photon production. Before exiting the gantry head of the machine, X-rays go through filtration and collimation as shown in figure 1.1. A flattening filter gives a homogeneous fluence by attenuating the center part of the forward peaked beam intensity relative to the edges. Two ion chambers inside the gantry head monitor the output of the machine and the flatness and symmetry of the X-ray beam. The collimation system includes two sets of jaws that can produce rectangular fields up to 40 cm x 40 cm at isocenter (100 cm from the electron target). A multi-leaf collimator (MLC) allows customized beam shaping with high precision as individually motorized leaves can move in and out of the field. The gantry is allowed to rotate to a full 360? around the isocenter providing the possibility for treatment at different beam angles. Radiation delivery during gantry rotation is also an option for a fast delivery. Usually, linacs are also equipped with imaging devices. Electronic portal imaging devices are detectors on an extendable arm that can be extended behind the couch 3 in the path of the megavoltage (MV) beam used for treatment. The direct use of the MV beam for imaging has an interesting advantage since it doesn?t require additional dose, but the images do not provide a good soft tissue and bone contrast due to the large Compton photon scatter component. Modern linacs have kilovoltage (kV) imagers as well. For example, Varian?s TrueBeam? (Varian Medical Systems, Palo Alto, CA) has an on-board imager (OBI) consisting of two additional arms on the side of the gantry: an X-ray tube and a detector. It allows snapshots orthogonal to the treatment beam during treatment as well as full cone beam computed tomography (CT). Other systems include Brainlab?s ExacTrac? (BrainLAB AG, Heimstetten, Germany) consisting of two pairs of fixed X-ray tubes and detectors installed in the treatment room to provide precise positioning assistance. Figure 1.1 Schematic of the head of a linear accelerator. To conclude, modern linear accelerators have many assets, from fluence modulation to adjunct imaging, allowing for a wide variety of treatment techniques. Target Flattening filter Primary collimator Ion chambers Multileaf collimator Secondary collimator (jaws) Electrons X-rays 4 1.1.3 Delivery techniques As mentioned above, IMRT and VMAT are treatment techniques that improve the dose distribution conformity. Intensity-modulated radiation therapy (IMRT) uses dynamic collimation of the beam at different static angles to obtain a highly conformal 3-dimensional dose distribution. Five to seven fields with highly irregular fluence patterns across the field are usually delivered using a multi-leaf collimator. Due to the complexity resulting from the displacement of individually motorized leaves, inverse planning is typically used for IMRT6-8. In conventional ?forward planning? where fields are added to the plan, the MLC leaves are manually shaped and the resulting dose distribution is checked for dose constraints in a ?trial and error? way. Inverse planning on the other hand starts with the specification of the dose constraints and a fluence map is optimized by the computer software. MLC leaf positions are then automatically generated to obtain the optimized fluence map. Volumetric modulated arc therapy (VMAT) uses a continuous modulated delivery with gantry rotation to form a complete arc9. Variable MLC leaf position, gantry speed and dose rate provide VMAT with a conformal delivery equivalent or superior to IMRT, but with a decreased treatment time and monitor units (MU)10. Inverse planning is also necessary for this technique to optimize all variables. In contrast to conventional radiation therapy where the total dose is divided into multiple fractions to maximise normal tissue repair, stereotactic ablative radiotherapy (SABR) is a technique which involves using highly focused megavoltage photon beams to deliver a high dose in a small number of fractions (hypo-fractionation) to treat small cancerous lesions in the body. However, SABR requires small margins for acceptable normal tissue toxicity. Thus, immobilization and image-guidance (IGRT) is very important for SABR in order to attain precision on the order of a millimetre to allow for reduced margins. For cranial immobilization, a personalized moulded mask is attached to the treatment couch. The mask is a more comfortable alternative to the former invasive way of immobilizing, where bolts were screwed to the skull and attached to a frame. In addition, IGRT uses kV imaging inside the 5 treatment room to assist with patient positioning. The acquired images are used to align the bony anatomy or internal markers with the position used during treatment planning. In addition to the aforementioned techniques, there have been recent developments allowing for higher dose rates and reduced treatment time. As mentioned in the description of linear accelerators, the radiation beam is modified by the insertion of a flattening filter before collimation to get a more uniform beam profile. Recently, the effect of removing this flattening filter is being investigated11, 12. The flattening filter free (FFF) mode can be used for small tumours since FFF beams are reasonably flat for small fields. It is also possible to use FFF fields to treat larger tumours with techniques like IMRT and VMAT since these techniques manipulate the fluence patterns as an integral part of the technique. FFF mode could allow reduction of the dose outside the field by reducing the scatter from the head of the linac. Without the attenuation due to the flattening filter, much higher dose rates can be achieved. It has been shown that treatment times for SABR VMAT can be considerably reduced using the FFF mode (3 to 4 times faster) with dosimetrically equivalent plans13. In brief, there exist several techniques to improve the conformity and the precision of treatment delivery to spare healthy tissue. However, conformity and precision become more difficult to achieve when the target is in motion. Target motion is therefore an additional aspect to take into consideration. 1.2 Motion management Accounting for motion is an important aspect of radiation therapy as a moving target has an increased uncertainty in its position. Different methods can be employed to account for motion depending on its nature and amplitude. For most cases, a bigger volume that encompasses all of the target motion is delineated during planning to ensure that the target receives the prescribed dose. However, when the motion is considerably large due to respiration, the adverse side effects may increase from treating a larger volume of healthy tissue around the target. For such cases, it may 6 be desirable to use either breath-hold, respiratory gating or tracking as other means of motion management. 1.2.1 Accounting for motion during treatment planning The International Commission on Radiation Units and Measurements (ICRU) defines different volumes used in treatment planning14. The gross tumour volume (GTV) is identified from the CT images based on the visible extent of the tumour. Then, a margin is added to the GTV to include microscopic spread of the disease, forming the clinical target volume (CTV). To account for motion, an internal margin is added to the CTV to create the internal target volume (ITV). A last margin accounting for set-up errors forms the planning target volume (PTV). These definitions are illustrated on figure 1.2. Figure 1.2 Schematic of the different target volumes. Advancements in imaging provide planners with tools to assess the dimension of the motion and precisely delineate the ITV. Four-dimensional computed tomography (4DCT) is an imaging technique used during treatment simulation. The technique uses the Real Time Position Management (RPM) system (Varian, Medical Systems) to track the breathing motion during a CT scan. The RPM uses an infrared camera mounted in the treatment room with a reflective block that is placed on the chest of the patient. The camera detects the block?s motion in the three spatial directions. The 4DCT reconstruction algorithm then bins the different CT slices to the corresponding phases of the breathing cycle resulting in different three-dimensional CT sets as a function of the phase15. PTV CTV GTV ITV 7 For treatment planning purposes, a maximum intensity projection (MIP) can be created from the 4DCT. MIP images are a single CT set formed by taking the maximum intensity value for each corresponding voxel over the different phases of the 4DCT. For example, a moving target will appear bigger in the MIP showing the complete extent of the GTV during respiration, as illustrated on figure 1.3. The MIP is a useful tool for treatment planning since the ITV can be drawn directly on the MIP instead of having to add an internal margin to the CTV. Creating the ITV from the MIP adds asymmetric, patient specific and target specific margins which are more representative of the motion observed during simulation. Figure 1.3 Schematic of the four-dimensional computed tomography (4DCT) and the corresponding maximum intensity projection (MIP). The first four images from the left illustrate a 4DCT with four bins (phases 0%, 25%, 50% and 75%) of a simple volume moving up and down. Although the increased target volume ensures a high dose to the tumour it also increases the dose to normal tissue and may result in an increase of adverse side effects. Accounting for respiratory motion to reduce the PTV has thus been a major concern in radiation therapy and some techniques have been developed to address this issue, including breath-hold techniques, respiratory gating and tracking1, 16. 1.2.2 Breath-hold As a significant source of internal motion is breathing, a simple method to reduce this motion is to have the patient perform a breath-hold during irradiation. Breath-holds can be performed voluntarily or using a modified spirometer to control the patient?s breathing pattern (active breathing control or ABC17). A spirometer, a CT0 CT25 CT50 MIP CT75 8 device that measures airflow, can be modified for ABC to use a balloon catheter or a valve to completely block the air inlet forcing the patient to hold his breath with a reproducible amount of air in the lungs. Another commonly used breath-hold technique is called deep inspiration breath-hold (DIBH). Advantages of holding the breath at deep inspiration include organs at risk protection from changes in internal anatomy and improved reproducibility. After three normal breathing cycles, the patient is asked to take a deep breath and hold it for 10 seconds allowing enough time for the delivery of a typical field or for a CT during simulation. For the delivery of an IMRT field, a breath-hold of 20 seconds or two breath-holds are required due to the increased amount of monitor units1. After a breath-hold, the patient is free to recover with a few cycles. For all breath-hold techniques, patient selection and eligibility depends strongly on their ability to hold their breath and their comfort during breath-hold. Another respiratory motion management tool, although not a breath-hold technique per se, is called abdominal compression. It uses a stereotactic body frame with a plate compressed against the abdomen to force shallow breathing. It is mostly used for immobilization in SABR for lung and liver cancer when tumour motion is above 5 mm18. 1.2.3 Gating Respiratory gating accounts for motion by restricting the radiation exposure to a specific portion of the breathing cycle and was first introduced by Ohara et al19. The technique involves assessing the breathing trace of the patient during treatment and using it to trigger interruptions of the beam. When the breathing motion moves out of a certain window, signifying that the tumour has moved outside the treatment volume, the beam delivery is paused until it comes back within that enabling window, as shown in figure 1.4. 9 The breathing trace detection is typically achieved with external surrogates. For example, the RPM system described previously uses a reflective block (i.e. the surrogate) placed on the chest of the patient. It is important when using external surrogates to know the correlation between the surrogate motion and the internal tumour motion. Internal fiducial markers such as 2 mm-diameter gold spheres inserted near the tumour can also be used as surrogates. Such markers are detected using additional kV-imaging during treatment. Although internal markers give directly the tumour motion, external surrogates are often preferred since they are non-invasive and do not require additional dose for imaging. In respiratory gating there are two methods to trigger the beam holds: amplitude gating and phase gating, as illustrated on figure 1.4. Amplitude gating sets an upper and a lower limit in the position of the surrogate to enable the beam. The range in amplitude constrained by those upper and lower limits defines the gate window. Phase gating sets limits in the phase of the breathing cycle and enables beam on at a specific part of the cycle, regardless of the absolute position of the surrogate. The two methods have been compared and neither technique is significantly superior to the other20, 21. Phase gating can be advantageous if the breathing trace is regular and periodic. However, when irregularities in the breathing trace are introduced, the system doesn?t recognize as easily the phase of the cycles as the period can vary. Amplitude gating has the advantage of restricting the exposure to a certain position of the target. The disadvantage resides when thresholds are barely reached, resulting in a short and undesired interruption of the beam. 10 Figure 1.4 Two gating methods: (a) amplitude gating with a 5 mm gate window on a sinusoidal motion and (b) phase gating between phases 30% and 70%. The beam is on only when the position is within the enabling window (illustrated here with the green and red lines). The downsides of both techniques can be mitigated by using respiratory coaching. Telling the patients when to breathe in and breathe out can improve the stability of the period. Showing the patients their real-time breathing trace can also help keeping the end of exhale (or end of inhale) position within the amplitude window. It has also been shown that the use of audio-visual feedback reduces the treatment time of gated radiation therapy22. In addition, performing breath-holds during gating a) b) 11 can result in reduced treatment times compared with free breathing23. Breath-holds are also easier to perform with visual feedback. Amplitude gating works better when using breath-holds since these interrupt the periodicity required for phase gating and can remain in the gating window. Although a greater fraction of patients can comfortably achieve respiratory gating compared to breath-hold or abdominal compression, it considerably increases the treatment time due to the many interruptions of the beam. 1.2.4 Tracking Real-time tumour tracking involves repositioning the treatment beam as to follow the target during delivery. It can be performed using a linear accelerator mounted to a robotic arm or by changing the MLC leaf positions so that the radiation field follows the PTV during a respiratory cycle. Tracking is by far the most complex motion management technique since it involves real-time imaging and tumour identification, tumour motion anticipation and adaptation of the dosimetry. Although tracking can eliminate the need for an ITV, the real-time imaging contributes to an increase of the dose during treatment. Tumour motion anticipation is necessary since there is a time delay between tumour identification and repositioning the beam. This delay is due to the image processing and motion of the MLC (or robotic arm) and is typically in the order of tenths of a second. Because of the irregularity of free breathing it is difficult to accurately predict the motion more than 0.5 seconds ahead. It is recommended that the delay should be kept under 0.5 seconds for that reason1. With increasing complexity in motion management and reduction of margins come more opportunities for errors and the need for appropriate control of quality. 1.3 Quality assurance Quality assurance (QA) is the development and realization of protocols in order to assess and ensure the precision and accuracy of a machine or a technique. They 12 include periodic checks (daily, weekly, monthly, and annually) and corrective actions. QA is a very important aspect of radiation oncology physics as it is essential to have well functioning equipment and precise techniques for clinical purposes. 1.3.1 QA of linear accelerators Many different checks are required to be performed on a linear accelerator before allowing its use for treatment. Recommendations on the tests, tolerance values and action levels are presented in the AAPM Task Group 142 report24. Examples of daily checks are output constancy, laser localization, collimator size and safety checks such as the door interlock. Monthly QA includes dose rate output and beam profile constancy, light and radiation field coincidence and couch position and gantry angle indicators. For SABR treatment machines, the same tests are performed although different tolerance levels are recommended. For instance, the coincidence of collimator, gantry, and couch axes with the isocenter is recommended to be within 1 mm for a stereotactic machine rather than within 2 mm for other machines24. For IMRT and VMAT, the constancy and the accuracy of the MLC leaves position need to be tested. The alignment of the leaves and the gaps between them can be affected by gantry rotation due to gravity. A common IMRT test uses a ?picket fence? pattern to check the leaves relative positions and alignment. For VMAT, Ling et al. proposed tests to verify the accuracy of dynamic MLC position and the ability to vary dose-rate, gantry speed and MLC leaf speed25. Those tests consist of irradiating different parts of a film with the same dose with different combinations of dose rate, gantry speed and leaf speed. For IMRT and VMAT, patient-specific QA is also recommended. It consists of delivering a patient plan to a phantom to perform dosimetrical verification using a measurement device. Another option is to perform independent dose calculation (e.g. Monte Carlo simulations26) to compare with the predicted doses from the TPS. 13 The dose comparison can be done using plan files from the TPS or actual linear accelerator log files collected after running the patient plan without the patient present. 1.3.2 QA for respiratory gating In addition to testing the functionality of the devices used for motion management, there are various patient-specific QA procedures when dealing with respiratory motion. Since free breathing is subject to intra and inter-fraction variability, it is strongly recommended to train the patient before and during simulation to achieve reproducible breathing motion. Reproducibility is key to ensure consistency between simulation and treatment. For gating using external surrogates, the correlation between internal and external motion needs to be well known. Since this correlation may change over the course of treatment, it is also suggested to verify the tumour position using kV imaging at the start of every fraction, especially for hypofractionated treatment such as SABR because it requires small margins and missing the target would significantly decrease tumour dose coverage and increase normal tissue toxicity. A dynamic phantom that can reproduce internal motion with realistic breathing patterns is useful for patient-specific verifications. Oliver et al.27 developed a Monte Carlo program to simulate segmental IMRT delivery to a moving phantom. Lee et al.28 have tested gated IMRT with different gate window widths using film measurements on a phantom undergoing sinusoidal motion. These investigations provide an assessment of the effect of residual motion within the gate which is essential prior to treatment. 1.4 Thesis objectives 1.4.1 Hepatocellular carcinoma The technique development and testing in this study can be applied to any site requiring gating as a motion management tool. However, it was initially motivated by 14 a large clinical trial developed at the British Columbia Cancer Agency for the treatment of hepatocellular (liver) carcinoma (HCC). HCC is a significant global health problem with an estimated 748,300 new cases diagnosed in 200829. Globally, it is the sixth most common cancer and is the third most common cause of cancer death29. In the United States, the 1-year survival rate is below 50%30. It is well established that the liver moves significantly as a patient breathes creating difficulty in treating the target volume to the desired dose while simultaneously keeping an acceptable level of normal tissue toxicity1, 31-34. Due to this problem, liver SABR treatments have been highly restricted to a small number of patients having very small lesions. It is anticipated that the results of this project will ultimately lead to improvements in treatment for a significant number of HCC patients who would previously not have been eligible for SABR. This will be made possible by the development of gated VMAT for SABR liver treatment. The use of VMAT has been validated for treatment of HCC with conventional fractionation35, 36. The addition of respiratory gating as motion management will possibly allow for hypo-fractionated (SABR) treatments of larger lesions. 1.4.2 Gated volumetric modulated arc therapy To compensate for the increased treatment time of respiratory gating, the idea of coupling gating with a fast delivery technique like volumetric modulated arc therapy is being investigated37-40. Qian37 and Nicolini38 showed the feasibility of gated VMAT and the dosimetric accuracy of deliveries to a static phantom under different gating periods. However, using respiratory gating with intensity modulation, like IMRT or VMAT, presents an additional challenge as dose gradients are no longer limited to the edges of the field. In those cases, tumour and organs at risk are likely to move through high gradients even within the gate. For this reason it is important to also consider the residual motion for dosimetric measurements. Li39 and Choi40 evaluated the geometric accuracy of gated VMAT using kV imaging during treatment on a dynamic phantom and real patients. 15 1.4.3 Study objective Overall, the objective of this present work is to develop QA procedures for gated VMAT for SABR treatment of liver cancer and investigate the gating parameters for acceptable plan delivery. The use of a dynamic phantom allows dosimetric verification on a moving target including the effect of residual motion within the gate. Optimization of the gating parameters, mostly gate width (thus allowed residual motion) and breathing pattern, was done regarding delivering dosimetrically accurate plans with a reduced treatment time. Where the geometric accuracy of gated VMAT has been investigated using a dynamic phantom39, 40, no one has shown the dosimetric accuracy of gated VMAT while considering the residual motion with a moving phantom. Only gated IMRT has been verified with film measurements on dynamic phantoms undergoing sinusoidal motion27, 28. However, for the development of a more rigorous QA protocol we propose the addition of ion chamber measurements as well as the use of real patient breathing traces (including a coached breath-hold) for our verification. 16 Chapter 2. Materials and Methods 2.1 Methods 10 patient plans for VMAT SABR liver treatments were created using the Eclipse? Treatment Planning System (TPS). The verification plans were then transferred to a CT-scanned Quasar? phantom (Modus Medical Devices Inc., London, ON) and delivered on a TrueBeam? linac using a 10 MV FFF (flattening filter free) beam and Varian?s Real Time Position Management System (RPM) for respiratory gating. The Eclipse Anisotropic Analytical Algorithm (AAA) was used for dose calculation. Table 2.1 presents a summary of each plan?s characteristics. Detailed step by step procedures are presented in appendix B. Table 2.1 Summary of the plans? characteristics. All plans had a prescribed dose of 300 cGy and maximum gantry speed of 4.8 ?/sec. Field size Plan # PTV size X Y MU Arc angle Approximate dose rate (cm3) (cm) (cm) (MU) (?) (MU/min) 1 60 6.7 6.7 740 180 -> 315 950 2 378 13.3 11.0 1121 90 -> 180 1200 3 2388 16.5 19.5 957 90 -> 181 1000 4 154 8.8 8.6 757 180 -> 180 600 5 280 11.7 12.3 1183 180 -> 180 970 6 871 15.7 15.7 591 100 -> 181 600 7 186 10.2 10.4 925 180 -> 252 700 - 1500 8 942 16.8 14.3 969 100 -> 181 1000 9 216 11.3 12.0 957 179 -> 181 770 10 1401 16.5 17.9 774 180 -> 181 400 - 1000 Ion chamber and Gafchromic? film measurements were acquired for a gated delivery while the phantom mimicked a breathing patient. The same plan was delivered under static (non-moving, non-gated) conditions. This enabled a direct film to film comparison of the effect of the gating vs. static conditions with the expectation that the two film deliveries will be identical. It also allows for a comparison without any confounding effects from the dose calculation engine in the TPS. However, the 17 non-gated static deliveries were still compared to the TPS to verify that the treatment was delivered as planned. Gated deliveries were executed using amplitude gating and the gate window was set to obtain a range of residual target motion from 2-5 mm. 2.2 Moving phantom For this study, the Quasar respiratory motion phantom was used and is shown in figure 2.1. The Quasar phantom has a roughly elliptical cross section with an interchangeable cylindrical movable insert. This insert moves in the inferior-superior direction in a linear or helical motion according to the breathing trace input, simulating the internal organ and tumour motion. Different inserts can be used for ion chamber or film measurements and with different densities to simulate lung or normal tissue (wooden or acrylic inserts, respectively). The RPM block is placed on a moving platform reproducing the chest movements in the anterior-posterior direction during breathing. Figure 2.1 Picture of the Quasar phantom. The insert containing the film moves in the inferior-superior direction and the top platform moves in the anterior-posterior direction for the RPM block. Movable insert RPM block Film Inf-Sup Ant-Post 18 Since the relation between the motions of the insert and of the upper platform is set arbitrarily, a ratio of 1:1 was chosen in order to display the breathing motion and gating parameters on the RPM system at the same scale as of the motion of the insert. Fixing this ratio can only be done manually by setting the maximum amount of motion of the insert to 2 cm, corresponding to the fixed maximum amount of motion of the upper platform. 2.3 Breathing patterns Two kinds of breathing patterns were used: free breathing (FB) and an interrupted (~5 s pause) end of exhale coached breathing (CB) pattern. An example of both the FB and CB patterns is shown in figure 2.2. Breathing patterns were acquired during simulation using the RPM system. For FB, the patients were only instructed to breathe normally. During CB, a screen was mounted on the simulation couch to display the real-time breathing pattern to the patient. The patients are asked to hold their breath for around 5 seconds at end of exhale in a specified gate window, inhale, and repeat the 5 seconds hold at each exhale. Providing real-time visual feedback helps the patients to reproduce accurately the breath-hold in each cycle within the same window. Out of 14 patients, 11 were able to perform the desired CB pattern. Of the three who failed, two were diagnosed with lung cancer which could explain the difficulty to breathe in a certain way and the other seemed to have difficulties hearing the instructions. 19 Figure 2.2 Sample breathing patterns with a 2 mm gate window indicated in red: a) free breathing (FB) and b) interrupted end of exhale coached breathing (CB). The maximum position corresponds to exhale and the minimum to inhale. 2.4 Simulation and treatment of gated deliveries Simulation was performed using a GE Lightspeed 16 CT scanner. A 4DCT scan of the Quasar phantom was taken with the film insert. The verification plans were created on the CT 50 phase which corresponds to maximum exhale. Treatment was performed using a TrueBeam? linac at 10 MV FFF in QA mode. However, since the RPM system does not allow setting or modifying the gating parameters in QA mode, the plans were first opened in treatment mode to record a reference breathing curve and set the gate parameters. When the system was a) b) 20 brought back into QA mode, the RPM system takes a few moments to relearn the breathing pattern and automatically sets the centre of the gate window at average maximum exhale. The resulting gate window cannot be modified and its position can vary depending on which portion of the breathing trace was used during the learning period. In brief, there is no way of controlling the position of the gate window in QA mode and it is set automatically by the RPM system. For this reason, the position of the centre of the gate window was recorded for every gated delivery. In addition to the automatic setting of the gate window, the RPM system often takes longer to ?learn? the CB pattern as it expects a more regular breathing, even if the reference curve is recorded with a CB pattern. To account for this uncertainty in the position of the gate during comparison with the TPS, regular CT scans were also performed with the ion chamber insert at different positions of displacement (every 1 mm from maximum exhale to 5 mm towards inhale). The sensitive volume of the ion chamber was contoured on each of these CT sets. The verification plans were also transferred to these CT sets and the average dose to the volume of the cavity was calculated. Comparison of the ion chamber measurements with the TPS can then be performed with the correct position of the insert corresponding to the position of the centre of the gate window noted during treatment. 2.5 Dosimetry 2.5.1 Gafchromic film Film measurements were performed using Gafchromic EBT3? film (International Specialty Products, Wayne, NJ, USA). Gafchromic film measurements are preferred for 2D dose distribution comparison since they provide very high resolution. For this reason they are often used for IMRT and VMAT verification. Films are easy to insert in phantom although care must be taken with respect to the film orientation at all times. The major downside is the delay prior to analysis due to the post-irradiation darkening of the film. The manufacturer recommends waiting at least 24h before scanning the film. Scanning was performed using an Epson? Expression 10000 XL 21 flatbed scanner following a validated protocol developed by Bouchard et al.41. RGB images were collected in reflective mode with 16 bits per color channel and a spatial resolution of 150 dpi corresponding to a pixel size of 0.17 x 0.17 mm and saved in tiff format. Film calibration and analysis were performed using Bouchard?s protocol and software41, which uses the red and blue channels. First, strips of uniformly irradiated film are used to characterize the uniformity of the scanner bed. All subsequent scans can then be corrected for scanner inhomogeneity. For every scan, the optical density of an unirradiated piece of film is subtracted to obtain the net relative optical density. The calibration curve is created based on pieces of film of 2 x 3 cm2 irradiated with a known dose by a standard 10 x 10 cm2 field on a solid water phantom at a depth of 10 cm. A typical calibration curve is shown in figure 2.3. Figure 2.3 Film calibration curve. 22 From figure 2.3 it can be seen that better dose resolution is obtained for doses below 400 cGy. For this reason, the prescription dose was reduced to 300 cGy for all plans. The other characteristics of the plans have been verified on the TPS and only the dose rate was decreased from the original plans. The MLC leaves positions and gantry speeds are kept the same. For each patient plan, six film measurements were performed: coached breathing with 2 and 5 mm window, free breathing with 2 and 5 mm window and without gating, and one static delivery without gating (CB-2mm, CB-5mm, FB-2mm, FB-5mm, FB-NonGated and static). All film measurements were done on pieces of 7 x 11 cm2 placed in the appropriate movable insert in the Quasar phantom. The phantom is aligned so that the isocenter is positioned in the middle of the film when at maximum exhalation, as shown in figure 2.4. Figure 2.4 Schematic top view of the Quasar phantom during film measurements. 23 2.5.2 Ion chamber While film measurements are good for relative dosimetry as well as comparisons of 2D dose distributions, they are not very accurate for absolute dose measurements as we have observed a difference of up to 20% with the TPS when comparing absolute dose. Therefore, an ion chamber was used to obtain a point dose measurement at which the film measurement can be normalized. Ion chamber measurements were performed using a Farmer-type ion chamber with an active volume of 0.6cc (PTW-Freiburg, N23333-1875). The ion chamber was placed in a movable insert of the Quasar phantom that only permitted the chamber to be positioned at a distance of 1.5 cm left of the isocenter. Despite this shift, the point of measurement was kept within the high dose region of the dose distribution in all cases. It is necessary to avoid placing the chamber near an expected dose gradient since the chamber is not completely stationary even during very narrow gates (residual motion). For each patient plan, ion chamber measurements were performed with the same conditions tested with film (CB-2mm, CB-5mm, FB-2mm, FB-5mm, FB-NonGated and static). For each of these conditions, at least two consecutive measurements were taken. Ion chambers are known to depend on ambient conditions like temperature, pressure and humidity. Instead of calculating correction factors to account for this, a cross calibration can be made. After each set of measurements, ion chamber cross calibration was done by delivering a 6?6 cm2 open field to a static chamber from the left of the phantom (gantry at 90?), as shown in figure 2.5. The average of three consecutive measurements was used. The dose at that point is then calculated with the TPS and is used for calibration. 24 Figure 2.5 Schematic top view of the Quasar phantom during calibration of the ion chamber. 2.5.3 Gamma analysis The gamma analysis is a dosimetric method for comparing dose distributions using a metric that considers dosimetric and geometric agreement between dose distributions. During the analysis, every pixel of a reference dose distribution is attributed with a gamma index based on the minimum dose difference and distance difference with every pixel of the compared dose distribution, up to a specific threshold of dose difference (?DM) and the distance to agreement (?dM). In other words, the gamma index (?) of a reference point (rr,Dr) is defined as the minimum value of the following function, ?r, calculated for every point of position rc and dose Dc in the compared distribution: 25 where ?r is the distance between the reference and compared point (i.e. |rr ? rc|) and ?D is the dose difference between the two points (i.e. Dc(rc) ? Dr(rr))42. When the gamma index for a pixel is above one it is said to fail the test. The percentage of agreement is then the percentage of pixels that pass the test. For example, in IMRT dose verification, an acceptable agreement with gamma at (5%, 5 mm) may be 95%, signifying that 95% of the pixels can find a corresponding pixel within 5% difference in its value in a 5 mm radius. When performing gamma analysis, it is possible to obtain different results depending on which dose distribution is used as a reference. Typically, more significant results are obtained when choosing the distribution with the less noise as the reference. A noisy reference would result in unrealistically high percentage of agreement as each pixel in the compared distribution would be able to employ the noise to find a pixel within tolerance. In this study, gamma analysis was used to compare the film measurements with the TPS calculated dose maps as well as to compare the gated deliveries with the static non-gated deliveries. The parameters were chosen to be (3%, 3 mm), (2%, 2 mm) and (1%, 1 mm) for all the pixels (threshold = 0%) of a 6 x 10 cm region of interest which was the same size or smaller than the dose distribution for all cases. Gamma analysis was chosen because the percentage of agreement gives a quantitative measure that can be used to set tolerance values for QA procedures. 26 Chapter 3. Results and Discussion 3.1 Film measurements 3.1.1 Film to film comparison Film to film comparison is advantageous as it isolates the effects of gating by not having to consider positioning errors and any confounding effects from the dose calculation engine in the TPS. Figure 3.1 shows a typical film to film comparison for plan 4 with a prescribed dose of 300 cGy delivered with coached breathing and 2 mm gate width compared to static conditions. In this case, and in most cases, the dose distributions appear to be very similar and no major differences can be seen between the different gating conditions and the static deliveries. The relative difference does not exceed 5% and no apparent shift or blurring of the dose is being observed compared to the static deliveries. Table 3.1 and figure 3.2 present the gamma analysis percentage of agreement averaged over all the ten plans (raw data can be found in appendix A). Gamma analysis (3%, 3 mm) exhibits an average percentage of agreement over 99.9% while (2%, 2 mm) yields greater than 98.7% agreement. This suggests dosimetric equivalence of all gated deliveries. Figure 3.3 shows the same data represented with box plots. It can be seen that there is no significant difference in the percentage of agreement between the different gating conditions. Furthermore, the non-gated deliveries under a free breathing motion resulted in a lower agreement: 93% and 81% for gamma (3%, 3 mm) and (2%, 2 mm) respectively. These results put some perspective to the percentage of agreement obtained for the gated deliveries. 27 Figure 3.1 Typical example of a film to film comparison. The dose maps for plan 4 with prescribed dose of 300 cGy with coached breathing with 2 mm gate width (a) and static (b) conditions are shown with the X and Y profiles (c) on the central axis. 28 Table 3.1 Gamma analysis results of the gated deliveries compared to the static delivery. Average of the percentage of agreement over all plans with range in brackets. Breathing pattern Gate width 3%, 3mm 2%, 2mm 1%, 1mm (mm) (%) (%) (%) CB 2 99.9 ? 0.1 (99.10 - 100.00) 99.0 ? 0.4 (95.68 - 99.98) 89 ? 2 (74.60 - 97.32) CB 5 99.9 ? 0.1 (99.11 - 100.00) 98.9 ? 0.4 (95.75 - 99.90) 87 ? 2 (80.87 - 97.80) FB 2 99.94 ? 0.03 (99.76 - 100.00) 99.3 ? 0.4 (96.24 - 99.93) 91 ? 2 (71.25 - 97.77) FB 5 99.89 ? 0.06 (99.36 - 99.98) 98.7 ? 0.7 (92.66 - 99.82) 83 ? 4 (59.32 - 94.48) FB Non gated 93 ? 2 (82.61 - 99.07) 81 ? 3 (66.90 - 94.61) 56 ? 3 (44.14 - 72.11) 0102030405060708090100CB - 2mm CB - 5mm FB - 2mm FB - 5mm FB - NGPercent of agreement (%)3%,3mm2%,2mm1%,1mm Figure 3.2 Gamma analysis results. Average of the percentage of agreement over all plans. 29 99.299.499.699.8100CB-2mmCB-5mmFB-2mmFB-5mm3%, 3 mmPercentage of agreement (%)949698100CB-2mmCB-5mmFB-2mmFB-5mm2%, 2 mmPercentage of agreement (%)60708090CB-2mmCB-5mmFB-2mmFB-5mm1%, 1 mmPercentage of agreement (%) Figure 3.3 Gamma analysis results shown with box plots. Points are drawn as outliers if they are larger than q3 + 1.5(q3 ? q1) or smaller than q1 ? 1.5(q3 ? q1), where q1 and q3 are the 25th and 75th percentiles, respectively. The plotted whiskers extend to the most extreme data values that are not outliers. Note that the three plots are displayed using different scales. Table 3.2 Gamma analysis results of the gated and static deliveries compared to the treatment planning system. Average of the percentage of agreement over all plans with range in brackets. Breathing pattern Gate width 3%, 3mm 2%, 2mm 1%, 1mm (mm) (%) (%) (%) CB 5 92 ? 1 (86.52 - 96.62) 81 ? 2 (72.08 - 90.32) 53 ? 2 (43.53 - 66.24) Static -- 93 ? 2 (83.11 - 99.51) 84 ? 3 (70.03 - 96.35) 55 ? 2 (45.36 - 67.59) 30 3.1.2 Comparison with the treatment planning system Even though film to film comparison was performed in order to avoid confounding effects of the TPS and to isolate the effects of gating, comparison to the TPS is still desirable to ensure the treatments are delivered as planned. However, since absolute dosimetry is difficult to achieve with high precision using film, relative doses were compared with the treatment planning system. The static dose maps and the CB ? 5 mm were then normalized relative to the point dose where the ion chamber measurements were taken. Not all breathing patterns and gate windows were tested against the TPS since they all had good agreement with the static deliveries and it was assumed that if the static was also in good agreement with the TPS there was no need for comparing all plans with the TPS. However, CB ? 5 mm was compared with the TPS to verify this assumption. To account for the lack of control in the position of the gate window discussed in section 2.4, a shift (usually around 1 mm) was applied in the inferior-superior direction on the TPS dose distribution. This shift corresponds to the average position of the insert within the gate window and is usually the centre of the gate position noted during treatment. For some cases with FB ? 5 mm where the maximum exhale is fairly consistent, the average position of the insert becomes roughly at a quarter of the gate window instead of the centre because the motion rarely exceeds the consistent maximum exhale. For these special cases, it is interesting to note that the residual motion was only half of the gate window. In other words, a consistent maximum exhale will lead to a residual motion of only 2.5 mm when the gate width is set to 5 mm. Table 3.2 shows the gamma analysis results for the CB ? 5 mm and static deliveries compared to the TPS. Gamma analysis (3%, 3 mm), (2%, 2 mm) and (1%, 1 mm) gave agreement of 93%, 84% and 55% on average, respectively, between the static delivery without motion and the TPS. Very similar results were obtained for the CB ? 5 mm, as expected. Although the agreement is lower than with the film to film 31 comparison, these results still represent a good agreement considering several sources of uncertainty in the film analysis. One source of uncertainty is related to the change in optical density of an unirradiated piece of film over a few months time. This change results in an increase in the background optical density that must be subtracted from the optical density of the irradiated film. After making this subtraction, the available range of net relative optical density is much decreased and it results in even lower dose resolution for doses higher than 300 cGy. It is supposed that this increase in background optical density could be related to the aging of the film or of the scanner. A second scanner bed homogeneity correction matrix had to be created approximately a year after the first patient plan was verified. To reduce uncertainties during film analysis it is thus recommended to perform calibration and homogeneity correction frequently. In addition, it is known that the Eclipse AAA algorithm is not very accurate in regions of electronic disequilibrium such as at field edges43. This could also explain some of the discrepancy seen in the film/TPS comparison. 3.2 Ion chamber measurements Ion chamber measurements are a good complement to film as they are more accurate for measuring absolute dose. Point dose measurements of gated deliveries were compared to the point doses calculated with the TPS and to the static point dose measurements. The results are summarized in table 3.3 and shown in figure 3.4. Table 3.3 Percentage difference in dose from ion chamber measurements. Breathing pattern Gate width Compared to static Compared to TPS (mm) (%) (%) CB 2 0.28 ? 0.07 (0.04 - 0.78) 1.2 ? 0.2 (0.36 - 1.94) CB 5 0.46 ? 0.14 (0.03 - 1.09) 1.4 ? 0.2 (0.24 - 2.57) FB 2 0.34 ? 0.12 (0.01 - 1.36) 1.2 ? 0.2 (0.30 - 2.29) FB 5 0.68 ? 0.16 (0.12 - 1.71) 1.3 ? 0.3 (0.04 - 2.59) Static -- -- 1.0 ? 0.2 (0.15 - 1.50) 32 00.511.5CB - 2mmCB - 5mmFB - 2mmFB - 5mmPercentage difference compared to staticPercentage difference (%) 00.511.522.5CB - 2mmCB - 5mmFB - 2mmFB - 5mmStaticPercentage difference compared to TPSPercentage difference (%) Figure 3.4 Percentage difference in dose from ion chamber measurements. The box plots parameters are the same as for figure 3.3. The ion chamber measurements were within 2.5% of the planned dose as calculated by the TPS in all cases and 1.3% on average which shows very good dosimetric agreement. Furthermore, no significant difference can be seen when comparing the results between the different breathing patterns and gate windows. 3.3 Treatment time Although respiratory gating produces dosimetrically equivalent deliveries to static, it significantly increases the treatment time due to the interruptions of the beam. The number and duration of the interruptions are highly dependent on the breathing trace and the gating parameters. Figure 3.5 shows the average treatment time for different breathing patterns and gate windows (the individual values are presented in appendix A). These times start from the moment that the therapist presses the ?Beam On? button and stop when all the MUs have been delivered. They do not include patient setup and coaching or the time necessary to the RPM system to ?learn? the breathing pattern. 33 1112765313675050100150200250300350400450CB - 2mm CB - 5mm FB - 2mm FB - 5mm Non-gatedTreatment time (s)Plan 1Plan 2Plan 3Plan 4Plan 5Plan 6Plan 7Plan 8Plan 9Plan 10Average Figure 3.5 Average treatment time per plan. The overall average for all plans is displayed in yellow with its value on the graph. As one could expect, treatment time increases when decreasing the gating window, simply because a shorter portion of the breathing cycle allows for the beam to be on. With the free breathing pattern, delivery takes on average twice as much time when reducing the gate width from 5 mm to 2 mm. Using the coached breathing pattern with a 5 second breath-hold at end of exhale can reduce the treatment time by more than a factor of two compared to free breathing with a 2 mm window. It can be seen from figure 3.5 that the coached breathing pattern can also reduce the variability in treatment time that results from the high variability of the free breathing patterns. Not only the free breathing patterns differ from patient to patient but they can also vary for the same patient when comparing different treatment sessions. This is shown by the error bars on figure 3.5 showing greater treatment time variability per patient for FB ? 2 mm. For six out of ten plans, the beam-on time was also recorded allowing the calculation of the duty cycle defined as the ratio of the beam-on time over the treatment time. It is interesting to note that the beam-on time for a gated delivery is often longer than the delivery time of a non-gated delivery. This is due to the extra time it takes for the dose rate to ramp up every time the beam comes on after an interruption. This ramp 34 up takes around 0.6 seconds as measured from the trajectory log files. The added effect is then larger when there are many interruptions and when the individual beam-on windows are short, as is the case for FB ? 2 mm where the beam-on time can take up to 25% longer than the corresponding non-gated delivery. On the other hand, using CB ? 5 mm resulted in beam-on times identical to the non-gated treatment times. Figure 3.6 shows the average duty cycle per plan for different breathing patterns and gate windows. 607647290102030405060708090100CB - 2mm CB - 5mm FB - 2mm FB - 5mmDuty Cycle (%)Plan 5Plan 6Plan 7Plan 8Plan 9Plan 10Average Figure 3.6 Average duty cycle per plan. The overall average for all plans is displayed in yellow with its value on the graph. To verify that the dose reduction to 300 cGy does not have an effect on the treatment time, some of the original plans were also delivered without reducing the dose. The non-gated treatment times were observed to be exactly the same whether the dose was reduced or not. For the gated treatment times, the differences between the original and the reduced plans were within the variability observed within a same plan. As a matter of fact, it happened that the delivery of the original plan was shorter than the reduced plan, as the gated treatment times depend mostly on the breathing trace. When considering treatment time, it is important to note that a fast delivery is not only for convenience but also has potential dosimetric implications. After maintaining 35 a consistent breathing pattern for a longer time, a patient can experience discomfort and start squirming, which can increase the residual motion without modifying the external surrogate motion. Reducing the treatment time also means reducing the time when the target moves thus reducing the overall residual motion. Treatment time is therefore an important factor to consider when optimizing gating parameters. To conclude in terms of gating parameter optimization, coached breathing with a 5 to 10 second breath-hold at end of exhale displays some advantages compared to free breathing. In addition to result in shorter treatment times, CB reduces the amount of residual motion within the gate window. Regardless of the width of the window, the short breath-hold during CB keeps the target to a relatively fixed position. Therefore, even if a 2 mm window physically limits the residual motion to 2 mm, a consistent CB pattern will have very little residual motion within a 5 mm window. Furthermore since the CB ? 5 mm can be delivered faster, reducing the overall residual motion, it has been chosen as the optimal setting for gated treatment. 3.4 Dose rate dependence As mentioned in chapter 1, VMAT uses a variable dose rate as the gantry rotates around the patient. The maximum dose rate and the variability in dose rate could potentially have a significant impact on the measurements. Thus, the question of dose rate dependence of both film and ion chamber needs to be addressed. In the case of the ion chamber, a variable dose rate can affect the ion collection efficiency. The correction for ion chamber collection inefficiency is well established and is a part of the TG-51 protocol as the recombination correction factor, Pion, and is calculated for pulsed beams as follows: where VH is the operating voltage of the ion chamber, VL is a reduced bias voltage (typically half of VH), MHraw and MLraw are the raw reading using VH and VL, respectively. 36 For a representative patient plan in this study (plan #9), ion chamber measurements were performed with an applied bias of 300 volts (which is the operating voltage) and 150 volts in order to estimate the average recombination factor per delivery. The calculated values of Pion are presented in table 3.4. Table 3.4 Average ion recombination factor Pion during delivery of plan #9. Breathing pattern Gate width Pion Difference with calibration (mm) % CB 2 1.005 1.0 CB 5 1.010 0.5 FB 2 1.015 0.0 FB 5 1.008 0.7 Calibration 1.015 -- Since a cross calibration was performed using a dose rate similar to the average dose rate of the delivery, the effect of the dose rate should be mitigated to some extent. In fact, table 3.4 shows that Pion for all measurements is within 1% with Pion measured during the calibration. As for the dose rate dependence of the film, it has been well established that gafchromic films are dose rate independent as it is believed that the chemical changes in the active layer of the film should not change as the dose per pulse or the frequency of pulses changes. Tests have been performed for conventional dose rates for up to 1000 MU/min and doses up to 4 Gy and showed no difference in the optical density of the EBT3 film44. However, with the recent development of the flattening filter free mode dose rates can go up to 2400 MU/min, as is the case for VMAT SABR. To address this concern, we performed a simple test where we irradiated some pieces of film with known doses at different dose rates on a cubic water equivalent phantom. Some pieces were irradiated in the exact same conditions also to assess the reproducibility of film measurement. Table 3.5 shows that the dose rate has no 37 significant impact on the dose measured by the film and that its effect is contained within the uncertainty due to reproducibility of measurements. Table 3.5 Dose rate effect and reproducibility of film measurements. Dose rate Measurements Average Standard deviation (MU/min) (cGy) (cGy) (cGy) 800 304 311 309 308 4 1200 304 317 305 309 7 2400 313 304 307 308 5 38 Chapter 4. Conclusion This thesis presented the development of quality assurance procedures for gated volumetric modulated arc therapy for stereotactic ablative radiation therapy of liver cancer at the BC Cancer Agency ? Vancouver Centre. Chapter 1 introduced the reader to different aspects of radiation therapy with a focus on the current motion management techniques. In chapter 2, a description of the method and materials used in this work was presented. Ion chamber and film measurements results are given in chapter 3. 4.1 Summary Respiratory gating is a motion management technique that allows margin reduction to decrease normal tissue toxicity and acute side effects of radiation therapy. When coupled with volumetric modulated arc therapy and flattening filter free mode, treatment times can be kept reasonable despite the reduced duty cycle due to the beam interruptions during gating. Small margins are necessary for acceptable normal tissue toxicity of hypofrationated SABR treatments. Gated VMAT thus allows treatment of hepatocellular carcinoma patients previously ineligible for SABR. Assuring high quality of treatment is an essential part of radiation therapy, especially when developing new techniques. Thus a new QA protocol for gated VMAT SABR has to be developed before the technique becomes clinically available. Over 120 gated arcs were successfully delivered in the work described in this thesis and have been shown to be dosimetrically equivalent to the static deliveries with differences in point dose measurements (using an ion chamber) under 1% and average gamma (2%, 2 mm) agreement above 98.7% (using gafchromic film). 39 Comparison with the treatment planning system resulted also in good agreement, with differences in point dose measurements under 2.5% and average gamma (3%, 3 mm) agreement of 92%. The comparison between the gated and static deliveries is a good way of isolating the effects of gating and demonstrating dosimetrical equivalence of the technique. The Quasar? phantom is a useful tool for dosimetric measurements of an object in motion while reproducing real patient breathing patterns. Thus, film to film comparison using the Quasar? phantom can be used for patient-specific quality assurance of gated VMAT plans. In respect to the second thesis objective in optimizing the gating parameters, it has been found that the use of a coached breathing with a 5 to 10 seconds breath-hold at end of exhale can significantly decrease the treatment time compared to free breathing. Since no significant difference in dose distribution was observed between a 2 mm and a 5 mm gate window, a 5 mm window could be chosen to keep the treatment time shorter. In addition, the breath-holds in the CB pattern help reduce residual motion even in a 5 mm window. Therefore, the coached breathing with breath-holds and a 5 mm gate window have been chosen for clinical implementation of gated VMAT SABR. 4.2 Future work 4.2.1 Imaging However, before gated VMAT can be clinically implemented, development of imaging techniques during treatment needs to be performed. With the RPM system only, gating would achieve tumour control in the assumption that there is no shift between the anterior-posterior motion of the chest and the internal tumour and organ motion. An on-board imager (OBI) like on the Varian TrueBeam? allows taking orthogonal images during treatment and positioning internal markers. It is possible to take an image with the OBI at every breathing cycle just before the beam would come on during gating. Thus, the therapist or a fast automated image processing 40 algorithm could stop the beam if the image displays the internal target at a wrong position and a couch shift could be performed to correct the position of the patient. 4.2.2 Moving platform Additional future work includes further testing of gated VMAT with a newly acquired moving platform which can hold any currently used static phantom. This moving platform can also use recorded patient breathing traces as an input to then reproduce the motion to other phantoms or detectors. More dose map measurements can be performed using moving diode or ion chamber arrays. Those tests would add to the robustness of the results discussed in this thesis and would possibly simplify the quality assurance procedure since film dosimetry can be quite complex and tedious and a point dose measurement with an ion chamber gives limited information. 4.2.3 Sensitivity tests In order to assess the sensitivity of the method, one could perform some complementary tests. The first test is a film measurement of an anterior open field with some MLC leaves moved a few millimetres inside the field, as shown in figure 4.1. Static and gated deliveries are then compared to a static open field. Other tests involve increasing the complexity of the previous field by delivering it as a full arc with the collimator at 0? and 45?. The hanging leaves can be positioned so that their individual effects on the dose distribution would not overlap due to the arc. Figure 4.2 shows a beam?s eye view (BEV) of the plan that can be used for the arc at gantry of 180? with the collimator at 45?. Furthermore, a standard patient plan can be modified by introducing some deliberate errors in the MLC motion45. The procedure would then pass or fail the test based on the size of the introduced errors. This test would also be useful in determining acceptable limits for a gamma analysis and establishing tolerance levels for QA procedures. 41 Figure 4.1 Anterior BEV of a static 4 x 8 cm2 field defined by the MLC with some leaves blocking a few millimetres, used as a first sensitivity test. The numbers in yellow on the figure indicate the length of the hanging leaves in millimetres. Figure 4.2 BEV of a simple field delivered as an arc with the collimator at 45? for a second sensitivity test. In this figure, the MLC leaves are hanging of 2 mm. 1234576- 5-1(mm) 42 Even though this is presented as future work, the first sensitivity test (static field) has been performed already and demonstrates some limitations of the gamma analysis. It can be seen from figure 4.3 that even though gamma analysis 3%, 3 mm gives an agreement of 99.2% there are obvious differences in dose that can be seen by eye. Figure 4.3 First sensitivity test results. Gamma (3%, 3 mm) gives 99.2% agreement. If the current method was to be replaced by another detector placed on the moving platform mentioned in section 4.2.2, similar sensitivity tests would also need to be performed. a) Open field b) Errors c) Difference d) Gamma 43 References 1. Keall PJ, Murray BR, Ramsey CR, et al. The management of respiratory motion in radiation oncology report of AAPM Task Group 76. Med Phys. 2006;33:3874-3900. 2. Joiner MC, Van der Kogel A, eds. Basic Clinical Radiobiology. 4th edition ed. CRC Press; 2009. 3. Johns HE, Cunningham JR, eds. The Physics of Radiology. 4th edition ed. Springfield, IL: Thomas; 1984. 4. Attix FH. Introduction to Radiological Physics and Radiation Dosimetry. New York: Wiley; 1986. 5. Karzmark CJ, Numan CS, Tanabe E, eds. Medical Electron Accelerators. New York: McGraw-Hill; 1993. 6. Bortfeld T, B?rkelbach J, Boesecke R, Schlegel W. Methods of image reconstruction from projections applied to conformation radiotherapy. Phys Med Biol. 1990;35:1423. 7. Mageras GS, Mohan R. Application of fast simulated annealing to optimization of conformal radiation treatments. Med Phys. 1993;20:639. 8. Webb S. Optimisation of conformal radiotherapy dose distributions by simulated annealing. Phys Med Biol. 1989;34:1349. 9. Yu CX. Intensity-modulated arc therapy with dynamic multileaf collimation: an alternative to tomotherapy. Phys Med Biol. 1995;40:1435-1449. 10. Otto K. Volumetric modulated arc therapy: IMRT in a single gantry arc. Med Phys. 2008;35:310-317. 11. Georg D, Kn??s T, McClean B. Current status and future perspective of flattening filter free photon beams. Med Phys. 2011;38:1280. 12. Hrbacek J, Lang S, Kl?ck S. Commissioning of Photon Beams of a Flattening Filter-Free Linear Accelerator and the Accuracy of Beam Modeling Using an Anisotropic Analytical Algorithm. Int J Radiat Oncol Biol Phys. 2011;80:1228-1237. 44 13. Reggiori G, Navarria P, Cozzi L, et al. Can volumetric modulated arc therapy with flattening filter free beams play a role in stereotactic body radiotherapy for liver lesions? A volume-based analysis. Med Phys. 2012;39:1112. 14. International Commission on Radiation Units and Measurements. Prescribing, recording, and reporting photon beam therapy (Supplement to ICRU Report 50). ICRU. United States: ICRU; 1999;ICRU Report 62. 15. Low DA, Christensen G, Politte DG, et al. A method for the reconstruction of four-dimensional synchronized CT scans acquired during free breathing. Med Phys. 2003;30:1254. 16. Giraud P, Yorke E, Jiang S, Simon L, Rosenzweig K, Mageras G. Reduction of organ motion effects in IMRT and conformal 3D radiation delivery by using gating and tracking techniques. Cancer / Radioth?rapie. 2006;10:269-282. 17. Wong JW, Sharpe MB, Jaffray DA, et al. The use of active breathing control (ABC) to reduce margin for breathing motion. Int J Radiat Oncol Biol Phys. 1999;44:911-919. 18. Negoro Y, Sasai K, Shibamoto Y, et al. The effectiveness of an immobilization device in conformal radiotherapy for lung tumor: reduction of respiratory tumor movement and evaluation of the daily setup accuracy. Int J Radiat Oncol Biol Phys. 2001;50:889-898. 19. Ohara K, Okumura T, Akisada M, et al. Irradiation synchronized with respiration gate. Int J Radiat Oncol Biol Phys. 1989;17:853-857. 20. Berbeco RI, Nishioka S, Shirato H, Chen GTY, Jiang SB. Residual motion of lung tumours in gated radiotherapy with external respiratory surrogates. Phys Med Biol. 2005;50:3655-3667. 21. Vedam SS, Keall PJ, Kini VR, Mohan R. Determining parameters for respiration-gated radiotherapy. Med Phys. 2001;28:2139. 22. Linthout N, Bral S, Van de Vondel I, et al. Treatment delivery time optimization of respiratory gated radiation therapy by application of audio-visual feedback. Radiotherapy and Oncology. 2009;91:330-335. 23. Berson AM, Emery R, Rodriguez L, et al. Clinical experience using respiratory gated radiation therapy: Comparison of free-breathing and breath-hold techniques. Int J Radiat Oncol Biol Phys. 2004;60:419-426. 24. Klein EE, Arjomandy B, Liu C, et al. Task Group 142 report: quality assurance of medical accelerators. Med Phys. 2009;36:4197-4212. 45 25. Ling CC, Zhang P, Archambault Y, Bocanek J, Tang G, Losasso T. Commissioning and quality assurance of RapidArc radiotherapy delivery system. Int J Radiat Oncol Biol Phys. 2008;72:575-581. 26. Teke T, Bergman AM, Kwa W, Gill B, Duzenli C, Popescu IA. Monte Carlo based, patient-specific RapidArc QA using Linac log files. Med Phys. 2009;37:116-123. 27. Oliver M, Staruch R, Gladwish A, Craig J, Chen J, Wong E. Monte Carlo dose calculation of segmental IMRT delivery to a moving phantom using dynamic MLC and gating log files. Phys Med Biol. 2008;53:N187-N196. 28. Lee SH, Kim KB, Kim MS, et al. Quantitative Analysis of Dose Distribution to Determine Optimal Width of Respiratory Gating Window Using Gafchromic EBT2 Film. J Korean Phys Soc. 2013;62:657-663. 29. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA: a cancer journal for clinicians. 2011;61:69-90. 30. Altekruse SF, McGlynn KA, Reichman ME. Hepatocellular carcinoma incidence, mortality, and survival trends in the United States from 1975 to 2005. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2009;27:1485-1491. 31. Weiss PH, Baker JM, Potchen EJ. Assessment of hepatic respiratory excursion. Journal of nuclear medicine : official publication, Society of Nuclear Medicine. 1972;13:758. 32. Suramo I, Paivansalo M, Myllyla V. Cranio-caudal movements of the liver, pancreas and kidneys in respiration. Acta Radiol Diagn. 1984;25:129-131. 33. Davies SC, Hill AL, Holmes RB, Halliwell M, Jackson PC. Ultrasound quantitation of respiratory organ motion in the upper abdomen. Br J Radiol. 1994;67:1096. 34. Harauz G, Bronskill MJ. Comparison of the liver's respiratory motion in the supine and upright positions: concise communication. Journal of nuclear medicine : official publication, Society of Nuclear Medicine. 1979;20:733. 35. Wang P, Hsu W, Chung N, Chang F, Fogliata A, Cozzi L. Radiotherapy with volumetric modulated arc therapy for hepatocellular carcinoma patients ineligible for surgery or ablative treatments. Strahlentherapie und Onkologie : Organ der Deutschen R?ntgengesellschaft. 2013;189:301. 36. Gong GZ, Bai T, Zhang G, et al. RapidArc combined with the active breathing coordinator provides an effective and accurate approach for the radiotherapy of 46 hepatocellular carcinoma. Strahlentherapie und Onkologie : Organ der Deutschen R?ntgengesellschaft. 2012;188:262. 37. Qian J, Xing L, Liu W, Luxton G. Dose verification for respiratory-gated volumetric modulated arc therapy. Phys Med Biol. 2011;56:4827-4838. 38. Nicolini G, Vanetti E, Clivio A, Fogliata A, Cozzi L. Pre-clinical evaluation of respiratory-gated delivery of volumetric modulated arc therapy with RapidArc. Phys Med Biol. 2010;55:N347-N357. 39. Li R, Mok E, Han B, Koong A, Xing L. Evaluation of the geometric accuracy of surrogate-based gated VMAT using intrafraction kilovoltage x-ray images. Med Phys. 2012;39:2686-2693. 40. Choi K, Xing L, Koong A, Li R. First study of on-treatment volumetric imaging during respiratory gated VMAT. Medical physics. United States: American Association of Physicists in Medicine; 2013;40:040701. 41. Bouchard H, Lacroix F, Beaudoin G, Carrier JF, Kawrakow I. On the characterization and uncertainty analysis of radiochromic film dosimetry. Med Phys. 2009;36:1931-1946. 42. Low DA, Harms WB, Mutic S, Purdy JA. A technique for the quantitative evaluation of dose distributions. Med Phys. 1998;25:656. 43. Kan MWK, Cheung JYC, Leung LHT, Lau BMF, Yu PKN. The accuracy of dose calculations by anisotropic analytical algorithms for stereotactic radiotherapy in nasopharyngeal carcinoma. Phys Med Biol. 2011;56:397. 44. Casanova Borca V, Pasquino M, Russo G, et al. Dosimetric characterization and use of GAFCHROMIC EBT3 film for IMRT dose verification. Journal of applied clinical medical physics / American College of Medical Physics. 2013;14:4111. 45. Mu G, Ludlum E, Xia P. Impact of MLC leaf position errors on simple and complex IMRT plans for head and neck cancer. Phys Med Biol. 2008;53:77-88. 47 Appendices Appendix A: Raw data Table A.1 Gamma analysis percent of agreement per plan. Plan CB - 2mm vs Static CB - 5mm vs Static FB - 2mm vs Static FB - 5mm vs Static 3%,3mm 2%,2mm 1%,1mm 3%,3mm 2%,2mm 1%,1mm 3%,3mm 2%,2mm 1%,1mm 3%,3mm 2%,2mm 1%,1mm 1 99.69 99.69 94.97 99.97 99.90 97.80 99.95 99.88 93.36 99.93 99.63 93.73 2 99.92 99.07 86.97 99.11 95.75 80.87 99.99 99.84 90.13 99.86 98.70 82.72 3 99.99 99.86 91.71 99.97 99.41 84.87 99.94 99.34 92.79 99.95 99.34 89.45 4 100.00 99.98 97.32 100.00 99.90 82.08 100.00 99.91 96.80 99.98 99.82 85.94 5 99.10 95.68 74.60 99.72 97.91 83.05 99.76 98.55 87.65 99.95 99.42 81.94 6 99.99 99.73 89.32 99.98 99.37 86.98 99.82 96.24 71.25 99.36 92.66 63.15 7 99.95 98.76 83.35 99.96 99.33 90.67 99.95 99.26 86.74 99.97 98.83 59.32 8 99.99 99.82 90.15 99.96 99.48 91.36 99.99 99.93 97.24 99.96 99.39 91.97 9 99.97 97.91 87.47 99.97 99.29 87.44 99.97 99.92 97.77 99.96 99.71 94.48 10 99.97 99.71 89.81 99.87 98.75 85.16 99.98 99.81 94.78 99.96 99.35 90.33 FB ? NG vs Static Static vs TPS CB - 5mm vs TPS Plan 3%,3mm 2%,2mm 1%,1mm 3%,3mm 2%,2mm 1%,1mm 3%,3mm 2%,2mm 1%,1mm1 98.82 89.45 63.09 99.51 96.35 56.82 94.19 83.12 55.39 2 87.22 74.69 48.22 97.65 89.76 59.07 91.04 81.49 54.70 3 97.51 90.70 67.42 91.33 76.89 46.78 92.99 79.91 50.11 4 99.07 94.61 72.11 97.90 90.80 65.24 96.62 90.32 66.24 5 -- -- -- 92.51 83.15 51.36 86.52 75.92 51.18 6 92.78 78.34 51.62 79.59 58.91 31.19 94.47 85.22 60.32 7 82.61 66.90 44.14 85.97 56.48 19.32 95.94 86.22 52.00 8 86.96 69.65 44.69 83.11 70.03 47.01 86.64 72.08 43.53 9 96.06 80.51 53.52 83.80 60.35 27.30 90.97 78.66 47.76 10 96.92 88.30 63.19 68.56 50.85 24.93 89.37 78.26 51.57 48 Table A.2 Treatment times per plan. CB - 2mm CB - 5mm FB - 2mm FB - 5mm Non-gated Plan (sec) (sec) (sec) (sec) (sec) 1 81 ? 4 61.0 ? 0.6 340 ? 20 133 ? 15 40.4 ? 0.42 130 ? 8 84 ? 2 370 ? 40 140 ? 5 46.7 ? 0.93 149 ? 13 72.7 ? 0.3 290 ? 20 114 ? 6 46.3 ? 0.34 146 ? 2 77.0 ? 1.0 294 ? 18 173 ? 8 61.8 ? 0.45 125 ? 1 89 ? 2 227 ? 3 131 ? 4 62.0 ? 0.06 70 ? 2 63.0 ? 1.5 127 ? 3 98 ? 2 48.6 ? 0.27 129 ? 11 71.8 ? 1.4 370 ? 30 168 ? 6 50.0 ? 0.08 83 ? 5 57.5 ? 1.3 204 ? 4 115 ? 5 48.3 ? 0.19 98 ? 2 85.3 ? 1.8 270 ? 30 149 ? 7 61.1 ? 0.310 100 ? 2 88.7 ? 1.3 265 ? 4 136 ? 9 62.0 ? 0.0 49 Appendix B: Procedures B.1 Gated VMAT QA Procedure 1. Getting the breathing trace from CT a. On the RPM computer in the simulator control room, open the desired patient. b. Click on Export and note the directory. c. Go to that directory and copy the file (.vxp) on the H drive. 2. Edit the breathing trace a. Open the Quasar Software. b. Click on Navigation ? Wave Editor c. Click on File ? Import d. Select the .vxp file you want to import. e. Delete the undesired portions of the breathing trace by selecting them and pressing the ?Delete? key. Make sure that the position at the beginning of the selection matches the position at the end for no sharp jump after deletion. f. Modify the amplitude so that it is around 1 cm (or previously check the real amplitude on the 4DCT scan). Keep in mind that 15mm on the software scale is actually 10mm of real motion. When importing the .vxp file, the software renormalizes the amplitude so that the lower and higher positions match -15mm and 15mm respectively. g. Adjust the position of the trace so that the maximum is at 15mm. h. On the left side, click on Filter and perform a smoothing operation to remove the noise (simply click on Apply with the default settings: Low Pass, 1 Hz, 60 BPM). i. On the left side, click on Test to verify that the motor will be able to perform the motion correctly, i.e. there is no sharp jump. j. Save the trace as a .qmr file by clicking on File ? Save k. If the trace was acquired during a scan, the couch motion may have induced a linear shift in the trace as a function of time. If desired, run in Matlab the code called ?Couch Correction.m? to get rid of the shift. 3. Create a verification plan a. In Eclipse, open the patient plan you want to transfer to the phantom. b. Copy the plan to a new course named ?Course_QA?. c. Right Click on the plan name and then click on Properties d. Under the ?Dose? tab, change the Prescribed Dose Per Fraction to 300cGy (this prevents the film to saturate using Gafgui) and click on OK. e. Click on Planning ? Create Verification Plan... f. Select ?Course_QA? and click Next. 50 g. Click on Select? then Change Patient? and select the patient ID of the Quasar phantom 4DCT: Patient ID: z20130308 h. Click on the CT50 phase, then OK and then click Next. i. Make sure the boxes are not checked, then click Next twice and then Finish. j. Right Click on the plan name and then click on Properties k. Under the ?General? tab, change the ID (by adding ?_NG? for Non Gated) and uncheck the box saying: ?Use Gated?. l. Right Click on the field (usually ?Field 1?) and then click on Properties m. Under the ?Geometry? tab, change the isocenter position to: X: 9.53 Y: -0.04 Z: 1.89 This brings the isocenter in the middle of the film insert of the phantom. n. Save. Calculate the dose volume. Save again. o. Approve the plan for delivery. p. Export the calculated dose plane: i. Right click on ?dose? on the left and click on Export dose plane. ii. Leave the settings to ?Absolute dose? and 10cm x 10cm. iii. Click Next twice and select the desired export location. Click on Finish. q. Reselect the original patient plan and repeat steps e) to n) (except step k) using the phantom CT with the ion chamber insert: Patient ID: vcc_quasar_chamber (select CT1) Find the dose to the ion chamber under the statistics tab at the bottom. Note it down. You do not need to approve this plan. 4. Cut the pieces of film a. Using gloves when manipulating the film, cut two pieces of 7cm by 11cm. (On a sheet you can fit 6 pieces of 6.8cm by 11cm, as shown in figure B.1) b. Keep a piece of unirradiated film. (usually there is an extra band after cutting the 6 pieces) c. Number the pieces on the bottom right for consistency. 51 Figure B.1 Schematic showing how to cut the films. 5. Measurements a. Set up the phantom on the couch. Do not center it; instead place it more on the right so that the insert is somewhat in the center of the couch laterally, i.e. the couch should then be around -10cm laterally during alignment. This will prevent the couch to be in the way of the gantry after making the isocenter shift. b. Align the phantom using the lasers. c. Apply this couch shift: Lat: -9.53 Long: -1.89 The isocenter should be in the middle of the film insert. d. For film measurements, place a piece of film in the insert by positioning the number as shown in figure B.2. Figure B.2 Schematic showing how to orientate the film in the phantom insert. For ion chamber measurements, place the ion chamber in the ion chamber insert. e. Make sure that the phantom is set to a total range of motion of 2cm. To do so, manually rotate the knob and verify measure the range of motion. Unscrew the knob to modify it if needed. f. To deliver a gated treatment: 11cm # # ## # # Extra band for background 6.8cm ? ??# Film Punch holes 52 i. Open the breathing trace with the Quasar software and press the play button under ?Phantom Control? and ?Oscillation?. ii. Click on the ?Repeat? button so that the trace loops and that the phantom never stops its motion. iii. On the treatment unit, open ?Treatment mode? and open the patient in ?Treat? mode (not ?QA?). This is only to setup the RPM system. Do not deliver the beam or move to requested couch positions. iv. After selecting the gated plan, click on Next on the right hand side screen to set the gating parameters. v. Select ?Amplitude gating?, click Next, select ?Planned period? and click Next again. vi. Set the thresholds to -0.25 and 0.25cm (for a gate window width of 5mm). vii. Click on Record Reference Curve on the bottom and wait for a few cycles before clicking on Stop. viii. Click on Done and Close patient on the left screen. ix. Open the same patient this time in ?QA? mode. ? For ion chamber measurements, I suggest selecting only the gated plan within the course. By selecting the non-gated plan also, it will jump to this plan after the first measurement and you will have to restart the RPM learning phase to make multiple readings. x. Check all boxes except the first one (gantry angle) to acquire current couch positions. Click on Acquire and Apply. xi. After selecting the gated plan on the left, click on the play/pause button on the right hand side screen to start the RPM system. (If the breathing trace has long breath-holds, it may take a while to ?learn? the trace. If so, you can select ?Rotation? on the Quasar software to make it move in a sine wave for a couple cycles and the RPM should be satisfied.) xii. Note down the position of the centre of the gate: 1. Change the scale of the RPM trace display to 0.5 to zoom in. 2. On the Quasar software, change the position until the detected trace on the RPM is at the center of the gate and note that position. 3. Go back in ?Oscillation?, start the motion and check the ?Autoscale? box of the RPM. xiii. You are now ready to deliver the beam. ? For ion chamber measurements, reset the ion collection on the electrometer before pressing ?beam on?. g. To deliver a non-gated treatment: i. Stay in ?QA? mode and select the non-gated plan. ii. In the Quasar software, set the position to 15mm. iii. Deliver the beam. 53 iv. During delivery, pay attention to the dose rate and note down an approximate average dose rate. h. After the ion chamber measurements, exit ?Treatment? mode and enter ?Research? mode to deliver a static square field for cross-calibration: i. 6 x 6 cm square field ii. Gantry 90 deg iii. Collimator 0 deg iv. 10 FFF v. 200 MU vi. Open the MLC vii. Select the approximate average dose rate previously noted. 6. Film analysis a. For film scanning and conversion to dose, follow the step by step instructions described in appendix B.2. ? When scanning the films: ? Make sure that the images are saved in a portrait mode. ? Place your films close to the central axis of the scanner (horizontally centered) as shown in figure B.3. Figure B.3 Schematic showing how to scan the films. b. On the OmniPro Im?RT software, click on File ? Import Data ? Generic ASCII c. Click on Add File(s)? and select the non-gated .opg dose map. Click OK. d. Choose Data Set 1 and set the scale value 300cGy to 100%. e. Repeat steps c) to e) but selecting the gated dose map and Data Set 2. f. Right click both images and click on More ? Smooth ? OK g. Right click on the gated map and click on Correct Origin. ?? Punch holes Film (facing down on the glass) Scanner Lamp Extra band for background ?# 54 h. From the position of the centre of the gate noted before delivery, calculate the shift: shift = (15-position)*2/3 (remember that there is a ratio of 10mm/15mm (or 2/3) between real motion and displayed motion on the Quasar software) i. Enter a Y deviation corresponding to this calculated shift. j. Verify that the Y profiles align well after making the shift. k. Click on Tools ? Options?. Under the ?2D Options? tab, set the Delta Dose and Distance to 2% and 2mm for Gamma analysis. Click on OK. l. Click on Tools ? Calculate Result ? 2 vs 1 to compare the gated map to the non-gated map. m. Click on Tools ? Calculate Result ? Gamma n. Right click on the gamma map and click on Histogram. o. Note the percent of agreement found in the statistics box beside ?Pixels in Ranges: 0 to 1?. p. Click on File ? Save Worspace to save your work. q. Click on File ? Import Data ? TPS Dose Planes r. Click on File? and select the exported dicom RT Dose file. Click OK. s. Choose Data Set 2 and set the scale value 300cGy to 100%. t. Click on Tools ? Calculate Result ? 1 vs 2 to compare the non-gated map to the TPS dose plane. u. Repeat steps n) to p) to get the gamma percent of agreement. v. If desired, you can also compare the TPS dose plane to the gated delivery. 7. Ion chamber analysis a. Repeat step 3.q) to create a verification plan on the CT that contains the chamber positioned at the center of the gate noted during measurement (at step 5.f.xii). ? CT1 is at 10mm, CT2 at 9mm, CT3 at 8mm, etc. (remember that there is a ratio of 10mm/15mm (or 2/3) between real motion and displayed motion on the Quasar software) b. Open the Excel file ?Patient QA.xls?. c. Right click on the sheet tab named ?Original? and click on Move or Copy d. Check the box ?Create a copy? and select ?Before Original?. e. Rename the copy with the patient ID. f. Enter the ion chamber readings and the TPS calculated doses in the yellow cells. g. The green cells should give you the percentage differences. 55 B.2 Film Analysis Procedure 1. Getting started a. Open Matlab and select the directory containing the files Gafgui.m and Gafgui.fig b. Type ?H=Gafgui;? or simply ?Gafgui? in the command window to launch the program. c. The window should appear with a yellowish color, meaning that the default film type is EBT2. You can change the film type under Parameters ? Film type For EBT3 films, simply select EBT2. 2. Averaging Images a. Click on Preparation ? Average images b. Open the five images to average (use the Shift key to select all five). c. After the images are averaged, save the new image wherever you like. (It is easier to save it in the current directory as it will open every time you need to load a file) 3. Homogeneity correction matrix a. To create the homogeneity correction matrix, click on Characterization ? Scanner homogeneity ? Create matrix b. It will ask you if you want to start from the beginning; select ?Yes?. c. Open the image with the horizontal strips to open. d. Enter the number of strips plus one for a blank selection. e. Select your preferred mode of selection (?Free? works well). f. Select each strip plus one blank strip by dragging a rectangle and double-clicking in it. The order doesn?t matter. g. It will ask you if you want to save this step; select ?No?. (If ?yes?, it will save your selection in an .m file that you can load by not starting from the beginning at step (b).) h. After the program is done with the calculation, save the correction matrixes (one polynomial and one filter). I recommend writing ?poly? or ?filter? in the file name. You can close the figures that popped out. 4. Correct images for homogeneity a. Click on Preparation ? Correct image b. Open the image to correct. c. Open the correction matrix. d. Select the method to use corresponding to the correction matrix. e. Save the corrected image. 56 5. Calibration curve a. To create the calibration curve, click on Characterization ? Film calibration ? Create curve b. It will ask you if you want to start from the beginning; select ?Yes?. c. Open the image with the unirradiated film scans. If you don?t have one, open the image with the irradiated film scans. d. Open the image with the irradiated film scans. e. Select ?Yes? to flip the image horizontally. f. Enter the number of films used for the calibration. g. Enter the size of the squares used for the selection (1cm by 1cm works well for pieces of 2cm by 3cm). h. Select each unirradiated film; To make a selection: ? Drag a rectangle to zoom on a piece; ? Double-click on it; ? Drag the fixed-sized square on the ROI; ? Press ?Enter?. (If you don?t have the unirradiated film scans, select N times the unirradiated piece of film in the irradiated film scans image.) i. Select each irradiated film using the above selection steps. j. Select ?Manual? to enter the dose values manually. k. Enter the dose values by respecting the order of the previous selection. (?UM? is French for MU, but you can enter the dose values in cGy.) l. It will ask you if you want to save this step; select ?No?. (If ?yes?, it will save your selection in an .m file that you can load by not starting from the beginning at step (b).) m. Choose the curve that has the lowest uncertainty on NOD and press ?Enter?. n. Save the curve. 6. Convert raw signal into dose maps a. Follow step 4 to correct the image for homogeneity. b. Click on File ? Import ? 48 bits .tif format c. Open the corrected image. d. To flip the image, click on Tools ? Rotate image ? Horizontal mirror e. Click on Data ? Define background f. Enter ?1?. g. If there is a piece of unirradiated film in the image, select ?Region? and follow the above selection steps. If you know the OD value of the background, select ?Manually? and enter the value. h. Click on Data ? Net OD/ROD i. Click on Data ? Dose 57 j. Open the calibration curve. 7. Export the dose map to a file readable by OmniPro I?mRT a. Click on File ? Export ? OPG format b. Select the location you want to save the file.