Volume 32, Issue 7, July 2005
Index of content:
- FIFTY‐FIRST ANNUAL MEETING OF THE CANADIAN ORGANIZATION OF MEDICAL PHYSICISTS AND THE CANADIAN COLLEGE OF PHYSICISTS IN MEDICINE
- CCPM Symposium: Optical Diagnostics and Therapeutics
CCPM Symposium I ‐ 01: Anatomic, functional, and molecular imaging using optical coherence tomography32(2005); http://dx.doi.org/10.1118/1.2030967View Description Hide Description
Optical Coherence Tomography (OCT) is a novel biomedical imaging technique which uses low‐coherence optical interferometry to obtain micron‐scale resolution tomographic images of sub‐surface tissue structure noninvasively. OCT has become a standard diagnostic tool in clinical ophthalmology, and is under investigation for other clinical applications including cancer detection and evaluation of cardiovascular disease. Within the past few years, dramatic technology advances have increased the performance of OCT systems many‐fold, and have also demonstrated the potential for micron‐scale functional and molecular imaging in living systems for the first time. We have developed spectral domain OCT scanners capable of imaging up to several times video rate, and applied them for real‐time two‐dimensional and near real‐time three dimensional imaging in human and small animal models. The applications of this new technology for high‐throughput noninvasive phenotyping and rapid 3D imaging in small animals and developmental biology models is particularly compelling. In addition, we have developed novel functional imaging extensions to OCT which take advantage of the altered spectral content of elastic and inelastically backscattered light to provide enhanced image contrast. These include the first demonstrations of molecular imaging with OCT, in which an imaging form of pump‐probe spectroscopy has been used to image the distributions of genetically expressed proteins with micron resolution in living animals, with sensitivity comparable to multiphoton microscopy.
CCPM Symposium I ‐ 02: Diffuse optical imaging of the neuro‐metabolic‐vascular relationship during brain activation32(2005); http://dx.doi.org/10.1118/1.2030968View Description Hide Description
The ability and interest in functional imaging of the human brain has grown with the advent of positron emission tomography(PET) and functional magnetic resonancetomography (fMRI). These imaging techniques are leading to a better understanding of the healthy, diseased, and injured functioning brain. Diffuse optical imaging is a non‐invasive, portable, and relatively inexpensive method that complements PET and fMRI with the ability to continuously monitor the hemodynamic, metabolic, and possibly neuronal activity state of the brain, and to measure populations of subjects and paradigms not amendable to PET or fMRI. All of these methods are predominantly sensitive to the hemodynamic parameters of the brain which arise from the neuronal and metabolic activity. During this talk, I will discuss the contributions of optical imaging to understanding the relationship between neuronal, metabolic, and vascular activity within the brain. Better knowledge of this relationship will guide the development of better treatments and improve the utility of diagnostic imaging methods.
CCPM Symposium II ‐ 01: Imaging breast tumor tissue in vivo with diffuse light: Tumor tissue characterization and monitoring32(2005); http://dx.doi.org/10.1118/1.2030969View Description Hide Description
Diffuse optical tomography with near‐infrared light has allowed characterization of breast tumortissue with a number of different constituent parameters, which could have relevance for diagnosis and therapy. Multi‐spectral tomography provides quantification of hemoglobin, oxygen saturation, water, and scatterer particle size and density. These parameters are shown for normal and diseased breast tissue, with an eye toward their pathobiological interpretation. The images of tumors present in breast cancer show significant increases in hemoglobin, water and scattering relative to the corresponding normal tissue, and the results of ongoing clinical trials are presented. The scattering particle size is shown to be correlated to the pathologically measured particle sizes in excised breast tissue, and further model‐based interpretation of the scatter signal may yield important structural information at the nanometer level in tissue, as measured macroscopically with NIR tomography.Imaging of fluorescence from tissue is also possible and is presented, along with a demonstration of how the technological design can be altered to allow video‐rate imaging similar to an ultrasound scanner.
32(2005); http://dx.doi.org/10.1118/1.2030970View Description Hide Description
Photodynamic therapy (PDT)‐ the use of light‐activated drugs‐ continues to develop as a viable treatment for solid tumors and dysplasias and for non‐oncologic applications. The underlying optical technologies for light generation, delivery and dosimetry are described. The last, in particular, remains challenging. This, together with the measurement of photosensitzer levels in tissue,tissue oxygenation and the biological effects of PDT, has required development of several different techniques and corresponding clinical and pre‐clinical instruments. These include fiberoptic‐based optical probes, near‐infrared luminescencemeasurement of excited singlet oxygen generated in PDT,measurements of photosensitzer photobleaching, and the use of bioluminescence to assess both tumor (or cell) destruction and gene regulation. Fundamentally new approaches to PDT include the use of ultrafast pulsed lasers for 2‐photon excitation of photosensitizers, the use of ‘metronomic’ (i.e. low dose‐rate) drug and light delivery, and tumor‐specific targeting using phototherapeutic ‘molecular beacons’. Each of these techniques requires multidisciplinary research and development spanning physics, optical engineering, photobiology and clinical specialties.
- Scientific Session 1: Diagnostic Imaging
32(2005); http://dx.doi.org/10.1118/1.2030971View Description Hide Description
Positron Emission Tomography(PET) is becoming more routine in the clinical work‐up of patients prior to cancer treatments. Especially when combined in a single session with X‐ray CT scanning the whole body PET scans are seen as increasingly valuable in delineating viable tumor margins. PET is a relative late arrival on the OncologicalImaging scene, but PET centers were well established in Canada well before MRI was invented. The first PET centre in Canada was established in Canada in 1975 with the “loan” of a 32‐detector instrument from the Brookhaven National Lab in 1975 to a team lead by (the late) Dr. Lucas Yamamoto at the Montreal Neurological Institute. This was followed by others in Hamilton and Vancouver about 1980. All of the early scanners at these centers were designed for brainimaging. Many important innovations in PET methodology have taken place in these centers. When one goes to a PET education seminar these days, introductory lectures usually stress the contributions made at St. Louis or UCLA to PET methodology and instrumentation, and the work of Canadian groups may not be mentioned. This presentation will discuss some of the highlights of the early days of PET in Canada especially the contributions made by those in the McConnell BrainImaging Centre of the MNI.
Sci‐PM Thurs ‐ 02: Spatial resolution in PET and the effect of gamma‐ray interaction depth in block detectors32(2005); http://dx.doi.org/10.1118/1.2030972View Description Hide Description
The spatial resolution in PET is poorer than that of CT or MRI. All modern PET scanners use block detectors, i.e. clusters of scintillation crystals coupled to four photomultiplier tubes(PMTs). Some of the loss of spatial resolution in PET is attributed to the use of block detectors, because a photon that interacts with one crystal in the cluster may be incorrectly positioned, resulting in blurring of the reconstructed image, called the “block effect”. We examined the effect of changing gamma‐ray interaction depth in the scintillation crystals in detectors from the CTI HR+ and GE Advance PET scanners. We postulated that the depth at which the gamma‐ray interacts may contribute to the “block effect” blurring. The “block effect” was measured for both detectors, and it was found to be 1.2 mm for the central crystals and negligible for the edge crystals in the CTI HR+ block. However, it was 0.9 mm in all crystals of the GE Advance detector. In the CTI HR+ detector, a depth dependence on the positioning of the event was observed, as was a dependence on the crystal location (edge vs. centre). In the GE Advance detector, no such dependence was observed. These results suggest that the depth of interaction of an annihilation photon may contribute to the block effect in detectors that use crystals cut from a single scintillation crystal (pseudo‐discrete crystals). In detectors that use discrete crystals no additional blurring as a function of gamma‐ray interaction depth in the detectors was observed.
Sci‐PM Thurs ‐ 03: 3‐D angiography: Effect of the maximum intensity projection on the measurement of artery diameters32(2005); http://dx.doi.org/10.1118/1.2030973View Description Hide Description
Our goal was to determine the impact of the maximum intensity projection (MIP) algorithm on quantitative analysis used for diagnosis and treatment planning of extracranial arterial disease. We performed 3‐D computed rotational angiography (CRA) on 26 consecutive symptomatic patients to verify an internal carotid artery (ICA) stenosis originally found using duplex ultrasound. These volumes of data were visualized using two different post‐processing projection techniques: MIP and digitally reconstructed radiographic(DRR) projection. A DRR is a radiographicimage simulating a conventional digitally subtracted angiogram (DSA), but it is derived computationally from the same CRA dataset as the MIP. By visualizing a single volume with two different projection techniques, the post‐processing effect of the MIP algorithm is isolated. Vessel measurements were made, according to the NASCET guidelines, and percentage stenosis grades were calculated. The NASCET‐type stenosis grades tended to be underestimated on average by 2.4% with the MIP algorithm, although this was not statistically significant (P=0.09). Moreover, the vessel measurements from the MIPs were consistently 0.17 mm larger than those from DRRs (P<0.0001). Thus, when applied to high‐contrast, high‐resolution CRA images, the MIP algorithm slightly increased the apparent dimensions of the arteries. This difference is not clinically relevant, due to the fact that it is only a fraction of the 0.38 mm voxel spacing. Moreover, the MIP dimensions could actually represent the truth, and that it is the sharp decrease near the edge of the intensity profile of the radiographicimages that causes an underestimation of the absolute dimensions.
Sci‐PM Thurs ‐ 04: A comparative study between multi‐station and moving‐table methods with steady‐state free precession32(2005); http://dx.doi.org/10.1118/1.2030974View Description Hide Description
Large field‐of‐view (FOV) imaging techniques, such as the multi‐station and moving‐table techniques, are necessary to image systemic diseases such as peripheral vascular disease and metastases. In the multi‐station technique, the full k‐space is acquired for each station, i.e., at each local FOV, and the images are combined offline. For the moving‐table method, the scanner bed is continuously moved through the local FOV during a single acquisition, creating a single large FOV image. Steady‐state free precession is a pulse sequence capable of rapid imagedata acquisition. This study compares large FOV images of healthy volunteers using both the moving‐table method and the multi‐station technique using an SSFP pulse sequence. In the multi‐station technique, six 32 s‐acquisition's are required to cover the large FOV. For the moving table method, the hybrid k‐space is collected during a single 150 s‐scan, creating a seamless large FOV image. Although the moving‐table method is more time‐efficient than the multi‐station technique, image quality is sacrificed. This quality reduction is due to non‐steady‐state conditions caused by table motion and because the k‐space data is partially sampled. By optimizing this moving‐table SSFP technique and integrating it with a tissue suppression algorithm, it may be possible to perform non‐contrast enhanced MR angiograms of the entire peripheral vasculature. Thus, the technique could provide a non‐invasive and time‐efficient method for producing seamless large FOV images for diagnosis of systemic diseases.
Sci‐PM Thurs ‐ 05: Improving background suppression in magnetic resonance‐guided endovascular therapy32(2005); http://dx.doi.org/10.1118/1.2030975View Description Hide Description
Vascular disease is a leading cause of death in Canada. Endovascular therapy represents a minimally invasive means of treating this disease. The current clinical standard for endovascular treatment uses x‐ray imaging as the modality to visualize the vasculature and devices introduced into the vascular system.Magnetic resonance(MR)imaging is a better modality in terms of patient safety and has potential for use in clinical endovascular therapy. Before that can happen, we must show that we can reliably visualize and track catheters within a slice of tissue under MR guidance. One way of increasing catheter conspicuity is the projection dephaser (PD) method of background suppression.(Dixon et al., MRM, 1986) We propose another approach, in which multiple phase cycles are applied over the slice thickness, such that upon projection into one plane, the background tissue signal adds destructively while the catheter signal is minimally affected. In a 3 Tesla MRscanner, we imaged a 4 French catheter (1.3 mm) in a pork chop phantom with a fast spoiled gradient echo sequence (repetition time/echo time/ flip angle/ slice thickness = 7.5 ms/ 3.6 ms/ 20°/ 70 mm). Visual analysis of acquired images shows that catheter conspicuity is significantly improved over the PD method. Quantitatively, catheter contrast,, in the PD‐suppressed image is 15% (Sc = catheter signal, Sb = background signal). With 35 phase cycles over the slice thickness, C is increased to 44%. These results show that this background suppression technique has potential for use in MR‐guided endovascular procedures.
32(2005); http://dx.doi.org/10.1118/1.2030976View Description Hide Description
DiffusionMRI has become one of the most powerful tools for the detection of acute stroke. The signal attenuation caused by the diffusion process is normally assumed to be exponential. The decay constant which is often called the “apparent diffusion coefficient” (ADC) is measured by 2 or 3 points in clinical applications yielding mono‐exponential decay curves. The diffusion signal attenuation of water molecules on human brain was measured with a certain pulse sequence. The sequence was modified to work over a range of diffusion times and high gradients. The decay was measured precisely for 96 b‐values up to the maximum possible gradient amplitude of 28.8 mT/m. A significant deviation from mono‐exponential behavior was observed consistent to the multi‐exponential model. The NNLS‐diff computational code, using a non‐negative least squares (NNLS) algorithm was developed for the data analysis. The diffusion time dependence of human braintissue was studied for diffusion times between 20 to 53 ms using 16 b‐values ranged from b=0 to the maximum possible b‐value in each case. At all diffusion times there was a diffusion coefficient at approximately 1×10−3 mm2/s and another at about 7×10−5 mm2/s. For some diffusion times a small contribution at about 1×10−2 mm2/s was also detected. Our results were consistent with work of others. However, we observe a small diffusion time dependence for the smallest diffusion coefficient which has not previously been reported. More work is required to identify the source of the new observation.
32(2005); http://dx.doi.org/10.1118/1.2030977View Description Hide Description
Minimally invasive cardiac surgery (MICS) has already been shown to reduce hospital stays, but the full potential is not yet realized, largely due to limitations in pre‐ and intra‐operative visualization inside the closed chest. To address these issues, we are developing the Virtual Cardiac Surgery Platform (VCSP) — a 4D (3D + time), virtual reality model of the patient specific thorax, derived from pre‐procedural images. In this abstract, we discuss the accuracy of our image registration‐based method for deforming geometrical template models of the heart sub‐anatomy (myocardium, right atrium + ventricle, left atrium + aorta, epicardium) to 10 different volunteers (“patients”). The template models are built by manually segmenting a high quality magnetic resonance(MR)image (this template image is an average of 20 acquisitions of the same volunteer, 1.53 mm3 voxels). The template image is mapped to a much lower quality patient image (1.5×1.5×6.0 mm3 voxels) obtained in a clinically feasible manner, by maximizing the normalized mutual information (NMI) between the two images. The resulting global (affine) and local (free form deformation) transformation is applied to one of the four template models to transform it into patient space. The registration accuracy is assessed by comparing the mapped template to the manual segmentation of the patient. On average, the customization process is accurate to within 2.4 ± 0.2 mm, whereas, the difference between two manual segmentations (gold standards) was 1.3 ± 0.2 mm. We believe our method adequately prepares templates for use within VCSP, prior to and during MICS.
Sci‐PM Thurs ‐ 08: Development of a lung tumour model for validating three‐dimensional thoracoscopic ultrasound imaging32(2005); http://dx.doi.org/10.1118/1.2030978View Description Hide Description
Introduction: Potential minimally invasive lungcancer therapies such as brachytherapy will require intraoperative imaging for both instrument tracking and tumour volume measurements for radiationdose planning. Ultrasound(US) is the modality of choice because it is inexpensive, real‐time, portable, and non‐invasive to the patient. Objectives: We have developed a lungtumourmodel, using excised porcine lung and agar tumours, to provide a means of verifying volume measurements of 3D USimages in ex vivo lungtissue.Methods: Spherical tumours were made from agar with diameters of 9.5mm to 25.4mm. The tumours were inserted through incisions on the underside of the excised porcine lung. The lung was placed in a box with ports and the thoracoscopic US probe was inserted through a port for imaging. One observer measured the tumourimage volumes five times, once every two days, using a radial segmentation algorithm with an interslice thickness of three degrees. Results: Both the coefficient of variation (COV) and percent difference decreased as the tumour size increased. The average COV and percent difference were 11.2% and 12.9%, respectively. Conclusions: 3D Thoracoscopic US can be used accurately and reproducibly to measure tumour volumes in vitro.
Sci‐PM Thurs ‐ 09: A semianalytic model to extract differential linear scattering coefficients of breast tissue from energy dispersive x‐ray diffraction measurements32(2005); http://dx.doi.org/10.1118/1.2030979View Description Hide Description
Our research group is focused on determining the potential applications of using x‐ray diffraction signals to diagnosis breast cancer. We have built a custom made x‐ray diffractometer system. Polyenergetic 50 kV beams collimated down to a 3 mm diameter are incident on 5 mm diameter 5 mm thick samples. A cadmiumzinc telluride detector is positioned at an angle θ with respect to the center of the target. We use our semianalytic model coupled with energy dispersive x‐ray diffractionmeasurements to extract differential linear scattering coefficients in units of m−1 sr−1. We optimize our system with as our target since good x‐ray diffraction data is available. A 2‐mm diameter aperture is positioned in front of our detector and the target to detector distance is ≈ 40 cm. We use a root‐mean‐square deviation to measure the overall agreement between our data and the gold standard. We get values of 1.6 and 2.0 m−1 sr−1 for scatter angles of 13° and 16° and values of 3.4, 2.5, 2.1, 2.8, 1.8 m−1 sr−1 for angles 5, 7, 8, 9, and 11. These values are obtained after correcting our data for fluorescence escape and hole tailing. However, the values are nearly the same even if we don't correct the data. At this stage, more optimization is required. Our preliminary results for breast tissue agree well with data measured by Kidane et al., Phys. Med. Biol. 44, 1791–1802 (1999). We intend to correlate the x‐ray diffraction and cellular pathology signals of breast tissue.
- Poster Session and Reception
32(2005); http://dx.doi.org/10.1118/1.2030980View Description Hide Description
The purpose of this project was to design, fabricate and test the data acquisition timing control, precision rotary stage control, and analog data multiplexer unit for a prototype megavoltage computed tomography (MVCT) detector. An 80‐element prototype detector array is made with (element size 0.275 × 0.8 × 1 cm3) scintillators and photodiodes on an arc (radius of 110 cm). In addition to designing and fabricating an in‐house data acquisition system (front‐end integrators, data multiplexer unit, and timing control), a precision rotary stage and its control are added to create a third generation MVCT scanner.Data acquisition is synchronized with radiation pulses from a linac. Response of detector as a function of dose rate was studied by varying the source to detector distance. A narrow slit beam, at five locations, was used to measure the pre‐sampled MTF. Detector signal in open beam was measured for a number of radiation pulses to use the periodogram method for NPS estimation. Using the measured MTF, NPS, and photon fluence impinging on detector,DQE was calculated. Detector response is linear as a function of dose rate, however shows a non‐linear component while measuring the attenuation by solid water due to the polyenergetic spectrum. Beam‐hardening correction is necessary before image reconstruction. MTF at the Nyquist frequency (0.16 mm−1) is 0.48. Zero‐frequency DQE in 6 MV at 21% is higher than any experimental MVCT detector. The basic performance of the prototype detector is satisfactory for producing reasonable low contrast resolution in MVCT images with low dose.
32(2005); http://dx.doi.org/10.1118/1.2030981View Description Hide Description
The objective of this work was to study the feasibility of acquiring mega‐voltage cone‐beam CT (MV CBCT)images of sufficient quality for setup verification from 2 D projection images. The generation of mega‐voltage cone‐beam CT with a standard radiation therapy Primus linear accelerator and an amorphous silicon electronic portal imaging device has been attempted in this work. 2 D projection images of the head section of a Rando phantom were acquired at an interval of 2° from gantry angle of 258° to 102°. Thus 103 2D projection images of the Rando phantom were acquired with 6 MV beam with 1 MU/projection exposure at 300MU/minute dose rate. The estimated dose to the center of the phantom placed at the isocentre of the linac was about 75cGy. The cone angle was only 14.2° and hence for simplicity the images were reconstructed assuming a parallel beam geometry using the IRADON function implemented in MATLAB. High contrast objects are clearly seen on the reconstructed images thus showing the potential of MV CBCT for patient alignment during external beam radiotherapy. From the available image data set, reconstructions were also performed with fewer projections i.e with 52 and 26. Image quality was acceptable with 103 and 52 projections. Reconstruction with Feldkamp cone beam algorithm or iterative algorithm such as Algebraic Reconstruction Techniques (ART) could further improve the image quality.
Po‐Poster ‐ 03: A comparison of intensity‐modulated radiotherapy versus 3D conformal radiotherapy for prostate cancer patients whose rectal reaction to prior conventional radiotherapy was known32(2005); http://dx.doi.org/10.1118/1.2030982View Description Hide Description
Background and purpose: To evaluate the difference in biological and physical endpoints between intensity‐modulated radiotherapy(IMRT) plans and conformal plans for thirteen prostate cancer patients whose rectal reaction to prior radiation therapy was known (five patients experienced Grade 0 late rectal complications, six Grade 1, and two Grade 2). Materials and methods: Conformal plans were generated for each patient with Eclipse version 7.1.59 while the Helios inverse optimization component within Eclipse was applied for computation of the IMRT plans. Physical endpoints applied to compare the IMRT and conformal plans included the target dose statistics (minimum, maximum, and mean dose) as well as dose‐volume histogram results for the normal tissues (rectum, bladder, and femoral heads). Biological endpoints used for comparison of the IMRT and conformal plans were tumour control probability (TCP), normal tissue complication probability (NTCP), and probability of uncomplicated local control (P+). Rectal NTCP was computed for both RTOG Grade 3 or lower complications and severe rectal complications (severe proctitis, necrosis, stenosis, and fistula). Results: On average, IMRT increased the TCP by 2.8% and increased the TDI by 7.7%. The average NTCPs for the Grade 3 or lower rectal complications were 15.1% and 4.4% for the conformal and IMRT plans, respectively. For severe rectal complications the average NTCP decreased from 2.1% (conformal) to 1.2% (IMRT). The average P+ increased from 59.4% (conformal) to 69.9% (IMRT).Conclusions:IMRT was found to reduce the rectal NTCP for RTOG Grade 3 or lower complications to ∼0.36 times the NTCP for the conformal plan.
32(2005); http://dx.doi.org/10.1118/1.2030983View Description Hide Description
A daily test was developed to assure intensity modulated fields are delivered with a high degree of accuracy and precision using the dynamic multileaf collimator delivery mode. Quality assurance of the MLC, particularly for dynamic mode of delivery, requires positional accuracy of individual leafs, as well as gap width defined by opposing leaf pairs. The daily test was designed to be quick, simple to use, and easy to analyze. With the secondary collimating jaws at 25 cm wide and 39 cm long, a 5mm sliding gap over 5 cm at isocenter placed between two static fields 4 cm off axis was measured dosimetrically using Kodak XV film. The sliding gap allows us to test dosimetrically the positional accuracy defined by Varian in motor encoder counts, as a function of beam on time. The speed of the leaves is controlled by the MUs delivered. Relative filmdosimetry with respect to a baseline film obtained after MLC alinment is performed daily and analyzed using RIT software. In order to facilitate the registration of the QA film to the baseline film, a mounting device utilizing the upper wedge block was used. The QA test was tested using beam shaper files with controlled errors introduced and dose differences of 5% are easily detected. In addition to the daily film QA, as part of morning photon output checks a sliding gap output is measured and these results will be presented as well.
Po‐Poster ‐ 05: An evaluation of treatment dose error due to beam attenuation from a carbon fiber table top32(2005); http://dx.doi.org/10.1118/1.2030984View Description Hide Description
The emergence of carbonfiber materials for use in radiation therapy was largely due to its high mechanical strength, low specific density, and its perceived radio‐translucence. These characteristics made it an ideal material for the patient support assembly utilized during treatments. Modern radiation therapy commonly employs beams delivered at oblique angles. With the introduction of carbon fiber table tops the attenuation of the couch is often ignored during treatment planning and there is little effort to avoid intersection of the beam with the table during patient setup. The perception that carbon fiber is relatively radio‐translucent has permitted it to be used while neglecting to consider the effects it may have on the dose to the patient. In this study we have measured the attenuation of the couch under various conditions for 6 and 18 MV photons. We have found dose reductions in phantom of greater than 7%. We further investigate the ability of a commercial treatment planning system (Theraplan Plus) to properly model this effect during the planning stage. Our results show that incorporating the carbon fiber couch in the patient model reduces the dose error to less than 2%. These results reveal that it is worthwhile addressing this real clinical problem in such a manner that it can be routinely considered for all patient treatments. Thus, practical suggestions are proposed for the incorporation of the treatment tabletop into patient treatment planningdose calculations.
32(2005); http://dx.doi.org/10.1118/1.2030985View Description Hide Description
Megavoltage cone‐beam computed tomography (MV‐CBCT) is a volumetric imaging method that can improve patient setup verification techniques. MV‐CBCT utilizes the treatment beam to obtain projections at every 1–2° around the patient. For this to be clinically acceptable, total dose received by the patient from all imaging sessions must be kept to a minimum and typically should be ⩽5% of the prescribed dose. This necessitates the use of extremely low doses (<<1MU) in the acquisition of each projection. At such low dose levels beam spot instability is known to exist and can compromise image quality. The purpose of this work is to quantify the beam spot motion of Siemen's accelerator for a conventional 6 MV beam and 5.4 MV “imaging beam” generated with a beryllium target. This was accomplished by using an a‐Si flat panel detector to image a cone‐beam geometric calibration phantom and using a calibration algorithm to derive the spot motion in reference frames fixed in space and/or attached to the gantry. Motion of the beam spot was observed immediately after beam startup primarily in the gun‐target direction. The maximum fixed motion of the 6MV beam spot was 1.1±0.3 mm and is similar to that observed for the low Z beam (1.2±0.1 mm). However, the beam spot position of the latter stabilized at about 0.5 MU compared to 6 MU for the 6 MV beam and had much less fluctuations once stabilized. The beam spot position of the conventional beam was much less reproducible than the low Z beam.
Po‐Poster ‐ 07: Commissioning of virtual linacs for Monte Carlo simulations by optimizing photon source characteristics32(2005); http://dx.doi.org/10.1118/1.2030986View Description Hide Description
In conjunction with rapidly expanding clinical Monte Carlo(MC) implementation, MC users are faced with the difficult and time consuming commissioning process of their virtual linac. After accurately configuring the treatment head according to manufacturer specifications, a rigorous and extensive benchmarking process is required to ensure that the MC virtual linac produces beams of essentially the same quality as those of the real treatment unit being modeled. Often, even after systematically varying the input parameters over a suitably chosen range, it is found that the shape of the measured profiles cannot be exactly matched. This limitation is attributed to the lack of accurate knowledge of the geometry and materials of some linac components, especially, the flattening filter. We have developed an automatic optimization method that allows a user with an arbitrary linear accelerator to commission a MC dose calculation engine that accurately reproduces the measured output of the accelerator in water. Using a simulated annealingoptimization algorithm our method converges MC dose distributions to experimental measurements by optimizing the weights of particles in a phase space. To achieve this, a phase space is divided up using LATCH variable assignment, each beamlet is transported into a water tank phantom, and the dose deposition is scored separately. Individual beamlet weights are then optimized such that the weighted sum of beamlet dose depositions converge toward our target dose distribution. The resulting beamlet weights are then assigned to all particles in the original phase space where they are incorporated into all future simulations.