Index of content:
Volume 33, Issue 6, June 2006
- Imaging Scientific Session: 330A
- Image Segmentation and Registration
33(2006); http://dx.doi.org/10.1118/1.2241449View Description Hide Description
Purpose: Gated PET acquisition has been traditionally used to address the respiratory/cardiac motion induced artifacts. However, gated PETimages usually have poor image quality due to insufficient photon counts. In this regard, various efforts have been made to estimate the motion information from the gated PETimages first, and then use this information to produce motion‐free, high quality images. The purpose of this study is to combine the two inter‐related tasks—motion estimation and image reconstruction, into one single process. A potential advantage of such an approach is a more accurate estimation of the motion information, as well as a better quality of the reconstructed image.Method and Materials:Motion estimation and image reconstruction are jointly performed by maximizing a modified likelihood function that incorporates motion information. A computer simulation was performed on a 2D digital phantom placed in a simulated PET scanner to test the feasibility of this approach. The digital phantom consisted of a hot sphere moving sinusoidally within a warm background. The motion of the sphere was discretized into 10 steps. This simulated a tumor in the patient body influenced by the respiratory motion during a gated data acquisition using 10 time frames. The simulated PET scanner consisted of two parallel gamma cameras each having 140 detectors, and could rotate in 280 angular steps. Projection data were modeled to follow a Poisson statistics, with 20% of random counts. Results: The jointly estimated image recovered the shape of the moving object while still preserved a good image quality. Quantification error was reduced by 21.2% when compared to the static reconstruction, and SNR was improved by 185% when compared to the gated image.Conclusion: The joint estimation approach we have proposed has a good potential in motion artifacts correction for 4D PETimages.
MO‐E‐330A‐02: Automated Prostate Contour Drawing On Post‐Implant CT Images Based On Ultrasound Volume and Seeds Positions33(2006); http://dx.doi.org/10.1118/1.2241450View Description Hide Description
Purpose: It is difficult to locate precisely the prostate on CTimages. We propose an automated contouring help method based on data acquired during intra‐operative brachytherapy procedure. The algorithm uses ultrasound volume and seeds positions to draw a preliminary contour on CTimage.
Method and Materials: The data acquired during the clinical protocol are the intra‐operative ultrasound volume and the seeds positions based on the detected needle insertion, and the seed positions and the CTimages thirty days post‐implant. In the first step the US volume and seed cloud are matched. For each z position of a clinical CTimage, the US contour and seeds are extracted. The seeds positions are automatically detected on CTimage and are used to calculate a transformation matrix to translate the seeds positions from day 0 to 30. The contours are reshaped using active contour. The reshaping process begins with the center slice and progress on both sides. The active contours are initialize with an expanded US volume. Implementation (results): There are no parameters to adjust in the first part of the algorithm. In the second part, there are six snakes parameters : continuity, curvature, convergence, gradient contraction minimum and contraction maximum. There is also a parameter controlling the resize factor of the US contours. The preliminary tests are conducted on ten clinical cases. Most of the contours were the final contour. The others needed a small physician action to correct the contours. Conclusion: This algorithm will be a useful tool to help physicians in tedious work to draw prostate contours on CTimages. This automated approach presents the physician with intra‐op US volume fused to the 30 days CT exams and proposes a new set of contours based on the morphology of the seed distribution.
33(2006); http://dx.doi.org/10.1118/1.2241451View Description Hide Description
Purpose: To automate segmentation of lesion and organs at risk in lung patients by using Gabor filtering and fuzzy clustering for 4D CTimage data and to check the usefulness of segmentation by using these methods. Method and Materials:Image was pre‐processed with contrast enhancement and median filtering in order to prepare it for feature extraction. By filtering the enhanced image by Gabor filtering, texture features were extracted. Three fuzzy segmentation algorithms were selected including Fuzzy C‐Means (FCM), Gustaffson Kessel and Gath Geva (GG). Optimal number of clusters is determined by using validity criterion including Partition coefficient (F), Classification entropy (H), Xie Beni index (XB) and Partition index (SC). Images were then clustered and clusters visualized by using Principle Component Analysis (PCA). Finally, Performance of the algorithms was calculated by plotting the Receiver Operating Curves (ROC). Results: It was found that different clustering algorithms are best suited for different tissue/Organ segmentations. Furthermore, ROC curves indicated that FCM segmented the left lung with a True Positive (TP) rate of 94.72%, GK segmented Lesion with a TP rate of 88% and GG segmented heart with a TP rate of 74.8%. Fuzzy clustering methods in conjunction with Gabor filters are seen to completely outperform traditional Hard clustering methods like K‐Means and KMediod. Finally, there is a marked difference seen between the fuzzy clustered regions and oncologist marked GTV. Conclusion: Texture based segmentation using Gabor filtering and fuzzy clustering perform quite well for Lesion and organs at risk segmentation, however, different fuzzy clustering methods should be combined for optimal segmentation. As Gabor filters use different frequencies and orientation for filtering, they reveal hidden structure, which may contribute to a more precise delineation of tissue/lesion structure.
33(2006); http://dx.doi.org/10.1118/1.2241452View Description Hide Description
Purpose: Contouring is very labor intensive, especially in image‐guidedradiotherapy involving many image sets. To expedite this task, a semiautomatic image segmentation algorithm was developed to obtain lung and tumor volumes at the end‐of‐exhalation from contours drawn on an image set taken at the end‐of‐inhalation, and vice versa. Method and Materials: The gradient vector flow (GVF) algorithm was first proposed by Xu and Prince. This snake model can dynamically conform to object shapes in response to two kinds of forces: internal force from the contour itself, and external forces from image gradients. The energy cost function composed of these two forces can be minimized based on local GVF information. Different parameters were used for elasticity and rigidity of lung and tumor tissue. A variation of the GVF method is proposed to speed up the convergence. A momentum item ξΔUt EXP(−t/T0) was added to the gradient vector calculation, in which ξ is a momentum constant and T 0 is the initial annealing temperature. Results: Two CTimage sets were used to verify the algorithm. The lung and tumor contours drawn at one phase were used as an initial guess for the automatic drawing in the second phase. The GVF was calculated for each slice and some important parameters, like elasticity, rigidity and mass viscosity etc were assigned specific values for the optimization. The GVF calculation converged quickly to 10−3 in less than 5 seconds for a 512 × 512 image. The ratio of overlapping areas drawn by the algorithm to those of the manual approach exceeded 98%. Conclusion: The close match between computed and manually drawn contours, as well as the short computing time indicates the feasibility of this method for clinical use. This method should also be expandable to other applications like image guided radiotherapy. More work is needed for contouring 4D image sets.
MO‐E‐330A‐05: Evaluation of Similarity Measures for Use in Intensity‐Based Registration in Radiotherapy33(2006); http://dx.doi.org/10.1118/1.2241453View Description Hide Description
Purpose: In radiotherapy, the implementation of automated patient positioning depends on optimal selection of a similarity measure. Visual inspection of a registration is insufficient for evaluating performance of a measure, since high precision registration is usually involved. A general methodology is presented for evaluation of a measure based on convergence, precision and systematic errors for 2D and 3D imaging.Method and Materials: The following common metrics were studied: mean pixelwise product (PROD), normalized correlation coefficient (NCC), partitioned intensity uniformity (PIU), mutual information (MI), normalized mutual information (NMI), entropy of the difference image (ED), gradient correlation (GC), gradient difference (GD), and pattern intensity (PI). Imaging from an anthropomorphic phantom and clinical data were used. Portal images were acquired with a video cameraEPID and a flat‐panel imager. Kilovoltage digitally reconstructedradiographs(DRR) and megavoltage DRRs (MVDRR) were used for reference images. The following shifts were carried out: translations from −10 mm to +10 mm with 0.2 mm increments, and/or rotations from −10° to +10° with 0.2° increments. The effect down‐sampling and image type were also investigated. Results: The selection of a measure depended on image type and down‐sampling schedule. PROD introduced the largest systematic error. NCC, PIU, MI, NMI, and GD typically had systematic errors less than 1 mm and 1°. Down‐sampling has less effect on NCC, PIU, and GC. MI, NMI, GC, GD, and PI have sharp “peaks” around their global optima. Asymmetry in the objective function was observed in NCC, PIU, MI, NMI, and ED. Registrations using GC, GD, and PI were less sensitive to the variation of anatomical structures. Using MVDRRs instead of DRRs increased objective function smoothness. Conclusion: A general methodology was presented for evaluation and selection of a measure. Selection was application dependent, indicating necessity of a validation study, as presented here, prior to clinical use.
MO‐E‐330A‐06: Cone Beam CT (CBCT) Image Fusion for Prostate Localization: Interobserver Variability Study Using Manual and Automatic Registration33(2006); http://dx.doi.org/10.1118/1.2241454View Description Hide Description
Purpose: Applying manual registration of CBCT to a reference image for prostate localization increases treatment time affecting prostate motion and patient setup. A reliable automatic registration can be more time efficient and accurate. This preliminary study demonstrates variability of manual fusion by multiple observers comparing manual to an in‐house automatic fusion. Methods and Materials: 7 patients with 7 CBCT scans each were studied by 4 observers. Off‐line manual fusion between CBCT and Simulation CT was performed using On‐Board Imaging (OBI) software by Varian version 1.2. We pre‐defined our Region of Interest (ROI) and Organ of Interest (OOI) to be the pelvic bones and the prostate respectively. Offsets were recorded in 4 degrees of freedom (3 translations and couch rotation). To increase precision, one outlier observation was excluded. Interobserver variation in manual registration was studied by determining percentage of observations that fall within 1 mm and 1 degree standard deviation (SD). Automatic registration, utilizing the uphill simplex, gradient correlation and mutual information similarity measures, was compared to the manual matched results using frequency histograms of the differences between the two methods. System robustness is checked using prostate calcifications. Results: Comparing different observers, 69% cranial‐caudal (C‐C), 83% anterior‐posterior (A‐P), 100% right‐left (R‐L), and 100% (couch rotation) of the observations fall within 1 mm,1 degree SD for ROI. 69% C‐C, 75% A‐P, 90% R‐L, and 100% couch rotation were observed for OOI. ROI/OOI automatic fusion showed 92%/85% overall agreement with manual fusion within 2 mm respectively. Conclusion: Manual registration is time consuming and demonstrates observers' variations. The automatic method localizes the prostate within 15 seconds. More observers and measurements, methods utilizing drawn contours and implanted fiducial markers, will be required for system accuracy and precision. Our automatic registration with feed back from observers' experiences would enable on‐line correction for prostate motion.
33(2006); http://dx.doi.org/10.1118/1.2241455View Description Hide Description
Purpose: In this work we develop a strategy of automatic contouring to relieve the effort of organ segmentation in 4D radiation therapy. The method adopts a novel technique of control volumes to achieve robust contour mapping among a series of 4D CTimages. Methods and Materials: For a given patient, segmentation of tumor and sensitive structures was manually performed for one of the breathing phases by a physician. Along the segmented contours a number of small control volumes (∼ 1cm) were selected. To obtain contours on another CT phase we mapped the control volumes collectively to this phase using rigid transformation, which served as a good starting contour for further adjustment. The final positions of mapped control volumes were determined by minimizing the energy function consisting of two terms: intensity similarity between the mapped volumes and the original volumes in the selected phase; elastic potential energy preventing control volumes from movement. The approach was tested with the 4D CTimages of 5 lungcancer patients. Results: For the patients the knowledge‐based approach of automatic contouring worked well even for CTimages with significant deformations. In the lung case the contours have the average error of less than 2mm and a maximum error of 5mm for noisy anatomical structures. A significant reduction of time compared with manual contouring was achieved. Conclusions: The auto‐mapping of contours in 4D radiation therapy was implemented with control volumes. The method provides an efficient way for 4D segmentation with high accuracy.
- Small Animal Imaging
33(2006); http://dx.doi.org/10.1118/1.2241739View Description Hide Description
Purpose: The ability to image 3D global structure (e.g. microvasculature) and function (e.g. gene expression) in whole unsectioned tumors, in high‐resolution and with high‐contrast, is of significant present interest in cancer research. We have developed novel optical imaging techniques capable of reconstructing the 3D distribution of absorbance or fluorescence‐emitting sources in tumortissue. Primary advantages include high spatial resolution and contrast, and ability to image a wide range of molecular probes.
Method and Materials: Constitutive RFP expressing HCT116 (human colon carcinoma) xenograft tumors were implanted in hind legs of nude BALBC mice. The tumor microvasculature was dual‐stained by in vivo tail vein injections of light attenuating isotonic ink and FITC lectin. The tumor was excised post‐staining and subjected to a fluorescence‐friendly clearing procedure to render it amenable for optical imaging. Quantitative images of microvascular density and viable‐cell distribution were reconstructed using tomographic algorithms from a set of transmission and fluorescent projection images acquired at multiple angles using appropriate filters, focusing optics and a CCD camera.
Results: Co‐registered, multimodal, high‐resolution and high‐contrast 3D reconstructions were achieved through large, (∼1cm) whole mount, murine tumor samples. Quality assurance tests indicated encouraging accuracy of reconstructedattenuation coefficients and geometry. The 3D reconstructed slices were juxtaposed with histological slices of the same tumor to reveal striking correlation between viable‐cell distribution and well‐perfused regions. A fluorescence friendly clearing procedure was identified.
Conclusion: We present a novel technique for multi‐modality optical imaging of 3D microvasculature and/or viable cell distribution in tumortissue with unprecedented image resolution and contrast. The technique has potential to image a wide variety of tumor and normal tissue structure and function. Of particular interest is the potential to analyze the effect of anti‐angiogenic agents and radiation therapy on the response of the entire 3D tumor microvasculature network.
33(2006); http://dx.doi.org/10.1118/1.2241740View Description Hide Description
Purpose: This work explores the potential of optical‐computed and emission‐tomography (OCT/OET), when coupled with optical clearing techniques, for imaging aspects of biological structure and function for cancer research. OCT/OET are the optical analogues to x‐ray‐CT and SPECT respectively, and can yield high resolution high contrast 3D images of a variety of inherent or applied absorbing and fluorescing stains. Materials and Methods Several methods of staining tissue have been explored and applied to a range of tissue types. Xenograph tumor microvasculature was labeled with both passive absorbing stain and fluorescing active probes (e.g. lectin conjugated with FITC) by tail vein injection. Murine vasculature in major organs(lung,heart, brain) were stained with absorbing dye in a similar manner. OCT/OET imaging was performed using an in‐house custom built scanner. Isotropic 3D transmision and emission data were reconstructed using tomographic algorithms from equiangularly spaced projection images. Results: Isotropic high resolution 3D image data of the xenograph tumor showed extensive peripheral microvasculature with the occasional larger vessels penetrating to the tumor core. High‐quality 3D images of the lungs were achieved showing clear differences in perfusion between irradiated and unirradiated lung regions. Exquisite high contrastimages were acquired of the vasculature and myocardial perfusion of the murine heart.Conclusion Primary advantages of OCT/OET include the preservation of tissue structure in 3D (tissue sectioning is not required), and the ability to acquire co‐registered images of both structure (e.g.micro‐vasculature) and function (e.g. perfusion, gene expression). Higher spatial resolution and higher contrast is achieved when compared with alternative modalities like micro‐CT and micro‐MRI. The techniques are versatile as imaging can be performed on a wide variety of absorbing and fluorescing stains.
33(2006); http://dx.doi.org/10.1118/1.2241741View Description Hide Description
Purpose: To develop and evaluate an image‐guidedrobotic system for measuring pO2 in animal tumors.Method and Materials: The robot consists of an X‐Y horizontal platform holding the rodent bed, a dual‐axis vertical arm for positioning the measurement probe and cannula, and fiducial markers for registering the coordinates of the imaging and robotic systems. Anesthesized and immobilized tumor‐bearing rats are injected with the hypoxia tracer 18F‐FMISO and imaged in the rodent bed using an animal PET scanner. The reconstructedPETimages are then uploaded into the robot computer, the rodent bed affixed to the X‐Y platform, and the coordinates of the robot and PET registered. Based on the tumor hypoxia image displayed on the robot computer, motion of the robotic X‐Y platform and vertical arm are coordinated to guide the positioning and percutaneous insertion of a probe (OxyLite‐Optronix) to measure pO2 at various locations in the tumor. The pO2 readings (cps/voxel) are then compared to the respective image intensity of the measurement points. Results: The registration accuracy between the robot and image coordinate system was better than 0.2mm. Although the 18F‐FMISO in the bladder produced characteristic starburst artifacts in the surrounding, low‐intensity regions, we successfully measured the pO2 for three tracks. The 18F‐FMISO image voxel values were found to be inversely correlated with the intra‐tumoral pO2 of the three tracks with (Pearson product moment) correlation coefficients of −0.899, −0.420 and −0.857. The scatter plot of pO2 vs. image intensity resembled a sigmoidal, rather than a linear relationship. Conclusion: PET‐guided pO2measurement is feasible using this prototype image‐guidedrobot system. To our knowledge, this is the first system of its kind — allowing direct point‐by‐point correlation of physiological probe measurements and image voxel values or, more generally, a physical action at a set of anatomic points identified on a preoperative image.
WE‐D‐330A‐04: Serial Flat Panel Computed Tomography Quantification of Bleomycin‐Induced Murine Lung Damage In Vivo33(2006); http://dx.doi.org/10.1118/1.2241742View Description Hide Description
Purpose: To compare methods of quantifying lung damage based on noninvasive flat‐panel CTimages with the standard method based on histologic section analysis.Method and Materials:Lung damage was induced in mice with bleomycin and damage progression followed with fpCT on days 10, 14, and 21 and with postmortem histological examination on day 21. Ten mice, five control and five treated, were scanned under breath hold with flat‐panel CT (fpCT). Prior to scanning, a tail vein catheter was inserted for administration of IV contrast. The percent lung damage calculated from fpCT (PLDfpCT) imagedata sets was compared to that obtained from sequential histological sections (PLDH). Results: IV contrast was helpful in separating vessels from regions of lung damage. PLDfpCT calculations were on average 2 times greater than PLDH calculations. Linear trendlines were fit between PLDH and PLDfpCT to all data (n=10) and then to only the treated mice (n=5); the resultant R2 values were 0.87 and 0.89, respectively. The fpCT scans allowed for observation of the progression of damage over time versus the single snapshot sampled with histology. Less damage was observed at day 14 than at either day 10 or day 21, which could be due to a transition from an acute inflammatory reaction to inflammation with formation of collagen. Conclusion: The quantification of lung damage based on fpCT images was strongly correlated to the conventional histological section analysis. Although the overall change in damage is consistent between histology and fpCT the magnitude of measured damage differs which may be explained by an inflammatory response throughout the lung which is not detected with histology. A potential strength of these longitudinal studies is the ability to follow individual animals over time to investigate biological affects such as a transition from acute to chronic damage patterns.
33(2006); http://dx.doi.org/10.1118/1.2241743View Description Hide Description
Purpose: Registration of micro‐PET and micro‐CT images is important for the interpretation of multi‐modality images. Here we develop a voxel‐based image registration method to facilitate the co‐registration of multi‐modality imaging of small animals. Method: Rigid image registration (six degrees of freedom: three translational and three rotational parameters) was developed using an open‐source ITK/VTK platform for coregistering micro‐PET and micro‐CT data. Modules importing/exporting DICOM data (binary mask files of contours and image sets) and the scanner specific proprietary microPET/CT/MRI data were also developed. Automated registration procedures based on the normal cross‐correlation (NCC) metric and mutual information were implemented. To better “see” the bony landmarks in microPET, a strategy for enhancing the bone matrix uptake of radiotracer by adding a tracer amount of 18F during the FDG‐microPET imaging was investigated. The convergence behavior of the voxel‐based registration algorithms was analyzed and the global convergence of the calculation was demonstrated. The accuracy of the developed registration algorithm was assessed by measurements using dual‐modality external fiducial line sources incorporated into the mouse cradle. Results: A voxel‐based registration technique has been established for in vivo molecular imaging study. Application of the technique to a number of mice micro‐PET and micro‐CT registrations indicated that an accuracy of better than 0.2 mm is achievable with the help of 18F injection. Without lighting‐up of the bony structures, however, we found that it is difficult to obtain a registration better than 1mm for regular FDG‐microPET. Computationally, the setup based on the NCC metric performs better in comparison with the mutual information approach. Conclusion: A NCC metric coupled with the use of 18F injection allowed us to obtain sub‐millimeter accuracy in microPET‐microCT registration of mice. The robust registration procedure should be valuable for routine molecular imaging application.
33(2006); http://dx.doi.org/10.1118/1.2241744View Description Hide Description
Purpose: The use of small mammals as models of human diseases has increased in the last several years. These models provide a valuable research tool to characterize these diseases and to evaluate new therapies. Dedicated imagingsystems, with the resolution and sensitivity to allow in vivo studies in small animals, reduce the number of animals required for a given study and offer the possibility of serial studies in the same animal. A dual two‐dimensional imagingsystem has been developed to monitor tumor development in mice and other small animals.Method and Materials: This imagingsystem combines functional information gained from projection nuclear emission imaging with the anatomical context provided by an x‐ray transmission image. The system captures a nuclear image in one side of a single computed radiographydetector(CR plate) and then the plate is moved to acquire an x‐ray image in the opposite side. The nuclear image is formed by using a parallel hole collimator. Four fiducial markers, that are visible in both images, are attached to the system. A transformation algorithm uses the information from the fiducial markers to align the images. The nuclear emission image is color coded and overlaid onto the gray‐scale x‐ray image.Results: For the assessment of the imagingsystem, a tissue‐equivalent phantom was manufactured. Initial investigations used I‐125. The spatial resolution, contrast, and detected scattered photon levels have been measured for both x‐ray and nuclear imaging, and the sensitivity of the nuclear imagingsystem was also measured. Conclusion: The lower radiation levels, faster throughput, and lower cost of this imagingsystem would allow the evaluation of functional activity more frequently in serial studies. This imagingsystem would complement tomographic systems by allowing rapid pre‐screening of animals when a complete 3D volume data set is not necessary.
33(2006); http://dx.doi.org/10.1118/1.2241748View Description Hide Description
Purpose: A small animal radiation platform equipped with on‐board conebeam CT and conformal irradiation capabilities is being constructed for translational research. This work reports on the dosimetric characteristics of the x‐ray lens subsystem used for high‐resolution dose delivery. Method and Materials: A constant voltage 225 kVp x‐ray beam from a 0.4 mm focal spot is shaped and directed to enter a 1.5 cm long, 2 cm diameter cylindrical multi‐layer graphite x‐ray lens placed 16 cm downstream from the source. The lens emits a converging circular beam of 40–80 keV x‐rays which forms a narrow cylindrical beam centered at 33.5 cm downstream with a length of approximately 5 cm in air. The pencil beam disperses further downstream. We measured dosimetry of the beam in water equivalent plastic using Gaf‐chromic EBT films (<0.1 mm pixel size) over a SSD range of 30 to 33.5 cm in steps of 5 mm. For each SSD, 36 films were orientated orthogonal to the beam and at every 2 mm between slabs in the depth direction. We established the optical density‐to‐dose calibration with a 6 MV beam. Results: The beam has a highly peaked cross‐beam profile with a full width half maximum of about 1.5 mm in the focal zone. Irregular cylindrical symmetry is present due to imperfect lens construction. Dose deposition was defined as the average value over a circular area with a diameter of 1.5mm from the beam axis. Surface dose rates are 230 – 160 cGy/min over the 3.5 cm range of SSDs. The percent depth dose is approximately 34% at 2 cm depth for a beam at 33.5 cm SSD. Conclusion: Highly localized dose can be delivered with the focused pencil beam. Work is on‐going to improve dose uniformity by lens rotation, and to develop a pencil beam algorithm for planning purposes.
WE‐D‐330A‐08: The Small‐Animal Radiation Research Platform (SARRP): Commissioning a 225 KVp “small‐Field” X‐Ray Source for Monte Carlo‐Based Treatment Planning33(2006); http://dx.doi.org/10.1118/1.2241749View Description Hide Description
Purpose: To characterize the dosimetry of a small‐field 225 kVp x‐ray source and demonstrate the feasibility of the Pinnacle Monte Carlo Dose Computation engine (v7.9t alpha release) for treatment planning in the context of small‐animal radiation therapy.Method and Materials: The 225 kVp therapeutic x‐rays for the Small‐Animal Radiation Research Platform (SARRP) are produced by a GE 225 x‐ray tube (spot size 3 mm; Al filtration 4 mm). A simple small‐field collimation system was constructed; for commissioning purposes we machined a set of brass cutouts which define field sizes of 3×3, 5×5, 10×10, 30×30 and 60×60 mm2 at the nominal SARRP source‐to‐axis distance of 33.5 cm. The x‐ray beams were characterized by irradiating Gafchromic EBT film in water. The film plane was oriented parallel to the central beam axis (z); measurements were made in the x‐z and y‐z planes for all field sizes.Dosimetric data were extracted from the films using a commercial flat‐bed document scanner. Profiles and percent depth‐dose (PDD) curves for depths from 0–9 cm (SSD = 33.5 cm) were imported into our research planning software for comparison with simulated data. Results: Preliminary simulated profiles for the 60×60 mm2 cutout deviate, on average, by 1.7% and 2.8% in the high‐dose region (90% or greater) for depths of 0.5 and 3 cm, respectively. Average PDD discrepancies for depths from 0–3 cm and 3–6 cm are 4.3% and 18.4% for the 60×60 mm2 cutout, respectively, and 6.6% and 29.2% for the 5×5 mm2 cutout. Conclusion: These encouraging results identify the feasibility of using this planning system. Pinnacle supports the finer resolution required for modeling the small fields. Future development will involve incorporation of a measured photon energy spectrum. For purposes of comparison and validation, we will also incorporate an independently‐generated photon phase‐space file for the simulated x‐rays.
33(2006); http://dx.doi.org/10.1118/1.2241750View Description Hide Description
Purpose: The development of IMRT has increased the community's reliance upon dose‐volume constraints for normal tissue avoidance and the development of SBRT has increased the consideration of alternate fraction schedules. The lack of clinical data for supporting these changes in practice places additional pressure on the development of realistic animal models for fractionation and normal tissuedose‐volume studies. In support of these efforts, the construction of an image‐guidedradiation therapy unit for small animals has been initiated. The current status of this development and its design elements are described. Method and Materials: The system is comprised of a decommissioned radiation therapy simulator (Nucletron — Simulix) adapted to support a flat‐panel detector (Perkin Elmer, RID1640) and a 225kVp x‐ray tube (GE Siefert 225, f.s.=0.4–3mm). The pulsed radiographic exposures (200–600) collected over 360 degrees are reconstructed using a filtered back‐projection cone‐beam CT method. Soft‐tissue visibility and geometric targeting of the system components have been assessed. EGSnrc Monte Carlo simulations were performed to assist in the design of the optimal treatment geometry for rodent and rabbit models. The simulations for 225 kVp x‐rays (1 cm diam circular field size) have been considered to achieve high dose rate, small penumbra, and acceptable clearance during arc‐based delivery. Results: The cone‐beam CTimagingsystem generates soft‐tissue images of rodents with sub‐mm resolution. Monte Carlo results demonstrate penumbral performance (d90‐50) of the system with a 3mm focal spot and reduced collimator‐object distance will be under 1mm. A radial dose gradient can be established from 360 degree arc approach (9%/mm from D90‐d50) for small field sizes.Conclusion: The development of a cone‐beam CT guided radiation therapy unit with sub‐mm resolution and high 3D dose gradients is progressing. Initial simulations and imagingsystem development suggests that precisely located spherical dose distributions can be delivered with such a system.