Volume 36, Issue 6, June 2009
Index of content:
- Young Investigators Symposium: Ballroom C
- John R. Cameron: Young Investigators Symposium
SU‐GG‐BRC‐01: Modeling Myocardial Mn2+ Efflux Rates Using Manganese‐Enhanced MRI T1 Mapping in a Murine Myocardial Infarction Model36(2009); http://dx.doi.org/10.1118/1.3182177View Description Hide Description
Purpose: Alterations in myocyte handling appear to be centrally involved in the dysfunctional characteristics of the failing heart. This study uses quantitative manganese‐enhanced MRI techniques to detect changes in the efflux relative to fluctuations in mice following inhibition of the sodiumcalcium exchanger (NCX) with SEA0400. Furthermore, the technique was applied to a mouse myocardial infarction (MI) model. Segmentation and modeling analyses were used to examine regional changes in efflux rates, allowing for a study of the dynamics in the peri‐infarction zone. Methods and Materials: was infused via the tail vein into C57Bl/6 mice (n=88). T1‐maps were obtained both pre‐ and post‐infusion at multiple time points. Time dependent changes in the relaxation rate (ΔR1) were calculated in the myocardium. For the MI mice various affected zones within the myocardium were identified and analyzed using segmentation software. The results from control and SEA0400 treated mice were applied to a pharmacokinetic model in order to estimate the transfer rates. Results: The ΔR1 efflux half‐life was doubled following treatment with 50 mg/kg SEA0400, with the two compartment model predicting a reduction in the myocardial efflux rate by more than a factor of two. In the MI group constant ΔR1 values for viable and infarcted tissue were fit with radial analysis. A significant difference was observed between the efflux rates of the infarcted region to that of the viable region, with a continuous range of efflux rates in the peri‐infarcted region. Conclusions: Quantitative MEMRI with T1‐mapping has demonstrated the sensitivity to observe changes in efflux with the ability to model the relative efflux rates. The technique also provides enough sensitivity for identifying the potentially salvageable adjacent zone as well as examining regional alterations in fluxes leading to relative information, potentially applicable to monitoring disease progression.
SU‐GG‐BRC‐02: Theoretical Analysis of Fiber‐Optic Scintillation Glass Arrays for High Quantum Efficiency Megavoltage Imaging and Image Guided Radiotherapy36(2009); http://dx.doi.org/10.1118/1.3182178View Description Hide Description
Purpose: In this work, we present a comprehensive theoretical analysis of a high performance megavoltage imager based on a fiber‐optic scintillation glass array (FOSGA), a potential candidate for megavoltage computed tomography (MVCT) as part of image guided radiation therapy(IGRT). The FOSGA imager was studied in several prototype configurations using Monte Carlo simulations and linear cascaded systems analyses, and imaging performance was characterized by its modulation transfer function(MTF) and detective quantum efficiency (DQE). Method and Materials: Our imager consists of Tb‐doped scintillation glass fibers inserted into a polymer‐tungsten grid‐housing. Material and structural specifications of the array are specially geared towards patented fabrication technologies for automated mass production and cost‐benefit. Radiation detection characteristics were modeled using radiation transport simulations and coupled to a custom‐written optical transport simulation program developed to specifically model the fiber‐optic coupling mechanism in our detector. A linear cascaded systems analysis was conducted to evaluate the effects of all proposed system components, design parameters, and imaging performance metrics (MTF & DQE) were determined for these configurations. Results: Radiation transport calculations indicated high intrinsic imaging performance typical of thick high‐Z scintillators with > 40% quantum efficiency for 5 cm thick arrays. Light transport simulations indicated limited coupling loss and optical cross‐talk from scatter as optical conduction was dominated by Fresnel and total‐internal‐reflection events. Cascaded systems analyses revealed no significant quantum sinks while indicating that detector thickness was the limiting parameter affecting overall performance [DQE(0) > 0.25 and MTF > 0.5 at 1 cy/mm, significantly improved over current megavoltage detectors with DQE(0) ∼ 0.01]. Conclusions: With high DQE and significant cost‐benefit compared to current megavoltage detectors and contemporary prototypes, a FOSGA imager can provide high portal image quality at low dose, which facilitates its extension to 3D verification via MVCT as a cost‐effective IGRT solution.
36(2009); http://dx.doi.org/10.1118/1.3182179View Description Hide Description
Purpose: This is a proof of concept study with the objective of reconstructing the position of an HDR source in 3D in real‐time using a flat panel detector(FPD). It can potentially become the new standard in Quality Assurance (QA) for treatment delivery. Method and Materials: A matrix of markers (Ball Bearings 4mm in diameter) with precisely known locations was mounted on the cover of a flat panel detector(Acuity, Varian Inc) at variable height. Images acquired with the x‐ray source were used to calibrate the system. A plan with three dwell positions, well defined in 3D was created and delivered. Images were acquired with the FPD during the delivery of ‘treatment’. In house software was created to automatically segment and label the markers' images. A mathematical solution for the ‘near‐intersection’ of two 3D lines was implemented and used to determine the ‘true’ 3D source position. Each line was defined by the 3D positions of each marker and its projection on the FPD. A matrix with N markers will produce N*(N‐1)/2 points of intersection and their mean will result in a more accurate source position. The HDR source was placed on a 5cm solid water to mimic the patient and the FPD was placed at distances varying from 50 to 70cm. Results: The best imaging geometry was determined and images of markers obtained with the HDR source (strength of 6.2Ci) were properly segmented at all distances. During delivery, the source was located at [0,0,50], [0.5,0,50] and [2.0, 0, 50]. The reconstructed positions were [0,0,50.130], [0.497,−0.008,50.106] and [1.984, −0.005,50.053] with a standard deviation of [0.027,0.019,0.115]cm. When intersecting lines in 3D, the mean shortest distance between any two lines was 0.025cm with standard deviation 0.016cm. Conclusion: We proved that the accuracy of source position detection in 3D using a FPD is sub‐millimeter.
SU‐GG‐BRC‐04: Electronic Versus HDR Ir‐192 Brachytherapy: Organ Dose Comparisons for Breast Cancer Using a Monte Carlo Patient Phantom36(2009); http://dx.doi.org/10.1118/1.3182180View Description Hide Description
Purpose: To quantify and compare the dose delivered to multiple organs‐at‐risk (OARs) in a female patient undergoing Xoft Axxent electronic (KVB) and high‐dose rate Ir‐192 (IBB) intracavitary balloon brachytherapy for breast cancer.Materials and Methods: A previous study has indicated that the dose to OARs such as the lungs and heart play a critical role in treatment planning. The anatomy of a female patient was represented by an adult female computational phantom which consists of over 140 organs. A balloon was inserted into a lumpectomy cavity in the left breast of the virtual patient. The Monte Carlo N‐Particle eXtended (MCNPX) code was used to simulate photon transport through the patient for hypothetical KVB and IBB scenarios. MCNPX's F6 tally was used to calculate the absorbed dose in organs distant from the treatment site. Results: In general, the KVB organdoses were more than a factor of 2 smaller than those of IBB because the low‐energy x‐rays are less penetrating. The distribution of organdoses shows a profound pattern depending on the distance, location, and organ shape. The largest doses were observed for organs such as the left lung and heart which are closest to the radiation source. For KVB, the doses received by the left lung and heart wall were 9.0% and 5.5% of that received by the planning target volume. These values were 11.0% and 11.3% for the IBB scenario. Conclusions: This paper reports, for the first time, a systematic comparison of multiple organdoses received from KVB and IBB. KVB may have safety advantages because its dose rate falls off faster than for IBB. As a previous clinical study found the target dose to be similar for these two methods, information on how healthy organs are irradiated will help decide when each modality is appropriate.
36(2009); http://dx.doi.org/10.1118/1.3182181View Description Hide Description
Purpose: To assess the importance of hypoxia imaging for radiation treatment planning by evaluation of the effectiveness of modulated radiation treatments targeting hypoxia with increased doses.Method and Materials: Starting with a two‐dimensional diffusion model of tissue oxygenation, a planar tumor model was developed to simulate chronic and acute hypoxia as well as oxygenation/reoxygenation effects on the microscopic scale. Tumor cell radiation response was evaluated using linear quadratic model taking into account individual cell oxygenation levels. Cell repopulation/tumor shrinkage was not considered, as to represent a slow‐growing tumor. Using this model, we evaluated the effectiveness of intensity modulated radiation therapytreatments targeting hypoxia with increased radiationdoses. We compared different approaches to targeting hypoxia, including fixed dose modulation based on the pre‐treatment imaging of hypoxia and adaptive dose modulation based on interfraction imaging of hypoxia. Furthermore, to study the importance of hypoxia imaging resolution we compared two scenarios: 1. microscopic, where cell‐by‐cell imaging of hypoxia and dosedelivery is possible, 2. macroscopic, where both imaging and irradiation can be conducted with 1mm resolution. Results: For a total tumordose of 30Gy, the adaptive hypoxia‐targeting therapy provided the best tumor control for the microscopic scenario with compared to for the treatment with uniform dose of 2Gy per fraction. However, for a more realistic macroscopic scenario, the finite resolution of hypoxia imaging and dosedelivery significantly reduced the therapeutic advantage of adaptive hypoxia targeted therapy resulting in compared to 7.09⋅10−5 for the uniform dosetreatment.Conclusion: We have demonstrated that the adaptive dose modulation based on oxygenation map acquired before each fraction does results in slightly better tumor control compared to uniform dosedelivery. However, the tumor control improvement is yet to be proven significant enough to justify the clinical implementation of hypoxia‐targeting dose modulation.
36(2009); http://dx.doi.org/10.1118/1.3182182View Description Hide Description
Purpose:Medical procedures such as cardiac catheterization, angiography and the deployment of endovascular devices are routinely performed using x‐ray fluoroscopy, in which each image is obtained at very low x‐ray exposures. The imaging performance of current solid‐state flat panel detectors(FPD) is compromised by electronic noise at these low detector exposures (0.1–10 μR/frame). There is thus a clear need to develop an imaging detector with the quantum noise limited operation of an x‐ray image intensifier and the inherent advantages of a compact solid‐state device. Here we propose a technology that takes advantage of avalanche multiplication of charge in an amorphous selenium (a‐Se) photoconductor.Method and Materials: To determine whether this technology holds promise for next‐generation FPDs, we investigate the following: (1) device and material requirements for prevention of electrical breakdown, (2) leakage currents at high electric fields, (3) real‐time imaging capability and linearity, and (4) the compliance of an avalanche a‐Se photoconductor with low‐voltage image readout electronics. Results: Our results show that a distributed resistive layer coupled to the avalanchephotoconductor enables breakdown‐free operation. We report, for the first time, avalanche gains exceeding 104 in a solid‐state x‐ray detector, and leakage currents of only ∼10 pA/mm2. The detector has a voltage‐programmable avalanche gain and can be operated in a linear regime at 30 frames per second over a five order of magnitude x‐ray exposure range, including the lowest clinical exposures encountered in fluoroscopy. Furthermore it is compatible with existing thin film transistor technology on which current FPDs are based. Conclusion: This detector technology should enable the development of next‐generation dose‐efficientFPDs for interventional radiology as well as advanced applications such as cone‐beam computed tomography or tomosynthesis. Combined with techniques such as region‐of‐interest fluoroscopy, our detector technology could significantly reduce radiation dose to the patient and physician.
36(2009); http://dx.doi.org/10.1118/1.3182183View Description Hide Description
Purpose The present treatment‐planningdosimetry constraints for the lung are based on studies that assume lungtissue is homogeneous in its response to toxicity, irrespective of tissue location or underlying function. No human studies have investigated the relationships between local lung function, spatial radiation dose distribution, and radiation induced lung function changes. We are able to observe radiation‐induced changes in lungtissue function using 4DCT and image registration.Method and Materials Two 4DCT data sets before and after RT from one patient are used in this study. Nonlinear, 3D image registration was applied to register the maximum exhalation image to the maximum inhalation image for the calculation of local lung expansion as a measurement of regional pulmonary function. We compared the changes of pulmonary function before and after RT with planned radiation dose at different locations of the lung.Results The registration accuracy analysis indicated our registration error is on the order of 1 mm. The pulmonary function change in left lung where the tumor is located is larger than it is in right lung (maximum Jacobian change is respectively 0.23 and 0.15). The highest correlation of the pulmonary function change to the delivered radiation dose is in regions which are at the distance of 20 to 25 mm to the tumor region (linear regression, r = −0.73). Further plots between the radiation dose and the pulmonary function change suggest the lungtissue function change is not sole based on radiation dose.Conclusion We described a technique that uses 4DCT, image registration and biomechanical analysis to measure regional lung function. We compared the regional lung function before and after RT to the planned radiation dose distribution and examined the differences at the treatment location and in non‐treatment regions.
Conflict of Interest VIDA Diagnostics, Inc (shareholder).
36(2009); http://dx.doi.org/10.1118/1.3182184View Description Hide Description
Purpose: To determine the depth of the inflection point in the buildup region of photonbeam depth‐ionization curves for 11 cylindrical ionization chambers of differing designs and investigate dependencies of the inflection point on beam energy, electron contamination, and chamber design. Method and Materials: Each ionization chamber is carefully aligned to the water surface using a precision alignment telescope with a custom, high‐precision beam scanning system. Following alignment, depth‐ionization curves are obtained including scanning the chamber beyond the air‐water interface for 6, 10, and 25 MV photonbeams. Measurements are made with and without a 1 mm lead foil in the beam for a 10×10 cm2 field. The inflection point in the measured depth‐ionization data is determined by finding the maximum in its second derivative. Results: The location of each chamber‐specific inflection point is invariant to changes in beam energy, electron contamination, and chamber orientation within the 0.5 mm measurement resolution. Eight chambers, whose outer diameters are within 0.7 mm of each other, exhibit inflection points that are within 0.5 mm of one another. The location of each chamber's inflection point is most strongly impacted by its outer radius and corresponds with the shallowest depth at which the chamber is fully submersed in the water, within 0.5 mm. Conclusion: The buildup region inflection point for a given cylindrical chamber is invariant to the beam variations studied and occurs at a depth equal to the chamber outer radius, within measurement resolution, for the 11 ionization chambers studied. Using this dependence, the inflection point can be used as a robust offline check to identify the water surface location and to identify inter‐setup/inter‐user setup variations. Higher quality surface localization can ease beam modeling and improve the quality of treatment planning data.
SU‐GG‐BRC‐09: CT Detection of Primary and Metastatic Lesions with a Nano‐Agent in Rabbits: Validation with FDG‐PET36(2009); http://dx.doi.org/10.1118/1.3182185View Description Hide Description
Purpose: To evaluate the performance of a recently engineered nano‐sized liposomeCT agent to detect primary and metastatic tumor lesions in a VX2‐sarcoma rabbit model as compared to FDG‐PET. Method and Materials: Nine New Zealand White rabbits bearing VX2‐sarcoma in their left lateral quadriceps received a single intravenous injection of 80 nm liposomes co‐encapsulating 185 ± 37 mg/kg iodine in the form of iohexol (Omnipaque®) and 7 ± 1 mg/kg gadolinium in the form of gadoteridol (Prohance®). The CT/PET (GE Discovery ST) imaging session took place twelve days after the tumor inoculation procedure, five days post liposomecontrast administration and one hour post‐FDG injection (30.3 ± 5.1 MBq/kg). Following CT/PET imaging, the rabbits were sacrificed and the primary and metastatic lesions were examined by a pathologist. The measurement of tumor size was performed on the CT data set. The registration of the CT and PET images was performed using MIPAV. Results: Liposome‐CT demonstrated the same sensitivity and specificity as FDG‐PET for the detection of the 9 primary tumors (volumes = 25 – 7280 mm3, SUVmax = 1.5 – 10.9, HUmax = 173 – 596). In addition, liposome‐CT detected 13 metastatic muscle lesions (volumes = 14 – 2732 mm3, HUmax = 254 – 493) that were histologically malignant, while FDG‐PET identified 7 (volumes = 54 – 1044 mm3, SUVmax = 2.7 – 7.1). For the 16 lesions detected by both imaging modalities, there was a positive correlation between the PET SUVmax and the CT HUmax. Conclusion: In this investigation, increased contrast of primary and metastatic lesions in CT was achieved with the administration of the liposome nano‐agent. This demonstrates the feasibility of employing the liposome‐CT method for effective tumor detection and localization.
SU‐GG‐BRC‐10: Shape Matters: Utilization of a Conformal Voxel Technique to Acquire Robust in Vivo Prostate MRSI at Short Echo Times36(2009); http://dx.doi.org/10.1118/1.3182186View Description Hide Description
Purpose: We seek to improve the quality of in vivo prostate MRSIdata acquisition by utilizing an optimized conformal voxel technique coupled with a spatial‐spectral excitation PRESS pulse sequence for short echo time acquisitions. Method and Materials: All subjects were scanned on a GE 1.5T Signa MR scanner equipped with Echospeed gradients. A standard endorectal coil in combination with a torso phased‐array coil was used. The PRESS pulse sequence was modified to include the optimized conformal voxel MR spectroscopic imaging technique (CV‐MRS). This method uses up to twenty Very Selective Saturation (VSS) pulses, automatically positioned in three dimensions, to “conform” the excitation volume to the shape of the prostate, effectively nulling signal from periprostatic lipids. Subjects were scanned using both the standard PRESS and the optimized CV‐MRS techniques at long and short echo times (TE). In vivo prostate spectra were collected and processed using a modified version of LCModel. Results: We observed an average lipid reduction of 60±18% for 17 subjects over the entire prostate when using the optimized CV‐MRS technique as compared to standard MRSI techniques. In specific regions along the peripheral zone, we observed lipid reduction greater than 95%. The effect of reducing the lipid contamination has resulted in a ∼70% improvement in peak identification of key prostate metabolites, based on goodness‐of‐fit parameters. Furthermore, short TE acquisitions have resulted in a substantial increase in the citrate signal, full visualization of the citrate multiplet and other metabolites not seen at long echo times. Conclusion: In vivo implementation of this optimized MRSI technique has confirmed the reduction in peripheral lipid contamination, and improved the quality of spectra throughout the prostate. Furthermore, this is the first demonstration of short TE in vivo prostate MRSI acquisitions, which provides significant signal increase and reveal short TE metabolites to potentially improve prostate cancer detection.