Volume 36, Issue 6, June 2009
Index of content:
- Joint Imaging/Therapy Scientific Session: Room 210A
- Translational Research
TH‐D‐210A‐01: Exploration of Soft‐Tissue Visualization at Low Dose Using Flat‐Panel Imagers Incorporating Thick, Segmented Scintillators for Megavoltage Cone‐Beam CT36(2009); http://dx.doi.org/10.1118/1.3182703View Description Hide Description
Purpose: Megavoltage (MV) active matrix, flat‐panel imagers (AMFPIs) have become the gold standard in radiotherapy portal imaging by virtue of their many advantages. Nevertheless, conventional MV AMFPIs are very inefficient, detecting only ∼2% of the incident radiation at 6 MV. However, recent theoretical and empirical studies have demonstrated that the incorporation of thick, segmented scintillators can significantly improve the DQE of MV AMFPIs, leading to improved image quality at very low dose. This creates the possibility of using megavoltage cone‐beam computed tomography (MV CBCT) to provide soft‐tissue visualization at clinically practical doses.Method and Materials: Prototype AMFPIs incorporating segmented scintillators based on CsI:Tl or BGO crystals with thicknesses ranging from ∼11 to 40 mm have been constructed and are under evaluation for MV CBCT. Each prototype incorporates a detector consisting of a matrix of 120x60 scintillator elements separated by 50 μm‐thick, reflective septal walls, with an element‐to‐element pitch of 1.016 mm. The prototypes are being configured to allow acquisition of tomographic images of low‐contrast, soft‐tissue‐equivalent objects, embedded in a water‐equivalent phantom, at the lowest available dose (one beam pulse per image).Results: For projection imaging, all prototypes offer substantial improvement in contrast sensitivity at low dose compared to conventional MV AMFPIs. In particular, the BGO prototype with a zero‐frequency DQE of ∼20% provides comparable contrast resolution at ∼20 times less dose. In view of such good performance, it is anticipated that reconstructed tomographic images of the contrast phantom, obtained at practical doses, will demonstrate soft‐tissue delineation. Conclusion: Prototype AMFPIs incorporating thick segmented BGO and CsI:Tl scintillators provide significant improvement in image quality at extremely low dose, facilitating the visualization of soft‐tissue at clinically practical doses. It is expected that future, optimized scintillators will lead to highly useful MV AMFPIs for projection and CBCTimaging.
36(2009); http://dx.doi.org/10.1118/1.3182704View Description Hide Description
Purpose: To develop a methodology of therapy portal imaging verification for volumetric modulated arc therapy (VMAT) and RapidArc treatments. Method and Materials: For Varian dynamic IMRT treatments the “DMLC aperture” or CIAO can be imaged by electronic portal imaging device(EPID) or with film. However, during VMAT and RapidArc treatment delivery, MLC leaves constantly move while gantry is rotating continuously. For quality assurance purpose for VMAT and RapidArc treatments we introduce the concept of “Dose Aperture”. In this approach a 3D isodose surface is created with a fixed beam direction, and a portal aperture is drawn around the projection of the isodose surface. For rotational modalities, such as VMAT and RapidArc, a standard orthogonal pair of beam directions can be selected on which the Dose Apertures are projected on the digitally reconstructed radiographs (DRRs). Portal images using these apertures can be obtained to verify the isocenter and location of the treatment with respect to the patient's bony anatomy.Results: For a VMAT plan the prescription isodose surface was generated using well established planning software tools. For the AP and LAT beams, two Dose Apertures were created, which were then projected on the corresponding DRRs for field matching and correlation. The Dose Apertures could be transferred to setup fields which are then readily available for portal imaging application of VMAT and RapidArc treatments. Conclusion: Although setup fields without the treatment apertures can be used to verify the location of the isocenter, the portal image generated with Dose Aperture is patient specific and can be used when weekly portal imaging is required.
36(2009); http://dx.doi.org/10.1118/1.3182705View Description Hide Description
We describe the fabrication and evaluation of a thick pixelated scintillator for megavoltage (MV) imaging composed of a ceramic containing over 99.9% gadolium oxysulfide. This sintered material offers a 59% increase in density over the Lanex Fast B (LFB) phosphor screens most commonly employed in MV imaging. The sintered pixelated array (SPA) is fabricated from a single slab of ceramic. This obviates the need to assemble over a million separate crystals in order to cover a 40.96cm × 40.96cm detector area. As a consequence, the design is amenable to fabrication using methods of mass production. Method and Materials: A 1.8mm‐thick 274 × 250 pixel SPA with 0.4mm pixel pitch is attached to the light‐sensitive surface of an amorphous silicon flat panel detector (Perkin Elmer XRD1640AN). Image quality is characterized using 1MU exposures of the 6MV beam of a Siemens Primus Linac. A QC‐3V phantom is employed to calculate the modulation transfer function(MTF) and contrast‐to‐noise ratio (CNR). The detective quantum efficiency (DQE) is computed. A LFB screen is then evaluated under identical conditions for comparative purposes. Cone beam CT(CBCT)imaging is performed with four arrays tiled side‐by‐side on the detector surface. Results: The half‐maximum value of MTF occurs at 0.32 and 0.34lp/mm for the SPA and LFB, respectively. The DQE(0.1lp/mm) of SPA is 5.8%, versus 1.0% for LFB. The SPA offers a 235% improvement in CNR over LFB. Previously undetectable low‐contrast phantom inserts are clearly visible in SPA MV‐CBCT images.Conclusions: The SPA appears to offer a practical and cost‐effective means of attaining major improvements in MV image quality. The measured MTF and DQE values underestimate the achievable performance, since the SPA and detector photodiode arrays were imperfectly aligned during these evaluations. A DQE(0) value closer to the maximum attainable 7.8% is expected.
Conflict of Interest: Sponsored by Siemens.
TH‐D‐210A‐04: Monte Carlo Calculations of Microscopic Dose Enhancement Factor for Gold Nanoparticle‐Aided Radiation Therapy36(2009); http://dx.doi.org/10.1118/1.3182706View Description Hide Description
Purpose: To quantify the energy deposition due to photoelectrons from goldnanoparticles on a micrometer scale and to calculate the corresponding microscopic dose enhancement factor during GoldNanoparticle‐AidedRadiation Therapy (GNRT). Method and Materials: The Monte Carlo code EGSnrc was modified to obtain the spectra of secondary electrons from atoms of gold and molecules of water under photon irradiation of a tumor infused with 0.7 wt. % gold. Six different photon sources were used: 125I, 103Pd, 169Yb, 192Ir, 50kVp, and 6MV x‐rays. Treating the scored electron spectra as point sources within an infinite medium of water, the event‐by‐event Monte Carlo code NOREC was used to quantify the radial dose distribution, giving rise to gold and water electron dose point kernels. These kernels were applied to a scanning electron microscope(SEM)image of a goldnanoparticle distribution in tissue. Treating each pixel in the image as a point source of gold or water emitting secondary electrons, the dose at each point was calculated, enabling the determination of the microscopic dose enhancement at each point. Results: For the lower energy sources 125I, 103Pd, 169Yb, and 50 kVp, the secondary electron fluence was increased by as much as two orders of magnitude, leading to a one‐to‐two order of magnitude increase in the electron dose point kernel over radial distances up to 50 μm. The dose was enhanced by 100% within 5 μm of the nanoparticles, and by 5% as far away as 30 μm. Conclusion: This study demonstrates a remarkable microscopic dose enhancement due to goldnanoparticles and low energy photon sources. Given that the dose enhancement exceeds 100% within very short (5 μm) distances from the nanoparticles, the maximum radiobiological benefit may be derived from active targeting strategies that concentrate nanoparticles in close proximity to the cancer cell and/or its nucleus.
TH‐D‐210A‐05: Characterization of the Spatial and Energy Distribution of Electrons Emitted From a Gold Nanoparticle Irradiated by X‐Rays Using Monte Carlo Simulations36(2009); http://dx.doi.org/10.1118/1.3182707View Description Hide Description
Purpose: Through Monte Carlo simulations, we investigate the spatial and energy characteristics of electrons formed from interactions of keV photon beams irradiating a goldnanoparticle to understand their role in enhancing cell kill. Methods and Materials: The GEANT4 toolkit was used for Monte Carlo simulation. A goldnanoparticle sphere with a 100 nm diameter was irradiated by x‐rays with energies of 35 to 6000 keV inside a tracking volume of water. For each electron emitted from the irradiatedgoldnanoparticle, the following parameters were tracked: i) the physics process that created the electron, ii) energy distribution of the electron, iii) range, and iv) deflection angle. The same simulations were performed by replacing the goldnanoparticle with water. Results: The energy distribution of the secondary electrons when the goldnanoparticle is irradiated by a monoenergetic photon beam with various energies is calculated. The number of interactions for 35 keV photons is about 157 times more than that for 660 keV, and 683 times more than the 6000 keV beam. When the goldnanoparticle is absent, the probability of creating an electron becomes much lower. Specifically, the ratio of total interactions with and without gold is 812, 137, 10, 7, and 2 for energies 35, 73.3, 660, 1200 and 6000 keV respectively. Furthermore, irradiating the goldnanoparticle at a higher photon energy increases the range over which the election can travel, and decreases the deflection angle. This represents a change in volume over which the electrons can deposit dose. Conclusions: We conclude that for a cell of typical size, a low energy (35 keV) photon beam generates a large number of secondary electrons when a goldnanoparticle is present compared to without, and will have sufficient range to cause damage in the cell in which the nanoparticle is uptaken.
TH‐D‐210A‐06: Drug Eluting Implanted Devices to Increase Biologic Effective Dose in Image Guided Radiation Therapy36(2009); http://dx.doi.org/10.1118/1.3182708View Description Hide Description
Purpose:Image guided radiation therapy(IGRT), such as lungSBRT using fiducials or prostate brachytherapy, implant objects essitial to the radiation treatment, but offering no therapeutic benefit. Such devices offer a vechicle for in‐situ delivery of radiosensitizer which could increase the biologic effective dose of the radiation. This work studies the achievable drug coverage as a function of thechemical and physical properties of the drugs and devices. Method and Materials: Drug concentration around a spherical eluter is parameterized in order to separate the physical properties of the eluter and effective diffusion elimination modulus (φb) of the drug. Computer modeling is used to simulate drug distributions produced by configurations of eluters that could be encountered in either steretotactic body radiation therapy(SBRT) for lung or ultrasound guided prostate implants. Measures of tumor coverage and normal tissue involvement are evaluated for multiple combinations of eluter sizes and φb. Elution kinetics from polymercoating of gold substrate was evaluated. Results: A small number of fiducials, as in lungSBRT, would require either a relatively large fiducial or chemical engineering of the drug to produce a low value of φb. The many spacers seen in prostate brachytherapy provide enough eluters to sensitize a portion of the prostate with values of φb in the vicinity of reported values. The use of drug eluting brachytherapy spacers would enable a more localized enhancement in biologic effective dose than possible by a local brachytherapy boost. Timed release of docirubicin from a polymercoating on a gold substrate is possible. Conclusion:IGRT techniques already implant devices in targets as part of the standard of care. These devices may be enhanced with drug eluting coatings to provide localized increase in biologic effective dose without increasing the physical dose involved in the therapy.
TH‐D‐210A‐07: Multi‐Modal Image Guidance in Neurosurgery: An Approach for Direct Targeting in Deep Brain Stimulation (DBS)36(2009); http://dx.doi.org/10.1118/1.3182709View Description Hide Description
Purpose: To develop and evaluate an integrated multi‐modal image guidance methodology for DBS. Method and Materials: An image guidance methodology was developed to integrate multiple imaging methods as well as atlas and intraoperatively acquired electrophysiology data. Imaging techniques were developed using a 3T MRI to provide unique contrast of specific structures in the brain. A new sequence, the Fast Grey matter Acquisition T1 Inversion Recovery (FGATIR), was developed to produce images that specifically nullified the white matter signal. In addition, diffusiontensorimaging (DTI) tractography was acquired to provide differentiation of target subregions through structural connectivity. These imaging techniques as well as standard (T1, T2 FLAIR, and CT)imaging protocols were integrated on a single platform. This integration was accomplished by the development of a graphical user interface (GUI) to allow for viewing and targeting based off these images along with a deformed atlas and intraoperatively acquired electrophysiology maps. Results: The developed image guidance system allowed for targeting based off all data types mentioned previously. The new imaging techniques developed were compared against intraoperatively acquired microelectrode data. The two novel imaging techniques, FGATIR and DTI tractography, provided elucidation of structure not seen by standard imaging methods and showed good correspondence with microelectrode maps. Contrast measurements for the FGATIR versus T1 and T2 FLAIR showed a higher contrast to noise ratio (CNR) for FGATIR scans. For example, the CNR for ventral lateral thalamus versus the remainder of thalamus was 2.19, 2.81, and 7.27 for T1, T2 FLAIR, and FGATIR, respectively. Conclusion: Our methodology for image guidance in DBS surgery has the potential to allow for direct visualization and targeting of regions that can not be visualized by standard techniques. The integration of all these methods within a single framework demonstrates the potential for multi‐modal image guidance in neurosurgery.
36(2009); http://dx.doi.org/10.1118/1.3182710View Description Hide Description
Purpose: To assess the effects of target, rectum and bladder anatomical deformations on targeting accuracy of prostate and postprostatectomy cancer patients undergoing IG‐IMRT. Methods: Localization with online kV‐CBCT was performed. The target and OARs positional/volumetric changes were evaluated and couch shifts were applied. For patients with large target/OARs volumetric changes compared to planning CT, arising from medications, diet, and/or ongoing RT, repeated localization CB scans were performed following an interventional procedure, shifts were then evaluated, and the IMRTtreatment was subsequently delivered. The interventional procedure involves the insertion of a rectal catheter or rectal deflation, and/or bladder filling. A total of 160 pre‐/post‐intervention shifts from 14 patients in the lateral/LR, vertical/AP, and longitudinal/SI directions were compared. The percentage of shifts larger than 5 mm in all directions was also compared. CTV‐to‐PTV expansion margins were estimated based on the pre‐ and post‐intervention localization data. Results: Systematic/S and random/s shifts from pre‐versus post‐intervention data (in mm) were: LR, 0.2±2.8 vs. 0.4±2.9; AP, −0.7±5.3 vs. −1.1±3.6; SI, 0.6±3.7 vs. −0.5±2.5. The mean 3D shift distance was 6.4±3.1 vs. 4.8±2.4 with a p‐value < 0.05. The percentage of preintervention shifts larger than 5 mm were 7.5%, 31.3%, and 16.3% in the LR, AP, and SI directions, respectively, compared to 8.8%, 15.0%, and 6.3% for post‐intervention. Large anatomical variations were observed for rectum and/or bladder, suggesting that localization without intervention may not be sufficient to ensure accurate targeting and sparing of rectum/bladder. Conclusion: Localization data from pre‐ and post‐intervention procedures show that for treatments that do not include intervention to correct for rectum/bladder anatomical variations, the CTV‐to‐PTV margin required is larger by 5 mm, and more rectum/bladder volumes are potentially at risk of radiation‐induced acute or late toxicity..
TH‐D‐210A‐09: Correlation Between Abdominal Organ Motion and An External Marker Toward Respiratory‐Gated Intensity‐Modulated Radiation Therapy for Pancreatic Carcinoma36(2009); http://dx.doi.org/10.1118/1.3182711View Description Hide Description
Purpose: To attain the respiratory‐gated intensity‐modulated radiotherapy(IMRT) for pancreatic carcinoma by using an external marker as a surrogate for tumor motion, the correlation between a motion of an external marker and displacement of a stent around pancreas was investigated. Methods and Materials: Three patients who had undergone the ERCP for biliary decompression with placement of a stent were applied in this study. An anterior‐posterior motion of an external marker placed on the patient's abdomen was monitored by the Varian® RPM system for two minutes, this was synchronously displayed on the X‐ray fluoroscopic screen which acquired the stent's superior‐inferior displacement. For each patient, the above measurement was performed at four sessions during the treatment course. Both the displacements of an abdominal marker and a stent were automatically extracted by in‐house software based on the template‐matching algorithm. The cross‐correlation coefficients between displacement of an abdominal marker and a stent motion and their intra‐fractional variations were evaluated. The predicted stent positions at 30%, 40% and 50% of respiratory phase, which were traced the regression line from 0%‐phase (Exhalation), were compared with actual positions. Results: The correlation coefficients () ranged from 0.77 to 0.97 and were inter‐fractionally reproducible. The predicted stent's positions at 30 %, 40 % and 50 % of respiratory phase were discrepant with actual positions with errors of 1.6 mm, 2.0 mm and 2.1 mm on average, respectively. Conclusions: The strong correlation between a motion of an external marker and a displacement of a stent around the pancreas was observed with a high reproducibility. An error of the predicted target position by an external marker was up to 3.1 mm for the expected‐gating duration (∼50 % respiratory phase) Our study demonstrated that the respiratory‐gated IMRT for pancreatic carcinoma may be clinically acceptable.