Funding Agencies

Projects

Click Here to view a PowerPoint presentation about phantom research.

1. The VIP-Man Model
2. Pregnant Female Models at 3-, 6- and 9-month Gestational Periods
3. Radiation Treatment Planning Using Monte Carlo Adjoint Method
4. Use of Rando Physical/CT Phantoms and MOSFET Dosimeters
5. 3D Ultrasound Imaging Software Ultra3D
6. Radiographic Imaging Simulator (ViPRIS)
7. CAD, MCNP and Phantom Fabrication
8. Virtual Reality Based Radiation Safety Simulator
9. Motion-Simulating 4-D Patient Model
10. Dosimetry for Secondary Cancers from Radiation Treatment
11. Nano-Structures Based X-ray Sources for Imaging and Radiotherapy
12. In-Situ Gamma Contamination Depth Measurement
13. Modeling of Medical Accelerators for Simulations of Secondary Doses and Activations
14. Deformable and Posture Changeable Adult Male and Female Phantoms
15. Modeling of CT Scanners and Development of CT Dose Reporting Software
16. GPU-based Monte Carlo Acceleration for Nuclear Reactor Analysis
17. Proton Telescope for Image-Guided Proton Radiotherapy
18. Applications of Compressive Sensing to Nuclear Science and Engineering
19. Animating Computational Human Phantoms Using 3-D Motion-Capture Data

The VIP-Man Model

MIRD-stylized mathematical models are easy to compute and standardize, but are simplified and crude compared to human anatomy. In this study, we have developed a whole-body tomographic model called VIP-Man using color photographs from the Visible Human Project. This VIP-Man model has been used for Monte Carlo organ dose calculations. The Monte Carlo codes that we use include: MCNP, EGS4, and MCNPX. At a voxel size of 0.33mm x 0.33mm x 1.0mm, the VIP-Man Model has the best resolution among all voxel models that are available today. VIP-Man also has many segmented and classified organs/tissues (such as skin, lens of the eye, red bone marrow) that were not available in other models. VIP-Man is now used to help improve the "ICRP Reference Man" and to re-evaluate a large amount of health physics dose quantities (both for internal and external dosimetry). The VIP-Man and the associated Monte Carlo algorithms are also used in patient treatment planning and radiographic image optimization. The organs of the VIP-Man can be adjusted and assembled to create new models as illustrated in this movie. VIP-Man model is part of the Consortium of Computation Human Phantoms. Click here for more information. Funding for this project includes National Science Fundation/CAREER and National Institute of Health (National Library of Medicine and National Cancer Institute). <top>

An adult male whole-body voxel model, called Visible Photographic Man (VIP-Man), has been constructed and implemented in several Monte Carlo codes.

Pregnant Female Models at 3-, 6- and 9-month Gestational Periods

Patient modeling for radiation dosimetry is undergoing a paradigm shift from simplified anthropomorphic models to more realistic, image-based models. We recently developed a new set of models of pregnant mother and fetus representing 3-month, 6-month and 9-month gestational periods for inclusion in the next release of OLINDA/EXM. These new models are made using advanced surface geometries such as NURBS and polygons. The 3D surface definitions allow the organ masses of both the mother and fetus to be adjusted to match those proposed in ICRP Publication 89 for reference individuals. Using medical images such as those from the Visible Human Project and a set of CT pregnant woman as anatomical reference, we defined surface geometries for each organ and the whole body for both the mother and fetus. Software was developed to convert the finished surface models into voxels at a desired voxel size. The voxelized models, whose masses were within 1% of the ICRP89 values, were first implemented into the EGS4 and MCNPX Monte Carlo code systems for intercomparison. The EGS4 code was used to calculate the final specific absorbed fractions (SAFs). Click here for more information. View a movie of the EON based RPI-P9 interactive program. This work is supported in part by grants 1R42CA115122-01 and 5R01CA116743-03 from the National Institutes of Health via subcontract from RADAR. <top>

Left: Whole body of the 9-month pregnant model.
Right: Close-up rendering showing the inner organs and fetus.
Bottom: Close-up of the fetus.

Radiation Treatment Planning Using Monte Carlo Adjoint Method

The objective of radiation therapy treatment planning is to find the optimal beams that maximize the dose to the tumor while minimizing doses to the adjacent healthy tissues. A method called Adjoint Monte Carlo (AMC) has been developed and used by the reactor physics community for years to derive importance functions much more efficiently than the forward Monte Carlo calculations. In this study, we apply the Adjoint Monte Carlo framework to a 3D anatomical model called the VIP-Man which was constructed from the Visible Human images. So far, this study clearly demonstrates the feasibility of the AMC in optimizing the beam directions during treatment planning based. Realistic three-dimensional anatomical patient data were used to perform these calculations. Click here for more information. <top>

Left: Adjoint MC simulates radiation transport in “backward-direction with the particles gaining energies.
Right: Optimal sources are picked and “forward-direction MC calculations are run to determine dose distribution.

Use of Rando Physical/CT Phantoms and MOSFET Dosimeters

The RANDO phantom has been widely used to measure doses in radiation protection and for the evaluation of radiotherapy treatment plans. While physical phantoms are indispensable in real dose measurements, computational phantoms are more flexible. To take advantages of both phantoms, a tomographic model has been developed from a series of CT scans of the RANDO phantom at Rensselaer. This model has been implemented into MCNP and benchmarked by comparing with the measurement results. At the same time, MOSFET dosimeters are increasingly utilized in radiation therapy and diagnostic radiology. However, it is difficult to characterize the dosimeter responses for monoenergetic sources. In this study, we have developed a detailed Monte Carlo simulation model of the MOSFET dosimeter using MCNP to characterize the energy and angular dependence. To benchmark the Monte Carlo dosimeter model, we have devised several experimental tools including phantoms and rotating devices for angular dependence setups. A Cs-137 photon source was used to determine the radiological characteristics such as linearity, angular dependence, and depth dependence. Click here for more information about RANDO or click here for more information about MOSFET. <top>

Left: Re-constructed 3D CT images of the RANDO phantom.
Right: 3D modeling of the MOSFET dosimeter in MCNP code.

Free-hand 3D Ultrasound Imaging Software Ultra3D

Ultrasound imaging has become one the most important modalities for diagnosis and staging of a number of diseases. Compared to other imaging modalities, ultrasound is safe, affordable, and mobile. Ultrasound is also one of the few devices that can provide patient images in real-time, thus allowing for cardiac and intra-operative guidance during surgery and radiation treatment. A free-hand 3D image processing and visualization software package, Ultra3D, has been developed and tested, especially to work with Terason’s miniaturized SmartProbe^TM for freehand 3D ultrasound imaging. Click here for more information.

In a video produced in 2002 by the Chinese TV Program called The Light of Science and Technology, Prof. Xu talks about 3d ultrasound imaging and digital human technologies in China. <top>

Left: 3D image in Ultra3D GUI showing the full-body of a fetus phantom.
Right: Ultra 3D user interface.

Radiographic Imaging Simulator (ViPRIS)

Optimization of the radiographic imaging process involves the maximization of image quality and minimization of patient dose. This research developed a computer simulator that models the imaging process. The simulator utilizes an image projection phantom and a dosimetry phantom involving the original CT data set of the Visible Human Project and the segmented VIP-Man/EGS model respectively. Monte Carlo modeling was used to simulate and quantify the scattered x-rays though the VIP-Man phantom. Effective doses were simultaneously calculated. Quantum noise, grid characteristics, detector response, implantation of a simulated lesion, and formation of the final x-ray image are simulated within ViPRIS. The simulated images allows for observer studies to evaluate the image quality and determine the minimum dose at which simulated lesions may be detected. Ratios of effective dose per unit energy imparted (E/e) for several common radiographic examinations and CT machines are calculated using the VIP-Man tomographic model and the EGS Monte Carlo code. Conversion coefficients and charts were published and made available for use in patient risk calculations. Click here to learn more about ViPRIS or click here to learn more about dose calculations. <top>

Left: Radiographic imaging chain containing the ViPRIS computational phantom.
Right: Resulting image generated from 50 keV photon x-ray and CT data set.
Bottom: X-ray spectrum data generated by XCOMP5R code.

CAD, MCNP and Phantom Fabrication

We are developing methods to use Computer Aided Design (CAD) geometries in: a) MCNP code and b) physical phantom fabrication. This project is funded by the National Institute of Standard and Technology and the Department of Energy/NEER. Click here for more information. <top>

Left: MCNP geometry input from a CAD drawing of a room where Cs-137 source is stored.
Right: The left lung of the VIP-Man model in voxel is converted to surface meshes.
Bottom: The STL file of the lung is used to fabricate a 3D mold first using 3D prototyping and then a casting device.

Virtual Reality Based Radiation Safety Simulator

We are developing a simulator of a nuclear power plant environment using virtual reality authoring tool, EON, to define the 3D layout in CAD for the calculation of effective dose equivalent (EDE) data. This software is used to train workers on how to improve radiation safety and work efficiency. This project is funded by the Electric Power Research Institute. Click here for more information. <top>

Left: A nuclear power plant reactor loading area is first modeled in a CAD software SolidWorks and then implemented in the EON virtual reality authoring environment.
Right: Various avatars are included in the augmented VR environment to simulate movement of workers and their radiation exposure using pre-calculated effective dose database.

Motion-Simulating 4-D Patient Model

Respiratory motions have a profound impact on the quality of radiation treatment of cancer. The current paradigm in radiation treatment involving external beams is still based on an assumption that both the tumor location and shape are known and remain unchanged during the course of radiation delivery. Such a favorable rigid-body relationship does not exist in anatomical sites such as the thoracic cavity and the abdomen, owing predominantly to respiratory motions. Computational models with 4D motion-simulating capabilities have shown a considerable advantage in understanding the effects of the respiratory motions and in developing new management strategies. The goal of this project is to develop physics-based and anatomically detailed motion-simulating human models for 4-dimensional Monte Carlo dosimetry simulations. This project is funded by the National Institute of Health/National Library of Medicine. Click here for more information. <top>

Motion-Simulating 4- Dimensional Human Phantom.

Dosimetry for Secondary Cancers from Radiation Treatment

Prof. Xu organized a symposium at the 2006 APPM Summer Meeting about non-target exposures from emerging modalities. Click here for more information. The change from 3D-CRT to IMRT has been accompanied by many new challenges including a potential for an increase in second malignancies due to more fields and longer exposure times. The goal of this project is to establish robust methods involving in-phantom measurements and Monte Carlo simulations of whole body doses from 3D-CRT and IMRT treatment plans. Measurements using a RANDO phantom and MOSFET dosimeters are re-constructed to determine organ doses from typical treatment plans, 4-field 3D-CRT, 6-Field 3D-CRT, and 7-Field IMRT for the prostate. Various accelerators are modeled in the MCNP code to estimate the scattered, secondary photons and neutrons. We use various computational phantoms including RANDO, VIP-Man, and pediatric phantoms through collaboration with the CCHP. This project is funded by the National Institue of Health/National Cancer Institute. <top>

Left: The RANDO physical phantom and MOSFET dosimeters are used to measure organ doses from various treatment plans.
Right: Organ doses for various treatment plan.
Bottom: Virtual models of the accelerator and a patient for Monte Carlo simulations.

Nano-Structures Based X-ray Sources for Imaging and Radiotherapy

The objective of this project is to discover nanostructures that possess optimized properties for a miniaturized X-ray cancer-killing device that is based on well established High Dose Rate Brachytherapy. The research plan has four major thrusts: (1) Fabrication and characterization of a suite of nanostructures including tubes using focused ion beam deposition and rods using oblique angle deposition. 2) Critical evaluation and comparison of nanostructures in terms of architecture, cost, threshold level, peak emission density, stability and lifetime. 3) Innovative design of a miniaturized X-ray source that is suitable for insertion into the prostate for HDR brachytherapy. 4) Demonstration of this cold, on/off controllable source for animal tests. These tasks combine the strengths of Rensselaer by involving two highly visible groups in nanotubes (Prof. Ajayan) and nanorods (Prof. Lu), with the objective being driven by the design of a functional brachytherapy system. <top>

Left: Arrays of multi-walled nanotubes by Prof Ajayan.
Right: Nano-rods, beams, and springs on metallic substrates by Prof. Lu.
Bottom: Electron field emission characteristics by Prof. Lu.

In-Situ Gamma Contamination Depth Measurement

In this project, we have developed novel detector systems and algorithms to characterize the depth of radiological contamination in buildings. This project involves in-situ gamma-ray spectroscopy. Our work in this area has resulted in a patent (U.S. Patent Serial No. 6,518,579, February 11, 2003). This project was funded by the Department of Energy/Environmental Management Program. <top>

For more information about this project, please refer to:

• Naessens EP, Xu XG. A Non-Destructive Method to Determine the Depth of Radionuclides in Materials In-Situ. Health Physics, 77(1):76-88, 1999. view
• Al-Ghamdi A and Xu XG. Estimating the Depth of Embedded Contaminants from In-Situ Gamma Spectroscopic Measurements. Health Physics, 84(5):632-636, 2003. view

Left: Experimental setup. Right: Sample of Gamma Spectroscopy Output.

Modeling of Medical Accelerators for Simulations of Secondary Doses and Activations

In this project, we have developed a detailed Monte Carlo model of the Varian Clinac accelerator for various studies on patient secondary organ doses and accelerator activation during radiation therapy treatments. Over 100 components including Beam-line and shielding components and multi-leaf collimator were modeled. Validations of the Linac model were performed both in-field and out-of-field. The model was useful in the radiation treatment planning, operational health physics, shielding design and retrospective risk assessment. <top>

For more information about this project, please refer to:

• Bednarz B, Xu XG. Monte Carlo modeling of a 6 and 18 MV Varian Clinac medical accelerator for in-field and out-of-field dose calculations: development and validation. Physics in Medicine and Biology, 54(4): N43-56, 2009 view
• Bednarz B, Xu XG. Monte Carlo modeling of a 6 and 18 MV Varian Clinac medical accelerator for in-field and out-of-field dose calculations: applications. Physics in Medicine and Biology. submitted in 2009.

Detailed model of Varian Clinac operating at 6- and18-MV in MCNPX

Deformable and Posture Changeable Adult Male and Female Phantoms

Whole-body computational phantoms of various sizes and postures are needed for the assessment of organ doses in Computed Tomography (CT) imaging, internal nuclear medicine, and external-beam radiation treatment procedures. The RPI reference adult male (RPI-AM) and female (RPI-AF) phantoms are deformable, mesh-based computational phantoms. The key feature of these phantoms is their ability to change their body size and organ shapes according to anthropometric data such as: 1) whole-body size data defined as height and weight by the National Health and Nutrition Examination Survey (NHANES) and 2) specific internal organ volume and mass data derived from a cumulative pattern analysis of the International Commission on Radiological Protection (ICRP) recommended reference values. The deformed phantoms coupled with the correct tissue density and elemental composition information can then be imported into a Monte Carlo radiation transport code. As few Monte Carlo codes currently accept the mesh-based phantoms directly, these phantoms are first converted to voxel geometry. Future research will be needed to develop a method to run Monte Carlo calculations in mesh geometry. The figures below demonstrate how one can take advantage of the deformability of the RPI reference adult phantoms to create percentiles- and postures- specific adult phantoms that can be used to represent a wide spectrum of patients or workers.<top>

            58.5kg    66.3kg    73.1kg     86.4kg    103.8kg                 46.5kg    55.8kg   64.0kg   78.9kg    95.9kg

RPI-AMs (in 176 cm height ) and RPI-AFs (163 cm height) with different weights

         (a) Standing                   (b) Raised-arms                        (c) Sitting                                   (d) Walking

RPI-AM(Right) and RPI-AF(Left) in different postures

Modeling of CT Scanners and Development of CT Dose Reporting Software

In recent years, Computed Tomography (CT) has become an increasingly popular diagnostic imaging modality. As CT usage grows and the technology and scan protocols evolve, it will be necessary to monitor for potential increases in the radiation risk associated with this modality. The purpose of this work was to develop and validate CT scanner models for implementation in the Monte Carlo N-Particle eXtended (MCNPX) radiation transport code. Detailed components were modeled for a multiple-detector CT (MDCT) scanner and two image-guided radiation therapy cone beam CT (CBCT) scanners. When coupled with computational patient phantom, such as the deformable RPI-AM and RPI-AF or RPI Pregnant Female models, these CT scanner models can be used to estimate the radiation dose received by various organs and tissues. The MDCT scanner model was validated by comparing simulated results against CTDI phantom measurements and dose profiles which were found in the literature. The MCDT source movement along the helical trajectory was simulated using the pitch. The results of the Monte Carlo dose calculations are stored in a database which can be accessed through an in-house developed CT dose reporting software. This software allows the user to enter parameters about the scan (e.g. current and scan length) and patient (e.g. gender, height, weight). The output of the software is a report which indicates the estimated dose received by various organs and the estimated effective dose received by the patient.<top>

For more information about this project, please refer to:

• Gu J, Bednarz B, Caracappa PF, Xu XG. The development, validation and application of a multi-detector CT (MDCT) scanner model for assessing organ doses to the pregnant patient and her fetus using Monte Carlo methods. Phys. Med. Biol. 54 (2009) 2699-2717. view
• Gu J, Bednarz B, Xu XG, Jiang SB. Assessment of Patient Organ Doses and Effective Doses using the VIP-MAN Adult Male Phantom for Selected CONE-BEAM CT Imaging Procedures during Image Guided Radiation Therapy. Radiat. Prot. Dosim., 131(4):431-443, 2008. view

Left: The geometry of an MDCT scanner source.
Right: Representation of the 16 x-ray sources and bowtie filters used to simulate a single axial CT scan.
Bottom: The GUI of the CT dose reporting software.

GPU-based Monte Carlo Acceleration for Nuclear Reactor Analysis

Monte Carlo simulation is ideally suited for solving Boltzmann neutron transport equation in inhomogeneous media. However, routine applications require the computation time to be reduced to minutes in a desktop system. The interest in adopting GPUs for Monte Carlo acceleration is rapidly mounting, fueled partially by the parallelism afforded by the latest GPU technologies and the challenge to perform full-size reactor core analysis on a routine basis.
     In this project, Monte Carlo codes for a fixed-source neutron transport and k-effective problems were developed for CPU and GPU environments to evaluate issues associated with computational speedup. A CPU implementation of the Monte Carlo neutron transport code was written in C++ and compiled with Microsoft Visual studio 2008. The GPU hardware used in this study is NVIDIA TeslaTM C2050 computing processor. Based on the new NVIDIA "Fermi" architecture, the TeslaTM C2050 is designed for solving large-scale computing problems more efficiently. The CPU code was run on a Windows Vista desktop with an Intel Xeon® E5507 Quad-core 2.26-GHz with 6-GB memory. The GPU-based code was optimized with 256 threads per block and 48 warps per SM to keep GPU fully occupied. Such baseline Monte Carlo neutron transport experiments were designed to evaluate the GPU acceleration afforded by the GPUs.
     The results suggest that a speedup factor of >30 in Monte Carlo radiation transport of neutrons is within reach using the state-of-the-art GPU technologies. For a task of voxelizing unstructured mesh phantom geometry that is more parallel in nature, the speedup of >45 is easily obtained. Successful implementation of Monte Carlo schemes on GPU will require considerable effort, especially when a production-scale Monte Carlo radiation transport code is considered. Given the prediction that future-generation GPU products will likely bring exponentially improved computing power and performances, innovative hardware and software solutions may make it possible to meet the "Kord Smith Challenge" in the next several years.<top>

NVIDIA C2050

• Ding A, Liu T, Liang C, Ji W, Shepard MS, Xu XG, and Brown FB. Evaluation of speedup of Monte Carlo calculations of simple reactor physics problems coded for the GPU/CUDA environment. International Conference on Mathematics and Computational Methods Applied to Nuclear Science and Engineering (M&C 2011),Rio de Janeiro, RJ, Brazil, May 8-12, 2011.
  

Proton Telescope for Image-Guided Proton Radiotherapy

One challenge in proton cancer treatment is to assess the proton range fluctuations due to changes in tissue density, tissue thickness, tumor size, or organ deformation. The variation may be magnified in the case of lung cancer treatment with proton beams. For example, a 5 mm water-equivalent distal margin may expand to 25 mm in lung for a 0.2 g/cc lung density. These pathlength variations underscore the need for feedback and control of the proton beam, to provide estimates of range variation on either a moment to moment or day to day timeframe. In this project, Monte Carlo methods are used to simulate and optimize a time-resolved proton range telescope (TRRT) in localization of intrafractional and interfractional motions of lung tumor and in quantification of proton range variations.
     The Monte Carlo N-Particle eXtended (MCNPX) code with a particle-tracking feature was employed to evaluate the TRRT performance, especially in visualizing and quantifying proton range variations during respiration. Protons of 230 MeV were tracked one by one as they pass through position detectors, patient 4DCT phantom, and finally scintillator detectors that measured residual ranges. The energy response of the scintillator telescope was investigated. Mass density and elemental composition of tissues were defined for 4DCT data.
     Proton water equivalent length (WEL) was deduced by a Most Likely Path (MLP) reconstruction algorithm that incorporates the Molière theory and Bethe-Block equation to improve the image quality. 4DCT data for three patients were used to visualize and measure tumor motion and WEL variations. The tumor trajectories extracted from the WEL map were found to be within ~1mm agreement with direct 4DCT measurement. Quantitative WEL variations studies showed that the proton radiograph is good representation of WEL changes from entrance to distal of the target.
     MCNPX simulation results showed that TRRT can accurately track the motion of the tumor and detect the WEL variations. Image quality was optimized by choosing proton energy, testing parameters of image reconstruction algorithm, and comparing with ground truth 4DCT. The future study will demonstrate the feasibility of using the time resolved proton radiography as an imaging tool for proton treatments of lung tumors.<top>

Conceptual design of the time-resolved proton range telescope system consisting of two pairs of position detecting GEM chambers and a range telescope that measures the residual range of protons.

Proton WEL maps are windowed and leveled to visualize tumor at each of the 10 respiratory phases of patient #1's (beginning at peak inhalation). The tumor is clearly seen in all phases.

• Han B, Chen GTY, Xu XG. Proton Radiography and Fluoroscopy of Lung Tumors: A Monte Carlo Study using Patient-specific 4DCT Phantoms. Medical Physics. In print and schedule for publish in Mar 2011.
  

Applications of Compressive Sensing to Nuclear Science and Engineering

Compressive sensing (CS) is a 5-year old theory that has already resulted in an extremely large number of publications in the literature and that has the potential to impact every field of engineering and applied science that has to do with data acquisition and processing. This paper introduces the mathematics, presents a simple demonstration of radiation dose reduction in x-ray CT imaging, and discusses the potential applications in nuclear science and engineering. The results show that by using a CS reconstruction algorithm the number of projection angles required to generate a decent image (and hence the radiation dose received by the patient) can be reduced by a factor of ten.<top>

Comparison of reconstructed images of the Shepp-Logan phantom for different number of projection angles using two algorithms: Filtered Back Projection (top row) and CS TV-regularization (bottom row).

• Mille M, Su L, Yazici B, Xu XG. Opportunities and Challenges in Applying the Compressive Sensing Framework to Nuclear Science and Engineering. International Conference on Mathematics and Computational Methods Applied to Nuclear Science and Engineering (M&C 2011), Rio de Janeiro, RJ, Brazil, May 8-12, 2011.

Animating Computational Human Phantoms Using 3-D Motion-Capture Data

The RPI Radiation Measurement and Dosimetry Group has begun research into using 3-D motion-capture technology to model animated computational human phantoms for use in radiation dosimetry simulation. The modeling of realistic human motion is imperative to properly accounting for bodily self-shielding when simulating radiation-transport. In the past, our group has done extensive work on developing human phantom models, as well as using phantoms in virtual reality environments with simulated radiation (for example, simulating a nuclear plant accident such as the accident at Japan's Fukushima Daiichi nuclear power plant in early 2011). We plan to implement our newly-developed animated phantoms on such VR platforms in order to provide tools that will be very useful for personnel, such as nuclear power plant operators and radiologic technicians, who work with radiation technologies on a daily basis. Such tools will also be useful for the purposes of emergency-planning and dose-reconstruction. Click here to watch a brief video demonstrating this work.<top>

Walking Computational Human Phantoms for Radiation Dosimetry Simulation