Development and validation of a measurement-based source model for kilovoltage cone-beam CT Monte Carlo dosimetry simulations

Generation of material-specific energy deposition kernels for kilovoltage x-ray dose calculations

PURPOSE

Dose calculation of kilovoltage x rays used in Image-Guided Radiotherapy has been investigated in recent years using various methods. Among these methods are model-based ones that suffer from inaccuracies in high-density materials and at interfaces when used in the kilovoltage energy range. The main reason for this is the use of water energy deposition kernels and simplifications employed such as density scaling in heterogeneous media. The purpose of this study was to produce and characterize material-specific energy deposition kernels, which could be used for dose calculations in this energy range. These kernels will also have utility in dose calculations in superficial radiation therapy and orthovoltage beams utilized in small animal irradiators.

METHODS

Water energy deposition kernels with various resolutions; and high-resolution, material-specific energy deposition kernels were generated in the energy range of 10-150 kVp, using the EGSnrc Monte Carlo toolkit. The generated energy deposition kernels were further characterized by calculating the effective depth of penetration, the effective radial distance, and the effective lateral distance. A simple benchmarking of the kernels against Monte Caro calculations has also been performed.

RESULTS

There was good agreement with previously reported water kernels, as well as between kernels with different resolution. The evaluation of effective depth of penetration, and radial and laterals distances, defines the relationship between energy, material density, and the shape of the material-specific kernels. The shape of these kernels becomes more forwardly scattered as the energy and material density are increased. The comparison of the dose calculated using the kernels with Monte Carlo provides acceptable results.

CONCLUSIONS

Water and material-specific energy deposition kernels in the kilovoltage energy range have been generated, characterized, and compared to previous work. These kernels will have utility in dose calculations in this energy range once algorithms capable of employing them are fully developed.

A new cone-beam computed tomography dosimetry method providing optimal measurement positions: A Monte Carlo study

PURPOSE

The conventional weighted computed tomography dose index (CTDI_w) may not be suitable for cone-beam computed tomography (CBCT) dosimetry because a cross-sectional dose distribution is angularly inhomogeneous owing to partial angle irradiations. This study was conducted to develop a new dose metric (f(0)^CB_w) for CBCT dosimetry to determine a more accurate average dose in the central cross-sectional plane of a cylindrical phantom using Monte Carlo simulations.

METHODS

First, cross-sectional dose distributions of cylindrical polymethyl methacrylate phantoms over a wide range of phantom diameters (8-40 cm) were calculated for various CBCT scan protocols. Then, by obtaining linear least-squares fits of the full datasets of the cross-sectional dose distributions, the optimal radial positions, which represented measurement positions for the average phantom dose, were determined. Finally, the f(0)^CB_w method was developed by averaging point doses at the optimal radial positions of the phantoms. To demonstrate its validity, the relative differences between the average doses and each dose index value were estimated for the devised f(0)^CB_w, conventional CTDI_w, and Haba's CTDI_w methods, respectively.

RESULTS

The relative differences between the average doses and each dose index value were within 4.1%, 16.7%, and 11.9% for the devised, conventional CTDI_w, and Haba's CTDI_w methods, respectively.

CONCLUSIONS

The devised f(0)^CB_w value was calculated by averaging four "point doses" at 90° intervals and the optimal radial positions of the cylindrical phantom. The devised method can estimate the average dose more accurately than the previously developed CTDI_w methods for CBCT dosimetry.

Validation of histogram-based virtual source models for different IGRT kV-imaging systems

PURPOSE

Image-guided radiotherapy (IGRT) improves tumor control but its intensive use may entrain late side effects caused by the additional imaging doses. There is a need to better quantify the additional imaging doses, so they can be integrated in the therapeutic workflow. Currently, no dedicated software enables to compute patient-specific imaging doses on a wide range of systems and protocols. As a first step toward this objective, we propose a common methodology to model four different kV-imaging systems used in radiotherapy (Varian's OBI, Elekta's XVI, Brainlab's ExacTrac, and Accuray's Cyberknife) using a new type of virtual source model based on Monte Carlo calculations.

METHODS

We first describe our method to build a simplified description of the photon output, or virtual source models (VSMs), of each imaging system. Instead of being constructed using measurement data, as it is most commonly the case, our VSM is used as the summary of the phase-space files (PSFs) resulting from a first Monte Carlo simulation of the considered x-ray tube. Second, the VSM is used as a photon generator for a second MC simulation in which we compute the dose. Then, the proposed VSM is thoroughly validated against standard MC simulation using PSFs on the XVI system. Last, each modeled system is compared to profiles and depth-dose-curve measurements performed in homogeneous phantom.

RESULTS

Comparisons between PSF-based and VSM-based calculations highlight that VSMs could provide equivalent dose results (within 1% of difference) than PSFs inside the imaging field-of-view (FOV). In contrast, VSMs tend to underestimate (for up to 20%) calculated doses outside of the imaging FOV due to the assumptions underlying the VSM construction. In addition, we showed that the use of VSMs allows reducing calculation time by at least a factor of 2.8. Indeed, for identical simulation times, statistical uncertainties on dose distributions computed using VSMs were much lower than those obtained from PSF-based calculations.

CONCLUSIONS

For each of the four imaging systems, VSMs were successfully validated against measurements in homogeneous phantoms, and are therefore ready to be used for future preclinical studies in heterogeneous or anthropomorphic phantoms. The cross system modeling methodology developed here should enable, later on, to estimate precisely and accurately patient-specific 3D dose maps delivered during a large range of kV-imaging procedures.

Characterization and comparison of imaging contrast enhancement with PEG-functionalized gold nanoparticles in kV cone beam computed tomography and computed tomography imaging

The purpose of this study is to define a simplified method to accurately predict and characterize kV cone beam computed tomography (kV CBCT) and computed tomography (CT) image contrast enhancement from gold nanoparticles (GNPs). Parameters of the kV CBCT of a Varian Novalis Tx linear accelerator and of a GE LightSpeed 4 Big Bore CT machine were modeled using the MCNP 6.2 Monte Carlo code. A 0.25 × 0.25 cm² source, defined with a 100 kVp energy spectrum with appropriate filtration, was implemented in the MCNP6.2 model for kV CBCT, which also contained x- and y-blades and a full bowtie filter. A 1 cm³ cube of GNP solution (modeled as a mass percentage of gold in water) was placed 100 cm below the source. For the CT-simulator model, a source was defined with energy spectra for 80 and 140 kVp x-rays with appropriate filtration and angular spectrum. A 1 cm³ GNP solution was modeled as before and a detector was placed 40 cm below that. Attenuation coefficients of four GNP solutions were computed and Hounsfield unit (HU) values were calculated. The computed HU values were compared against experimentally measured values obtained by scanning batches of GNPs of various sizes and concentrations using a GE LightSpeed 4 Big Bore CT scanner at 80 kVp and 140 kVp energies, as well as the kV CBCT capability of a Varian Novalis Tx linear accelerator. HU analysis was carried out using Velocity Medical Solutions clinical CT image analysis software. The MCNP calculated HU values matched the measured values to within ± 5%. Image contrast enhancement analysis showed a total increase in HU of up to 223. The sample having the highest gold mass percentage tested showed the greatest increase in HU number compared to water.

The potential impact of ultrathin filter design on dosimetry and relative biological effectiveness in modern image-guided small animal irradiators

OBJECTIVE:

Modern image-guided small animal irradiators like the Xstrahl Small Animal Radiation Research Platform (SARRP) are designed with ultrathin 0.15 mm Cu filters, which compared with more heavily filtrated traditional cabinet-style biological irradiators, produce X-ray spectra weighted toward lower energies, impacting the dosimetric properties and the relative biological effectiveness (RBE). This study quantifies the effect of ultrathin filter design on relative depth dose profiles, absolute dose output, and RBE using Monte Carlo techniques.

METHODS:

The percent depth-dose and absolute dose output are calculated using kVDoseCalc and EGSnrc, respectively, while a tally based on the induction of double-strand breaks as a function of electron spectra invoked in PENELOPE is used to estimate the RBE.

RESULTS:

The RBE increases by >2.4% in the ultrathin filter design compared to a traditional irradiator. Furthermore, minute variations in filter thickness have notable effects on the dosimetric properties of the X-ray beam, increasing the percent depth dose (at 2 cm in water) by + 0.4%/0.01 mm Cu and decreasing absolute dose (at 2 cm depth in water) by -1.8%/0.01 mm Cu for the SARRP.

CONCLUSIONS:

These results show that modern image-guided irradiators are quite sensitive to small manufacturing variations in filter thickness, and show a small change in RBE compared to traditional X-ray irradiators.

ADVANCES IN KNOWLEDGE:

We quantify the consequences of ultrathin filter design in modern image-guided biological irradiators on relative and absolute dose, and RBE. Our results show these to be small, but not insignificant, suggesting laboratories transitioning between irradiators should carefully design their radiobiological experiments.

A measurement-based generalized source model for Monte Carlo dose simulations of CT scans

The goal of this study is to develop a generalized source model for accurate Monte Carlo dose simulations of CT scans based solely on the measurement data without a priori knowledge of scanner specifications. The proposed generalized source model consists of an extended circular source located at x-ray target level with its energy spectrum, source distribution and fluence distribution derived from a set of measurement data conveniently available in the clinic. Specifically, the central axis percent depth dose (PDD) curves measured in water and the cone output factors measured in air were used to derive the energy spectrum and the source distribution respectively with a Levenberg-Marquardt algorithm. The in-air film measurement of fan-beam dose profiles at fixed gantry was back-projected to generate the fluence distribution of the source model. A benchmarked Monte Carlo user code was used to simulate the dose distributions in water with the developed source model as beam input. The feasibility and accuracy of the proposed source model was tested on a GE LightSpeed and a Philips Brilliance Big Bore multi-detector CT (MDCT) scanners available in our clinic. In general, the Monte Carlo simulations of the PDDs in water and dose profiles along lateral and longitudinal directions agreed with the measurements within 4%/1 mm for both CT scanners. The absolute dose comparison using two CTDI phantoms (16 cm and 32 cm in diameters) indicated a better than 5% agreement between the Monte Carlo-simulated and the ion chamber-measured doses at a variety of locations for the two scanners. Overall, this study demonstrated that a generalized source model can be constructed based only on a set of measurement data and used for accurate Monte Carlo dose simulations of patients' CT scans, which would facilitate patient-specific CT organ dose estimation and cancer risk management in the diagnostic and therapeutic radiology.

A new technique to characterize CT scanner bow-tie filter attenuation and applications in human cadaver dosimetry simulations

PURPOSE

To present a noninvasive technique for directly measuring the CT bow-tie filter attenuation with a linear array x-ray detector.

METHODS

A scintillator based x-ray detector of 384 pixels, 307 mm active length, and fast data acquisition (model X-Scan 0.8c4-307, Detection Technology, FI-91100 Ii, Finland) was used to simultaneously detect radiation levels across a scan field-of-view. The sampling time was as short as 0.24 ms. To measure the body bow-tie attenuation on a GE Lightspeed Pro 16 CT scanner, the x-ray tube was parked at the 12 o'clock position, and the detector was centered in the scan field at the isocenter height. Two radiation exposures were made with and without the bow-tie in the beam path. Each readout signal was corrected for the detector background offset and signal-level related nonlinear gain, and the ratio of the two exposures gave the bow-tie attenuation. The results were used in the geant4 based simulations of the point doses measured using six thimble chambers placed in a human cadaver with abdomen/pelvis CT scans at 100 or 120 kV, helical pitch at 1.375, constant or variable tube current, and distinct x-ray tube starting angles.

RESULTS

Absolute attenuation was measured with the body bow-tie scanned at 80-140 kV. For 24 doses measured in six organs of the cadaver, the median or maximum difference between the simulation results and the measurements on the CT scanner was 8.9% or 25.9%, respectively.

CONCLUSIONS

The described method allows fast and accurate bow-tie filter characterization.

Technical Note: spektr 3.0-A computational tool for x-ray spectrum modeling and analysis

PURPOSE

A computational toolkit (spektr 3.0) has been developed to calculate x-ray spectra based on the tungsten anode spectral model using interpolating cubic splines (TASMICS) algorithm, updating previous work based on the tungsten anode spectral model using interpolating polynomials (TASMIP) spectral model. The toolkit includes a matlab (The Mathworks, Natick, MA) function library and improved user interface (UI) along with an optimization algorithm to match calculated beam quality with measurements.

METHODS

The spektr code generates x-ray spectra (photons/mm(2)/mAs at 100 cm from the source) using TASMICS as default (with TASMIP as an option) in 1 keV energy bins over beam energies 20-150 kV, extensible to 640 kV using the TASMICS spectra. An optimization tool was implemented to compute the added filtration (Al and W) that provides a best match between calculated and measured x-ray tube output (mGy/mAs or mR/mAs) for individual x-ray tubes that may differ from that assumed in TASMICS or TASMIP and to account for factors such as anode angle.

RESULTS

The median percent difference in photon counts for a TASMICS and TASMIP spectrum was 4.15% for tube potentials in the range 30-140 kV with the largest percentage difference arising in the low and high energy bins due to measurement errors in the empirically based TASMIP model and inaccurate polynomial fitting. The optimization tool reported a close agreement between measured and calculated spectra with a Pearson coefficient of 0.98.

CONCLUSIONS

The computational toolkit, spektr, has been updated to version 3.0, validated against measurements and existing models, and made available as open source code. Video tutorials for the spektr function library, UI, and optimization tool are available.

Evaluation of gantry speed on image quality and imaging dose for 4D cone-beam CT acquisition

Background

This study investigates the effect of gantry speed on 4DCBCT image quality and dose for the Varian On-Board Imager®.

Methods

A thoracic 4DCBCT protocol was designed using a 125 kVp spectrum. Image quality parameters were evaluated for 4DCBCT acquisition using Catphan® phantom with real-time position management™ system for gantry speeds varying between 1.0 to 6.0°/s. Superior-inferior motion of the phantom was executed using a sinusoidal waveform with five second period. Scans were retrospectively sorted into 4 phases (CBCT-4 ph) and 10 phases (CBCT-10 ph); average 4DCBCT (CBCT-ave), using all image data from the 4DCBCT acquisitions was also evaluated. The 4DCBCT images were evaluated using the following image quality metrics: spatial resolution, contrast-to-noise ratio (CNR), and uniformity index (UI). Additionally, Hounsfield unit (HU) sensitivity compared to a baseline CBCT and percent differences and RMS errors (RMSE) of excursion were also determined. Imaging dose was evaluated using an IBA CC13 ion chamber placed within CIRS Thorax phantom using the same sinusoidal motion and image acquisition settings as mentioned above.

Results

Spatial resolution decreased linearly from 5.93 to 3.82 lp/cm as gantry speed increased from 1.0 to 6.0°/s. CNR decreased linearly from 4.80 to 1.82 with gantry speed increasing from 1.0 to 6.0°/s, respectively. No noteworthy variations in UI, HU sensitivity, or excursion metrics were observed with changes in gantry speed. Ion chamber dose rates measured ranged from 2.30 (lung) to 5.18 (bone) E-3 cGy/mAs.

Conclusions

A quantitative analysis of the Varian OBI’s 4DCBCT capabilities was explored. Changing gantry speed changes the number of projections used for reconstruction, affecting both image quality and imaging dose if x-ray tube current is held constant. From the results of this study, a gantry speed between 2 and 3°/s was optimal when considering image quality, dose, and reconstruction time. The future of 4DCBCT clinical utility relies on further investigation of image acquisition and reconstruction optimization.

Organ doses can be estimated from the computed tomography (CT) dose index for cone-beam CT on radiotherapy equipment

Cone beam computed tomography (CBCT) systems are fitted to radiotherapy linear accelerators and used for patient positioning prior to treatment by image guided radiotherapy (IGRT). Radiotherapists' and radiographers' knowledge of doses to organs from CBCT imaging is limited. The weighted CT dose index for a reference beam of width 20 mm (CTDIw,ref) is displayed on Varian CBCT imaging equipment known as an On-Board Imager (OBI) linked to the Truebeam linear accelerator. This has the potential to provide an indication of organ doses. This knowledge would be helpful for guidance of radiotherapy clinicians preparing treatments. Monte Carlo simulations of imaging protocols for head, thorax and pelvic scans have been performed using EGSnrc/BEAMnrc, EGSnrc/DOSXYZnrc, and ICRP reference computational male and female phantoms to derive the mean absorbed doses to organs and tissues, which have been compared with values for the CTDIw,ref displayed on the CBCT scanner console. Substantial variations in dose were observed between male and female phantoms. Nevertheless, the CTDIw,ref gave doses within  ±21% for the stomach and liver in thorax scans and 2  ×  CTDIw,ref can be used as a measure of doses to breast, lung and oesophagus. The CTDIw,ref could provide indications of doses to the brain for head scans, and the colon for pelvic scans. It is proposed that knowledge of the link between CTDIw for CBCT should be promoted and included in the training of radiotherapy staff.

A rapid noninvasive characterization of CT x-ray sources

PURPOSE

The aim of this study is to generate spatially varying half value layers (HVLs) that can be used to construct virtual equivalent source models of computed tomography (CT) x-ray sources for use in Monte Carlo CT dose computations.

METHODS

To measure the spatially varying HVLs, the authors combined a cylindrical HVL measurement technique with the characterization of bowtie filter relative attenuation (COBRA) geometry. An apparatus given the name "HVL Jig" was fabricated to accurately position a real-time dosimeter off-isocenter while surrounded by concentric cylindrical aluminum filters (CAFs). In this geometry, each projection of the rotating x-ray tube is filtered by an identical amount of high-purity (type 1100 H-14) aluminum while the stationary radiation dose probe records an air kerma rate versus time waveform. The CAFs were progressively nested to acquire exposure data at increasing filtrations to calculate the HVL. Using this dose waveform and known setup geometry, each timestamp was related to its corresponding fan angle. Data were acquired using axial CT protocols (i.e., rotating tube and stationary patient table) at energies of 80, 100, and 120 kVp on a single CT scanner. These measurements were validated against the more laborious conventional step-and-shoot approach (stationary x-ray tube).

RESULTS

At each energy, HVL data points from the COBRA-cylinder technique were fit to a trendline and compared with the conventional approach. The average relative difference in HVL between the two techniques was 1.3%. There was a systematic overestimation in HVL due to scatter contamination.

CONCLUSIONS

The described method is a novel, rapid, accurate, and noninvasive approach that allows one to acquire the spatially varying fluence and HVL data using a single experimental setup in a minimum of three scans. These measurements can be used to characterize the CT beam in terms of the angle-dependent fluence and energy spectra along the bowtie filter direction, which can serve as input for accurate CT dose computations.

Modeling a superficial radiotherapy X-ray source for relative dose calculations

The purpose of this study was to empirically characterize and validate a kilovoltage (kV) X‐ray beam source model of a superficial X‐ray unit for relative dose calculations in water and assess the accuracy of the British Journal of Radiology Supplement 25 (BJR 25) percentage depth dose (PDD) data. We measured central axis PDDs and dose profiles using an Xstrahl 150 X‐ray system. We also compared the measured and calculated PDDs to those in the BJR 25. The Xstrahl source was modeled as an effective point source with varying spatial fluence and spectra. In‐air ionization chamber measurements were made along the x‐ and y‐axes of the X‐ray beam to derive the spatial fluence and half‐value layer (HVL) measurements were made to derive the spatially varying spectra. This beam characterization and resulting source model was used as input for our in‐house dose calculation software (kVDoseCalc) to compute radiation dose at points of interest (POIs). The PDDs and dose profiles were measured using 2, 5, and 15 cm cone sizes at 80, 120, 140, and 150 kVp energies in a scanning water phantom using IBA Farmer‐type ionization chambers of volumes 0.65 and 0.13 cc, respectively. The percent difference in the computed PDDs compared with our measurements range from −4.8% to 4.8%, with an overall mean percent difference and standard deviation of 1.5% and 0.7%, respectively. The percent difference between our PDD measurements and those from BJR 25 range from −14.0% to 15.7%, with an overall mean percent difference and standard deviation of 4.9% and 2.1%, respectively — showing that the measurements are in much better agreement with kVDoseCalc than BJR 25. The range in percent difference between kVDoseCalc and measurement for profiles was −5.9% to 5.9%, with an overall mean percent difference and standard deviation of 1.4% and 1.4%, respectively. The results demonstrate that our empirically based X‐ray source modeling approach for superficial X‐ray therapy can be used to accurately compute relative dose in a homogeneous water‐equivalent medium. They also show limitations in the accuracy of the BJR 25 PDD data.

PACS number: 87.55.kh

Experimental validation of a kilovoltage x-ray source model for computing imaging dose

PURPOSE

To introduce and validate a kilovoltage (kV) x-ray source model and characterization method to compute absorbed dose accrued from kV x-rays.

METHODS

The authors propose a simplified virtual point source model and characterization method for a kV x-ray source. The source is modeled by: (1) characterizing the spatial spectral and fluence distributions of the photons at a plane at the isocenter, and (2) creating a virtual point source from which photons are generated to yield the derived spatial spectral and fluence distribution at isocenter of an imaging system. The spatial photon distribution is determined by in-air relative dose measurements along the transverse (x) and radial (y) directions. The spectrum is characterized using transverse axis half-value layer measurements and the nominal peak potential (kVp). This source modeling approach is used to characterize a Varian(®) on-board-imager (OBI(®)) for four default cone-beam CT beam qualities: beams using a half bowtie filter (HBT) with 110 and 125 kVp, and a full bowtie filter (FBT) with 100 and 125 kVp. The source model and characterization method was validated by comparing dose computed by the authors' inhouse software (kVDoseCalc) to relative dose measurements in a homogeneous and a heterogeneous block phantom comprised of tissue, bone, and lung-equivalent materials.

RESULTS

The characterized beam qualities and spatial photon distributions are comparable to reported values in the literature. Agreement between computed and measured percent depth-dose curves is ⩽ 2% in the homogeneous block phantom and ⩽ 2.5% in the heterogeneous block phantom. Transverse axis profiles taken at depths of 2 and 6 cm in the homogeneous block phantom show an agreement within 4%. All transverse axis dose profiles in water, in bone, and lung-equivalent materials for beams using a HBT, have an agreement within 5%. Measured profiles of FBT beams in bone and lung-equivalent materials were higher than their computed counterparts resulting in an agreement within 2.5%, 5%, and 8% within solid water, bone, and lung, respectively.

CONCLUSIONS

The proposed virtual point source model and characterization method can be used to compute absorbed dose in both the homogeneous and heterogeneous block phantoms within of 2%-8% of measured values, depending on the phantom and the beam quality. The authors' results also provide experimental validation for their kV dose computation software, kVDoseCalc.