Cone-beam Computed Tomography Projection Images Research Articles

Improved compact representation method for Monte Carlo phase spaces

Phase space files used in particle Monte Carlo simulations can be large, and thereby awkward to distribute and slow to load into memory. To circumvent this, compact source models and phase space representations have been developed. In this study, the aim is to increase the versatility of a previously presented method, based on principal component analysis whitening, by evaluating several modifications to the method on simulated cone-beam computed tomography projection images. Scaling the phase space components before whitening was seen to give the largest improvement and allowed for accurate modelling of phase spaces with a shifted detector and/or a bowtie filter present, configurations where the original method struggled. The modified method reproduced results using the original phase space to within a few percent both in energy fluence on the detector and in simulated projection and scatter images of a patient pelvis, without obvious position-dependence of the errors. Hence, the modified method was deemed sufficiently accurate for most applications relying on cone-beam computed tomography simulations. The final aim is to use the more versatile method for a broader range of phase space-based applications such as linear accelerator beam simulations and room shielding calculations.

Tumor phase recognition using cone-beam computed tomography projections and external surrogate information.

Directly extracting the respiratory phase pattern of the tumor using cone-beam computed tomography (CBCT) projections is challenging due to the poor tumor visibility caused by the obstruction of multiple anatomic structures on the beam's eye view. Predicting tumor phase information using external surrogate also has intrinsic difficulties as the phase patterns between surrogates and tumors are not necessary to be congruent. In this work, we developed an algorithm to accurately recover the primary oscillation components of tumor motion using the combined information from both CBCT projections and external surrogates. The algorithm involved two steps. First, a preliminary tumor phase pattern was acquired by applying local principal component analysis (LPCA) on the cropped Amsterdam Shroud (AS) images. In this step, only the cropped image of the tumor region was used to extract the tumor phase pattern in order to minimize the impact of pattern recognition from other anatomic structures. Second, by performing multivariate singular spectrum analysis (MSSA) on the combined information containing both external surrogate signal and the original waveform acquired in the first step, the primary component of the tumor phase oscillation was recovered. For the phantom study, a QUASAR respiratory motion phantom with a removable tumor-simulator insert was employed to acquire CBCT projection images. A comparison between LPCA only and our method was assessed by power spectrum analysis. Also, the motion pattern was simulated under the phase shift or various amplitude conditions to examine the robustness of our method. Finally, anatomic obstruction scenarios were simulated by attaching a heart model, PVC tubes, and RANDO® phantom slabs to the phantom, respectively. Each scenario was tested with five real-patient breathing patterns to mimic real clinical situations. For the patient study, eight patients with various tumor locations were selected. The performance of our method was then evaluated by comparing the reference waveform with the extracted signal for overall phase discrepancy, expiration phase discrepancy, peak, and valley accuracy. In tests of phase shifts and amplitude variations, the overall peak and valley accuracy was -0.009±0.18sec, and no time delay was found compared to the reference. In anatomical obstruction tests, the extracted signal had 1.6±1.2 % expiration phase discrepancy, -0.12±0.28sec peak accuracy, and 0.01±0.15sec valley accuracy. For patient studies, the extracted signal using our method had -1.05±3.0 % overall phase discrepancy, -1.55±1.45% expiration phase discrepancy, 0.04±0.13sec peak accuracy, and -0.01±0.15sec valley accuracy, compared to the reference waveforms. An innovative method capable of accurately recognizing tumor phase information was developed. With the aid of extra information from the external surrogate, an improvement in prediction accuracy, as compared with traditional statistical methods, was obtained. It enables us to employ it as the ground truth for 4D-CBCT reconstruction, gating treatment, and other clinic implementations that require accurate tumor phase information.

TH-CD-207A-04: Optimized Respiratory Gating for Abnormal Breathers in Pancreatic SBRT

Purpose: Pancreatic SBRT is uniquely challenging due to both the erratic/unstable motion of the pancreas and the close proximity of the radiosensitive small bowel. Respiratory gating can mitigate this effect, but the irregularity of motion severely affects traditional phase-based gating. The purpose of this study was to analyze real-time motion data of pancreatic tumors to optimize the efficacy and accuracy of respiratory gating, with the overall goal of enabling dose escalated pancreatic SBRT. Methods: Fifteen pancreatic SBRT patients received 30–33 Gy in 5 fractions on a Varian TrueBeam STx unit. Abdominal compression was used to reduce the amplitude of tumor motion, and daily cone-beam computed tomography (CBCT) scans were acquired prior to each treatment for target localization purposes. For this study, breathing data (phase and amplitude) were collected during each CBCT scan using Varian's Real-Time Position Management system. An in-house template matching technique was used to track the superior-inferior motion of implanted fiducial markers in CBCT projection images. Using tumor motion and breathing data, phase-based or amplitude-based respiratory gating was simulated for all 75 fractions, targeting either end-exhalation or end-inhalation phases of breathing. Results: For the average patient, gating at end-exhalation offered the best reductions in effective motion for equal duty cycles. However, optimal central phase angle varied widely (range: 0–92%, mean±SD: 49±12%), and phase-based gating windows typically associated with end-exhalation (i.e., “30–70%”) were rarely ideal. Amplitude-based gating significantly outperformed phase-based gating, with average effective ranges for amplitude-based gating 25% lower than phase-based gating ranges (as much as 73% lower). Amplitude-based gating was consistently better suited to accommodate abnormal breathing patterns. For both phase-based and amplitude-based gating, end-exhalation provided significantly better results than end-inhalation. Conclusion: Amplitude-based gating reliably outperformed phase-based gating, and end-exhalation was more suitable than end-inhalation. These results will be used to guide future dose-escalation trials. Research funding provided by Varian Medical Systems to Miften and Jones.

Robust breathing signal extraction from cone beam CT projections based on adaptive and global optimization techniques

We present a study of extracting respiratory signals from cone beam computed tomography (CBCT) projections within the framework of the Amsterdam Shroud (AS) technique. Acquired prior to the radiotherapy treatment, CBCT projections were preprocessed for contrast enhancement by converting the original intensity images to attenuation images with which the AS image was created. An adaptive robust z-normalization filtering was applied to further augment the weak oscillating structures locally. From the enhanced AS image, the respiratory signal was extracted using a two-step optimization approach to effectively reveal the large-scale regularity of the breathing signals. CBCT projection images from five patients acquired with the Varian Onboard Imager on the Clinac iX System Linear Accelerator (Varian Medical Systems, Palo Alto, CA) were employed to assess the proposed technique. Stable breathing signals can be reliably extracted using the proposed algorithm. Reference waveforms obtained using an air bellows belt (Philips Medical Systems, Cleveland, OH) were exported and compared to those with the AS based signals. The average errors for the enrolled patients between the estimated breath per minute (bpm) and the reference waveform bpm can be as low as −0.07 with the standard deviation 1.58. The new algorithm outperformed the original AS technique for all patients by 8.5% to 30%. The impact of gantry rotation on the breathing signal was assessed with data acquired with a Quasar phantom (Modus Medical Devices Inc., London, Canada) and found to be minimal on the signal frequency. The new technique developed in this work will provide a practical solution to rendering markerless breathing signal using the CBCT projections for thoracic and abdominal patients.

Automated patient setup and gating using cone beam computed tomography projections

In radiation therapy, fiducial markers are often implanted near tumors and used for patient positioning and respiratory gating purposes. These markers are then used to manually align the patients by matching the markers in the cone beam computed tomography (CBCT) reconstruction to those in the planning CT. This step is time-intensive and user-dependent, and often results in a suboptimal patient setup. We propose a fully automated, robust method based on dynamic programming (DP) for segmenting radiopaque fiducial markers in CBCT projection images, which are then used to automatically optimize the treatment couch position and/or gating window bounds. The mean of the absolute 2D segmentation error of our DP algorithm is mm for 87 markers on 39 patients. Intrafraction images were acquired every 3 s during treatment at two different institutions. For gated patients from Institution A (8 patients, 40 fractions), the DP algorithm increased the delivery accuracy ( versus , p < 0.01) compared to the manual setup using kV fluoroscopy. For non-gated patients from Institution B (6 patients, 16 fractions), the DP algorithm performed similarly ( mm versus mm, p = 0.48) compared to the manual setup matching the fiducial markers in the CBCT to the mean position. Our proposed automated patient setup algorithm only takes 1–2 s to run, requires no user intervention, and performs as well as or better than the current clinical setup.

Optimizing cone beam CT scatter estimation in egs_cbct for a clinical and virtual chest phantom

Cone beam computed tomography (CBCT) image quality suffers from contamination from scattered photons in the projection images. Monte Carlo simulations are a powerful tool to investigate the properties of scattered photons.egs_cbct, a recent EGSnrc user code, provides the ability of performing fast scatter calculations in CBCT projection images. This paper investigates how optimization of user inputs can provide the most efficient scatter calculations. Two simulation geometries with two different x-ray sources were simulated, while the user input parameters for the efficiency improving techniques (EITs) implemented inegs_cbct were varied. Simulation efficiencies were compared to analog simulations performed without using any EITs. Resulting scatter distributions were confirmed unbiased against the analog simulations. The optimal EIT parameter selection depends on the simulation geometry and x-ray source. Forced detection improved the scatter calculation efficiency by 80%. Delta transport improved calculation efficiency by a further 34%, while particle splitting combined with Russian roulette improved the efficiency by a factor of 45 or more. Combining these variance reduction techniques with a built-in denoising algorithm, efficiency improvements of 4 orders of magnitude were achieved. Using the built-in EITs inegs_cbct can improve scatter calculation efficiencies by more than 4 orders of magnitude. To achieve this, the user must optimize the input parameters to the specific simulation geometry. Realizing the full potential of the denoising algorithm requires keeping the statistical uncertainty below a threshold value above which the efficiency drops exponentially.

Three-dimensional, Time-Resolved, Intrafraction Motion Monitoring Throughout Stereotactic Liver Radiation Therapy on a Conventional Linear Accelerator

To investigate the time-resolved 3-dimensional (3D) internal motion throughout stereotactic body radiation therapy (SBRT) of tumors in the liver using standard x-ray imagers of a conventional linear accelerator. Ten patients with implanted gold markers received 11 treatment courses of 3-fraction SBRT in a stereotactic body-frame on a conventional linear accelerator. Two pretreatment and 1 posttreatment cone-beam computed tomography (CBCT) scans were acquired during each fraction. The CBCT projection images were used to estimate the internal 3D marker motion during CBCT acquisition with 11-Hz resolution by a monoscopic probability-based method. Throughout the treatment delivery by conformal or volumetric modulated arc fields, simultaneous MV portal imaging (8 Hz) and orthogonal kV imaging (5 Hz) were applied to determine the 3D marker motion using either MV/kV triangulation or the monoscopic method when marker segmentation was unachievable in either MV or kV images. The accuracy of monoscopic motion estimation was quantified by also applying monoscopic estimation as a test for all treatments during which MV/kV triangulation was possible. Root-mean-square deviations between monoscopic estimations and triangulations were less than 1.0 mm. The mean 3D intrafraction and intrafield motion ranges during liver SBRT were 17.6 mm (range, 5.6-39.5 mm) and 11.3 mm (2.1-35.5mm), respectively. The risk of large intrafraction baseline shifts correlated with intrafield respiratory motion range. The mean 3D intrafractional marker displacement relative to the first CBCT was 3.4 mm (range, 0.7-14.5 mm). The 3D displacements exceeded 8.8 mm 10% of the time. Highly detailed time-resolved internal 3D motion was determined throughout liver SBRT using standard imaging equipment. Considerable intrafraction motion was observed. The demonstrated methods provide a widely available approach for motion monitoring that, combined with motion-adaptive treatment techniques, has the potential to improve the accuracy of radiation therapy for moving targets.

Mitigation of motion artifacts in CBCT of lung tumors based on tracked tumor motion during CBCT acquisition

An algorithm capable of mitigating respiratory motion blurring artifacts in cone-beam computed tomography (CBCT) lung tumor images based on the motion of the tumor during the CBCT scan is developed. The tumor motion trajectory and probability density function (PDF) are reconstructed from the acquired CBCT projection images using a recently developed algorithm Lewis et al (2010 Phys. Med. Biol. 55 2505–22). Assuming that the effects of motion blurring can be represented by convolution of the static lung (or tumor) anatomy with the motion PDF, a cost function is defined, consisting of a data fidelity term and a total variation regularization term. Deconvolution is performed through iterative minimization of this cost function. The algorithm was tested on digital respiratory phantom, physical respiratory phantom and patient data. A clear qualitative improvement is evident in the deblurred images as compared to the motion-blurred images for all cases. Line profiles show that the tumor boundaries are more accurately and clearly represented in the deblurred images. The normalized root-mean-squared error between the images used as ground truth and the motion-blurred images are 0.29, 0.12 and 0.30 in the digital phantom, physical phantom and patient data, respectively. Deblurring reduces the corresponding values to 0.13, 0.07 and 0.19. Application of a −700 HU threshold to the digital phantom results in tumor dimension measurements along the superior–inferior axis of 2.8, 1.8 and 1.9 cm in the motion-blurred, ground truth and deblurred images, respectively. Corresponding values for the physical phantom are 3.4, 2.7 and 2.7 cm. A threshold of −500 HU applied to the patient case gives measurements of 3.1, 1.6 and 1.7 cm along the SI axis in the CBCT, 4DCT and deblurred images, respectively. This technique could provide more accurate information about a lung tumor's size and shape on the day of treatment.

Scatter kernel estimation with an edge-spread function method for cone-beam computed tomography imaging

The clinical applications of kilovoltage x-ray cone-beam computed tomography (CBCT) have been compromised by the limited quality of CBCT images, which typically is due to a substantial scatter component in the projection data. In this paper, we describe an experimental method of deriving the scatter kernel of a CBCT imaging system. The estimated scatter kernel can be used to remove the scatter component from the CBCT projection images, thus improving the quality of the reconstructed image. The scattered radiation was approximated as depth-dependent, pencil-beam kernels, which were derived using an edge-spread function (ESF) method. The ESF geometry was achieved with a half-beam block created by a 3 mm thick lead sheet placed on a stack of slab solid-water phantoms. Measurements for ten water-equivalent thicknesses (WET) ranging from 0 cm to 41 cm were taken with (half-blocked) and without (unblocked) the lead sheet, and corresponding pencil-beam scatter kernels or point-spread functions (PSFs) were then derived without assuming any empirical trial function. The derived scatter kernels were verified with phantom studies. Scatter correction was then incorporated into the reconstruction process to improve image quality. For a 32 cm diameter cylinder phantom, the flatness of the reconstructed image was improved from 22% to 5%. When the method was applied to CBCT images for patients undergoing image-guided therapy of the pelvis and lung, the variation in selected regions of interest (ROIs) was reduced from >300 HU to <100 HU. We conclude that the scatter reduction technique utilizing the scatter kernel effectively suppresses the artifact caused by scatter in CBCT.

SU-FF-J-84: Accuracy and Feasibility of Cone Beam Computed Tomography (CBCT) for Stereotactic Radiosurgery (SRS) Setup

Purpose: To investigate the accuracy and feasibility of cone beam computed tomography (CBCT) for stereotactic radiosurgery (SRS) setup. Method and Materials: A stereotactic BRW frame was attached to a Rando head phantom with imbedded radio-opaque markers (1mm in diameter) simulating the target locations. The head phantom with a CT localizer was first simulated for radiosurgery planning on a conventional CT scanner and then scanned on a treatment machine using CBCT. The differences between the conventional and CBCT localizations of the target positions were computed using a radiosurgery planning software. The translational corrections for target positions were calculated as the differences between the CBCT coordinates of the machine isocenter and the planned isocenters, which were the planned BRW coordinates from the conventional CT scan multiplied by the transformation matrix between the BRW and CBCT coordinate systems on the CBCT scan. The setup accuracy of CBCT was assessed from the analysis of orthogonal projection images for each radio-opaque target at the machine isocenter. Results: All nine fiducial markers of BRW localizer were successfully identified on all but one slice of the CBCT scan. The average localization difference between the conventional CT and CBCT BRW target coordinates was 0.28mm (SD 0.10mm). The mean distance error for all the radio-opaque targets localized using the CBCT and orthogonal projection images was 1.28mm (SD 0.61 mm). The major contributing factor to this cumulative setup error was the uncertainty in the superior-inferior direction due to the 2mm slice thickness in conventional CT. Conclusion: The CBCT image guidance can be used to setup SRS patients within accuracy comparable to the current SRS standard using an external fiducial system. The technique described here can serve as a gold standard for evaluating the accuracy of alternative immobilization and setup devices for SRS.

TU‐D‐I‐611‐05: Monte Carlo Investigation of Scatter Contribution to Kilovoltage Cone‐Beam Computed Tomography Images

Purpose: In this work, we evaluate the scatter radiation present in projection images produced by kilovoltage (kV) cone‐beam computed tomography (CBCT) using Monte Carlo (MC) calculations. We also demonstrate that MC simulations can be used to reduce scatter in kV CBCT reconstructed images. Method and Materials: All images were acquired using a kV CBCT bench‐top system composed of an x‐ray tube, a rotation stage and a CsI flat‐panel imager. The EGSnrc MC code was used to fully model the flat‐panel imager and the x‐ray source for energies of 80, 100, and 120 kVp. The x‐ray source model was validated using first and second half‐value layer (HVL) and profile measurements. MC simulations and measured images were obtained for a cylindrical water phantom and for the head of the anthropomorphic Rando phantom. A modified version of the DOSXYZnrc program was used to score phase space files with identified scattered and non‐scattered particles after the phantoms. The cone angle and the source‐to‐detector distance (SDD) were varied during the simulations to determine their effect on scatter radiation in projection images. MC predictions of the scatter distribution were subtracted from the CBCT projection images prior to the reconstruction to improve the reconstructed image quality. Results: The simulated and measured first and second HVL agree within 2% except for the 80 kVp beam where the agreement is within 9%. The measured and simulated profiles agree within 4%. Simulated and measured projection images are in good agreement. Decreasing the cone angle and increasing the SDD decrease the amount of scatter. On the other hand, the phantom geometry has little effect on the scatter radiation distribution. Conclusion: MC simulations were used to characterize the scatter radiation from different geometries. It was shown that it is possible to use MC simulations to reduce the scatter in the reconstructed image.