Reduction of motion blurring artifacts using respiratory gated CT in sinogram space: A quantitative evaluation

First clinical implementation of real-time, real anatomy tracking and radiation beam control

PURPOSE

We describe the acceptance testing, commissioning, periodic quality assurance, and workflow procedures developed for the first clinically implemented magnetic resonance imaging-guided radiation therapy (MR-IGRT) system for real-time tracking and beam control.

METHODS

The system utilizes real-time cine imaging capabilities at 4 frames per second for real-time tracking and beam control. Testing of the system was performed using an in-house developed motion platform and a commercially available motion phantom. Anatomical tracking is performed by first identifying a target (a region of interest that is either tissue to be treated or a critical structure) and generating a contour around it. A boundary contour is also created to identify tracking margins. The tracking algorithm deforms the anatomical contour (target or a normal organ) on every subsequent cine frame and compares it to the static boundary contour. If the anatomy of interest moves outside the boundary, the radiation delivery is halted until the tracked anatomy returns to treatment portal. The following were performed to validate and clinically implement the system: (a) spatial integrity evaluation; (b) tracking accuracy; (c) latency; (d) relative point dose and spatial dosimetry; (e) development of clinical workflow for gating; and (f) independent verification by an outside credentialing service.

RESULTS

The spatial integrity of the MR system was found to be within 2 mm over a 45-cm diameter field-of-view. The tracking accuracy for geometric targets was within 1.2 mm. The average system latency was measured to be within 394 ms. The dosimetric accuracy using ionization chambers was within 1.3% ± 1.7%, and the dosimetric spatial accuracy was within 2 mm. The phantom irradiation for the outside credentialing service had satisfactory results, as well.

CONCLUSIONS

The first clinical MR-IGRT system was validated for real-time tracking and gating capabilities and shown to be reliable and accurate. Patient workflow methods were developed for efficient treatment. Periodic quality assurance tests can be efficiently performed with commercially available equipment to ensure accurate system performance.

Influence of Head Motion on the Accuracy of 3D Reconstruction with Cone-Beam CT: Landmark Identification Errors in Maxillofacial Surface Model

PURPOSE

The purpose of this study was to investigate the influence of head motion on the accuracy of three-dimensional (3D) reconstruction with cone-beam computed tomography (CBCT) scan.

MATERIALS AND METHODS

Fifteen dry skulls were incorporated into a motion controller which simulated four types of head motion during CBCT scan: 2 horizontal rotations (to the right/to the left) and 2 vertical rotations (upward/downward). Each movement was triggered to occur at the start of the scan for 1 second by remote control. Four maxillofacial surface models with head motion and one control surface model without motion were obtained for each skull. Nine landmarks were identified on the five maxillofacial surface models for each skull, and landmark identification errors were compared between the control model and each of the models with head motion.

RESULTS

Rendered surface models with head motion were similar to the control model in appearance; however, the landmark identification errors showed larger values in models with head motion than in the control. In particular, the Porion in the horizontal rotation models presented statistically significant differences (P < .05). Statistically significant difference in the errors between the right and left side landmark was present in the left side rotation which was opposite direction to the scanner rotation (P < .05).

CONCLUSIONS

Patient movement during CBCT scan might cause landmark identification errors on the 3D surface model in relation to the direction of the scanner rotation. Clinicians should take this into consideration to prevent patient movement during CBCT scan, particularly horizontal movement.

Artifacts: The downturn of CBCT image

Cone-beam computed tomography (CBCT) has been accepted as a useful tool for diagnosis and treatment planning in dentistry. Despite a growing trend of CBCT in dentistry, it has some disadvantages like artifacts. Artifacts are discrepancies between the reconstructed visual image and the actual content of the subject which degrade the quality of CBCT images, making them diagnostically unusable. Additionally, structures that do not exist in the subject may appear within images. Such structures can occur because of patient motion, the image capture and reconstruction process. To optimize image quality, it is necessary to understand the types of artifacts. This article aims to throw light on the various types of artifacts associated with CBCT images.

An algorithm to extract three-dimensional motion by marker tracking in the kV projections from an on-board imager: four-dimensional cone-beam CT and tumor tracking implications

The purpose of this work is to extract three‐dimensional (3D) motion trajectories of internal implanted and external skin‐attached markers from kV cone‐beam projections and reduce image artifact from patient motion in cone‐beam computed tomography (CBCT) from on‐board imager. Cone beam radiographic projections were acquired for a mobile phantom and liver patients with internal implanted and external skin‐attached markers. An algorithm was developed to automatically find the positions of the markers in the projections. It uses normalized cross‐correlation between a template image of a metal seed marker and the projections to find the marker position. From these positions and time‐tagged angular views, the marker 3D motion trajectory was obtained over a time interval of nearly one minute, which is the time required for scanning. This marker trajectory was used to remap the pixels of the projections to eliminate motion. Then, the motion‐corrected projections were used to reconstruct CBCT. An algorithm was developed to extract 3D motion trajectories of internal and external markers from cone‐beam projections using a kV monoscopic on‐board imager. This algorithm was tested and validated using a mobile phantom and patients with liver masses that had radio‐markers implanted in the tumor and attached to the skin. The extracted motion trajectories were used to investigate motion correlation between internal and external markers in liver patients. Image artifacts from respiratory motion were reduced in CBCT reconstructed from cone‐beam projections that were preprocessed to remove motion shifts obtained from marker tracking. With this method, motion‐related image artifacts such as blurring and spatial distortion were reduced, and contrast and position resolutions were improved significantly in CBCT reconstructed from motion‐corrected projections. Furthermore, correlated internal and external marker 3D‐motion tracks obtained from the kV projections might be useful for 4DCBCT, beam gating and tumor motion monitoring or tracking.

PACS numbers: 87.57.Q, 87.57.C‐

A new fast and fully automated software based algorithm for extracting respiratory signal from raw PET data and its comparison to other methods

PURPOSE

Respiratory gating in PET is an approach used to minimize the negative effects of respiratory motion on spatial resolution. It is based on an initial determination of a patient's respiratory movements during a scan, typically using hardware based systems. In recent years, several fully automated databased algorithms have been presented for extracting a respiratory signal directly from PET data, providing a very practical strategy for implementing gating in the clinic. In this work, a new method is presented for extracting a respiratory signal from raw PET sinogram data and compared to previously presented automated techniques.

METHODS

The acquisition of respiratory signal from PET data in the newly proposed method is based on rebinning the sinogram data into smaller data structures and then analyzing the time activity behavior in the elements of these structures. From this analysis, a 1D respiratory trace is produced, analogous to a hardware derived respiratory trace. To assess the accuracy of this fully automated method, respiratory signal was extracted from a collection of 22 clinical FDG-PET scans using this method, and compared to signal derived from several other software based methods as well as a signal derived from a hardware system.

RESULTS

The method presented required approximately 9 min of processing time for each 10 min scan (using a single 2.67 GHz processor), which in theory can be accomplished while the scan is being acquired and therefore allowing a real-time respiratory signal acquisition. Using the mean correlation between the software based and hardware based respiratory traces, the optimal parameters were determined for the presented algorithm. The mean/median/range of correlations for the set of scans when using the optimal parameters was found to be 0.58/0.68/0.07-0.86. The speed of this method was within the range of real-time while the accuracy surpassed the most accurate of the previously presented algorithms.

CONCLUSIONS

PET data inherently contains information about patient motion; information that is not currently being utilized. We have shown that a respiratory signal can be extracted from raw PET data in potentially real-time and in a fully automated manner. This signal correlates well with hardware based signal for a large percentage of scans, and avoids the efforts and complications associated with hardware. The proposed method to extract a respiratory signal can be implemented on existing scanners and, if properly integrated, can be applied without changes to routine clinical procedures.

A phantom study on the effects of target motion in non-gated kV-CBCT imaging

Kilo-voltage cone beam computed tomography (kV-CBCT) integrated with a linac can produce online volumetric and anatomical images for patient set-up and dosimetric analysis in adaptive radiotherapy. However CBCT is prone to motion artifacts. This study investigates the impact of target motion in CBCT imaging. To simulate respiratory movement, a dynamic phantom was moved in three-dimensions with a period of 4 s and two different amplitudes (PA1 and PA2). The targets of well defined geometries were made using wax. A reference image of the static target was achieved with fan beam CT. Using CBCT, the targets in static and dynamic modes were imaged under full-fan beam conditions. The length of average HU spread was reduced in range from 19.35 to 44.44% along the cranio-caudal direction of targets. The percentage volume loss of dynamic targets imaged using CBCT (for Hounsfield Units with window width -500 to 0) ranged from 14.35 to 30.95% for PA1 and 21.29 to 43.80% for PA2 in comparison with static targets imaged with fan beam CT. A significant loss of volumetric information may result for non-gated CBCT imaging of moving targets and may result in a systematic error in re-contouring when CBCT images are used for radiotherapy re-planning.

Self-calibration of a cone-beam micro-CT system

Use of cone-beam computed tomography (CBCT) is becoming more frequent. For proper reconstruction, the geometry of the CBCT systems must be known. While the system can be designed to reduce errors in the geometry, calibration measurements must still be performed and corrections applied. Investigators have proposed techniques using calibration objects for system calibration. In this study, the authors present methods to calibrate a rotary-stage CB micro-CT (CBmicroCT) system using only the images acquired of the object to be reconstructed, i.e., without the use of calibration objects. Projection images are acquired using a CBmicrouCT system constructed in the authors' laboratories. Dark- and flat-field corrections are performed. Exposure variations are detected and quantifled using analysis of image regions with an unobstructed view of the x-ray source. Translations that occur during the acquisition in the horizontal direction are detected, quantified, and corrected based on sinogram analysis. The axis of rotation is determined using registration of antiposed projection images. These techniques were evaluated using data obtained with calibration objects and phantoms. The physical geometric axis of rotation is determined and aligned with the rotational axis (assumed to be the center of the detector plane) used in the reconstruction process. The parameters describing this axis agree to within 0.1 mm and 0.3 deg with those determined using other techniques. Blurring due to residual calibration errors has a point-spread function in the reconstructed planes with a full-width-at-half-maximum of less than 125 microm in a tangential direction and essentially zero in the radial direction for the rotating object. The authors have used this approach on over 100 acquisitions over the past 2 years and have regularly obtained high-quality reconstructions, i.e., without artifacts and no detectable blurring of the reconstructed objects. This self-calibrating approach not only obviates calibration runs, but it also provides quality control data for each data set.

Anniversary paper. Development of x-ray computed tomography: the role of medical physics and AAPM from the 1970s to present

The AAPM, through its members, meetings, and its flagship journal Medical Physics, has played an important role in the development and growth of x-ray tomography in the last 50 years. From a spate of early articles in the 1970s characterizing the first commercial computed tomography (CT) scanners through the "slice wars" of the 1990s and 2000s, the history of CT and related techniques such as tomosynthesis can readily be traced through the pages of Medical Physics and the annals of the AAPM and RSNA/AAPM Annual Meetings. In this article, the authors intend to give a brief review of the role of Medical Physics and the AAPM in CT and tomosynthesis imaging over the last few decades.

Technical note: A deformable phantom for dynamic modeling in radiation therapy

A deformable phantom was developed to aid in quality assurance for dynamic imaging and targeting techniques in radiation therapy. Made of simple materials combined with standard components for imaging and motion experiments, this phantom can be relatively easily constructed and used for both diagnostic imaging and dosimetry. Repeat imaging studies indicate that the phantom meets criteria of relative attenuation, deformation, and reproducibility of configuration, necessary for quality assurance of radiographic and tomographic targeting.

Image reconstruction for transmission tomography when projection data are incomplete

Two iterative methods are developed for forming a maximum-likelihood estimate of the attenuation density in a patient or object for transmission tomography when projection data are incomplete. The methods converge monotonically to the same limit points. Results of testing the methods with both simulated and real data are given.

Improvement in banding artefacts in four-dimensional computed tomography for radiotherapy planning

Respiratory-gated CT (RGCT) and four-dimensional CT (4DCT) scan techniques cover consecutive segments of the respiratory cycle. However, motion artefacts may occur in fast respiratory phases such as mid-inhalation and -exhalation. CT imaging involves the use of a number of x-ray tube positions for each couch position. We investigated the fundamental nature of motion artefacts using a constant-velocity moving phantom in motion in the CT plane or perpendicular to the CT plane, and in pigs to simulate a human model. Artefacts and movement distance were evaluated in a moving phantom and artificially ventilated pigs with a 256-multi-detector row CT (256MDCT). The phantom moved in the CT plane or perpendicular to the CT plane with a constant velocity. Backprojection used variable initial backprojection angles (IBAs). The phantom length for motion perpendicular to the CT plane was independent of IBA but was represented by phantom diameter plus the distance of movement per gantry rotation. In contrast, that for the motion in the CT plane was dependent on IBA, as represented by phantom diameter plus the distance of movement per rotation for IBA perpendicular to the phantom movement direction, and phantom diameter plus half the distance of movement per gantry rotation for other IBAs. Results for volumetric CT images with different IBAs showed the presence of banding artefacts. Similar findings were seen in artificially ventilated pigs. Motion artefacts are unavoidable in both conventional CT and 256MDCT. Banding artefacts will be improved if the same IBAs at each couch position are accounted for during image reconstruction. This improvement will be beneficial in respiratory gated and 4D radiation therapies.

Region-of-interest reconstruction of motion-contaminated data using a weighted backprojection filtration algorithm

The recently developed weighted backprojection filtration (WBPF) algorithm using data redundancy has capabilities that make this algorithm an attractive candidate for reconstructing images from motion-contaminated projection data. First, the WBPF algorithm is capable of reconstructing region-of-interest (ROI) images from reduced-scan fan-beam data, which have less data than the short-scan data required to reconstruct the entire field of view (FOV). Second, this algorithm can reconstruct ROI images from truncated data. Using phantom simulation studies, we demonstrate how these unique capabilities can be exploited to reduce the amount of motion-contaminated data used for reconstruction. In particular, we use examples from cardiac imaging to illustrate how off-center phantom positioning combined with phase-interval ROI reconstruction can result in the suppression of motion artifacts. In terms of temporal resolution, reduced-scan reconstruction with 45% of a full-scan dataset can be used to improve the temporal resolution of a short-scan reconstruction by 25.8% if ungated data are used. For data gated at 66 beats per minute, reduced-scan reconstruction with 45% of a full-scan dataset can be used to improve the temporal resolution of a short-scan reconstruction by 7.9%. As a result of our studies, we believe that the WBPF algorithm demonstrates the potential for reconstructing quality ROI images from motion-contaminated fan-beam data.