Using cone-beam CT projection images to estimate the average and complete trajectory of a fiducial marker moving with respiration

Motion estimation and correction in SPECT, PET and CT

Patient motion impacts single photon emission computed tomography (SPECT), positron emission tomography (PET) and X-ray computed tomography (CT) by giving rise to projection data inconsistencies that can manifest as reconstruction artifacts, thereby degrading image quality and compromising accurate image interpretation and quantification. Methods to estimate and correct for patient motion in SPECT, PET and CT have attracted considerable research effort over several decades. The aims of this effort have been two-fold: to estimate relevant motion fields characterizing the various forms of voluntary and involuntary motion; and to apply these motion fields within a modified reconstruction framework to obtain motion-corrected images. The aims of this review are to outline the motion problem in medical imaging and to critically review published methods for estimating and correcting for the relevant motion fields in clinical and preclinical SPECT, PET and CT. Despite many similarities in how motion is handled between these modalities, utility and applications vary based on differences in temporal and spatial resolution. Technical feasibility has been demonstrated in each modality for both rigid and non-rigid motion, but clinical feasibility remains an important target. There is considerable scope for further developments in motion estimation and correction, and particularly in data-driven methods that will aid clinical utility. State-of-the-art machine learning methods may have a unique role to play in this context.

Real-time intrafraction motion monitoring in external beam radiotherapy

Radiotherapy (RT) aims to deliver a spatially conformal dose of radiation to tumours while maximizing the dose sparing to healthy tissues. However, the internal patient anatomy is constantly moving due to respiratory, cardiac, gastrointestinal and urinary activity. The long term goal of the RT community to 'see what we treat, as we treat' and to act on this information instantaneously has resulted in rapid technological innovation. Specialized treatment machines, such as robotic or gimbal-steered linear accelerators (linac) with in-room imaging suites, have been developed specifically for real-time treatment adaptation. Additional equipment, such as stereoscopic kilovoltage (kV) imaging, ultrasound transducers and electromagnetic transponders, has been developed for intrafraction motion monitoring on conventional linacs. Magnetic resonance imaging (MRI) has been integrated with cobalt treatment units and more recently with linacs. In addition to hardware innovation, software development has played a substantial role in the development of motion monitoring methods based on respiratory motion surrogates and planar kV or Megavoltage (MV) imaging that is available on standard equipped linacs. In this paper, we review and compare the different intrafraction motion monitoring methods proposed in the literature and demonstrated in real-time on clinical data as well as their possible future developments. We then discuss general considerations on validation and quality assurance for clinical implementation. Besides photon RT, particle therapy is increasingly used to treat moving targets. However, transferring motion monitoring technologies from linacs to particle beam lines presents substantial challenges. Lessons learned from the implementation of real-time intrafraction monitoring for photon RT will be used as a basis to discuss the implementation of these methods for particle RT.

An interdimensional correlation framework for real-time estimation of six degree of freedom target motion using a single x-ray imager during radiotherapy

Increasing evidence suggests that intrafraction tumour motion monitoring needs to include both 3D translations and 3D rotations. Presently, methods to estimate the rotation motion require the 3D translation of the target to be known first. However, ideally, translation and rotation should be estimated concurrently. We present the first method to directly estimate six-degree-of-freedom (6DoF) motion from the target's projection on a single rotating x-ray imager in real-time. This novel method is based on the linear correlations between the superior-inferior translations and the motion in the other five degrees-of-freedom. The accuracy of the method was evaluated in silico with 81 liver tumour motion traces from 19 patients with three implanted markers. The ground-truth motion was estimated using the current gold standard method where each marker's 3D position was first estimated using a Gaussian probability method, and the 6DoF motion was then estimated from the 3D positions using an iterative method. The 3D position of each marker was projected onto a gantry-mounted imager with an imaging rate of 11 Hz. After an initial 110° gantry rotation (200 images), a correlation model between the superior-inferior translations and the five other DoFs was built using a least square method. The correlation model was then updated after each subsequent frame to estimate 6DoF motion in real-time. The proposed algorithm had an accuracy (±precision) of  -0.03  ±  0.32 mm, -0.01  ±  0.13 mm and 0.03  ±  0.52 mm for translations in the left-right (LR), superior-inferior (SI) and anterior-posterior (AP) directions respectively; and, 0.07  ±  1.18°, 0.07  ±  1.00° and 0.06  ±  1.32° for rotations around the LR, SI and AP axes respectively on the dataset. The first method to directly estimate real-time 6DoF target motion from segmented marker positions on a 2D imager was devised. The algorithm was evaluated using 81 motion traces from 19 liver patients and was found to have sub-mm and sub-degree accuracy.

Reconstruction of implanted marker trajectories from cone-beam CT projection images using interdimensional correlation modeling

PURPOSE

Cone-beam CT (CBCT) is a widely used imaging modality for image-guided radiotherapy. Most vendors provide CBCT systems that are mounted on a linac gantry. Thus, CBCT can be used to estimate the actual 3-dimensional (3D) position of moving respiratory targets in the thoracic/abdominal region using 2D projection images. The authors have developed a method for estimating the 3D trajectory of respiratory-induced target motion from CBCT projection images using interdimensional correlation modeling.

METHODS

Because the superior-inferior (SI) motion of a target can be easily analyzed on projection images of a gantry-mounted CBCT system, the authors investigated the interdimensional correlation of the SI motion with left-right and anterior-posterior (AP) movements while the gantry is rotating. A simple linear model and a state-augmented model were implemented and applied to the interdimensional correlation analysis, and their performance was compared. The parameters of the interdimensional correlation models were determined by least-square estimation of the 2D error between the actual and estimated projected target position. The method was validated using 160 3D tumor trajectories from 46 thoracic/abdominal cancer patients obtained during CyberKnife treatment. The authors' simulations assumed two application scenarios: (1) retrospective estimation for the purpose of moving tumor setup used just after volumetric matching with CBCT; and (2) on-the-fly estimation for the purpose of real-time target position estimation during gating or tracking delivery, either for full-rotation volumetric-modulated arc therapy (VMAT) in 60 s or a stationary six-field intensity-modulated radiation therapy (IMRT) with a beam delivery time of 20 s.

RESULTS

For the retrospective CBCT simulations, the mean 3D root-mean-square error (RMSE) for all 4893 trajectory segments was 0.41 mm (simple linear model) and 0.35 mm (state-augmented model). In the on-the-fly simulations, prior projections over more than 60° appear to be necessary for reliable estimations. The mean 3D RMSE during beam delivery after the simple linear model had established with a prior 90° projection data was 0.42 mm for VMAT and 0.45 mm for IMRT.

CONCLUSIONS

The proposed method does not require any internal/external correlation or statistical modeling to estimate the target trajectory and can be used for both retrospective image-guided radiotherapy with CBCT projection images and real-time target position monitoring for respiratory gating or tracking.

Comparison of 2D and 3D modeled tumor motion estimation/prediction for dynamic tumor tracking during arc radiotherapy

Many real-time imaging techniques have been developed to localize a target in 3D space or in a 2D beam's eye view (BEV) plane for intrafraction motion tracking in radiation therapy. With tracking system latency, the 3D-modeled method is expected to be more accurate even in terms of 2D BEV tracking error. No quantitative analysis, however, has been reported. In this study, we simulated co-planar arc deliveries using respiratory motion data acquired from 42 patients to quantitatively compare the accuracy between 2D BEV and 3D-modeled tracking in arc therapy and to determine whether 3D information is needed for motion tracking. We used our previously developed low kV dose adaptive MV-kV imaging and motion compensation framework as a representative of 3D-modeled methods. It optimizes the balance between additional kV imaging dose and 3D tracking accuracy and solves the MLC blockage issue. With simulated Gaussian marker detection errors (zero mean and 0.39 mm standard deviation) and ~155/310/460 ms tracking system latencies, the mean percentage of time that the target moved  >2 mm from the predicted 2D BEV position are 1.1%/4.0%/7.8% and 1.3%/5.8%/11.6% for the 3D-modeled and 2D-only tracking, respectively. The corresponding average BEV RMS errors are 0.67/0.90/1.13 mm and 0.79/1.10/1.37 mm. Compared to the 2D method, the 3D method reduced the average RMS unresolved motion along the beam direction from ~3 mm to ~1 mm, resulting in on average only  <1% dosimetric advantage in the depth direction. Only for a small fraction of the patients, when tracking latency is long, the 3D-modeled method showed significant improvement of BEV tracking accuracy, indicating potential dosimetric advantage. However, if the tracking latency is short (~150 ms or less), those improvements are limited. Therefore, 2D BEV tracking has sufficient targeting accuracy for most clinical cases. The 3D technique is, however, still important in solving the MLC blockage problem during 2D BEV tracking.

An image-based method to synchronize cone-beam CT and optical surface tracking

The integration of in-room X-ray imaging and optical surface tracking has gained increasing importance in the field of image guided radiotherapy (IGRT). An essential step for this integration consists of temporally synchronizing the acquisition of X-ray projections and surface data. We present an image-based method for the synchronization of cone-beam computed tomography (CBCT) and optical surface systems, which does not require the use of additional hardware. The method is based on optically tracking the motion of a component of the CBCT/gantry unit, which rotates during the acquisition of the CBCT scan. A calibration procedure was implemented to relate the position of the rotating component identified by the optical system with the time elapsed since the beginning of the CBCT scan, thus obtaining the temporal correspondence between the acquisition of X-ray projections and surface data. The accuracy of the proposed synchronization method was evaluated on a motorized moving phantom, performing eight simultaneous acquisitions with an Elekta Synergy CBCT machine and the AlignRT optical device. The median time difference between the sinusoidal peaks of phantom motion signals extracted from the synchronized CBCT and AlignRT systems ranged between -3.1 and 12.9 msec, with a maximum interquartile range of 14.4 msec. The method was also applied to clinical data acquired from seven lung cancer patients, demonstrating the potential of the proposed approach in estimating the individual and daily variations in respiratory parameters and motion correlation of internal and external structures. The presented synchronization method can be particularly useful for tumor tracking applications in extracranial radiation treatments, especially in the field of patient-specific breathing models, based on the correlation between internal tumor motion and external surface surrogates.

Automatic tracking of gold seed markers from CBCT image projections in lung and prostate radiotherapy

PURPOSE

To construct a method and software to track gold seed implants in prostate and lung patients undergoing radiotherapy using CBCT image projections.

METHODS

A mathematical model was developed in the MatLab (Mathworks, Natick, USA) environment which uses a combination of discreet cosine transforms and filtering to enhance several edge detection methods for identifying and tracking gold seed fiducial markers in images obtained from Varian (Varian Medical Systems, Palo Alto, USA) and Elekta (Kungstensgatan, Sweden) CBCT projections.

RESULTS

Organ motion was captured for 16 prostate patients and 1 lung patient.

CONCLUSION

Image enhancement and edge detection is capable of automatically tracking markers for up to 98% (Varian) and 79% (Elekta) of CBCT projections for prostate and lung markers however inclusion of excessive bony anatomy (LT and RT LAT) inhibit the ability of the model to accurate determine marker location.

A comparison of phase, amplitude, and velocity binning for cone-beam computed tomographic projection-based motion reconstruction

PURPOSE

We previously developed a motion estimation technique based on direct cone-beam projection analysis. It is able to reconstruct the complete motion trajectory of a radio-opaque marker, including cycle-to-cycle variability, using respiratory binning of the projection images. This paper investigates the use of phase, amplitude, and amplitude-velocity binning in the context of projection-based cone-beam motion estimation (CBME).

METHODS AND MATERIALS

We simulated cone-beam computed tomographic scans of 160 tumor trajectories estimated by a CyberKnife Synchrony System (Accuray, Sunnyvale, CA), and reconstructed the complete trajectory with CBME using phase, amplitude, and amplitude-velocity binning of the projection data. Various numbers of respiratory bins, from 1 (no binning) to 100, were used for phase and amplitude binning, while 1 to 100 amplitude bins with 4 velocity bins were used for amplitude-velocity binning. From this large pool of data, we correlated the reconstruction accuracy with bin type, total number of bins, number of breathing cycles per bin, and the position of the bin within the breathing cycle.

RESULTS

CBME predicted the true motion of the marker with a 3-dimensional (3D) mean root mean square (RMS) error of 0.24 mm for amplitude-velocity binning, 0.31 mm for amplitude binning, and 0.52 mm for phase binning. Reconstruction 3D RMS error increased to over 1 mm when less than 3 breathing cycles contributed to a bin. We found that reconstruction accuracy was optimized when about 20 bins were used. Accuracy also decreased in bins located around the inhale portion of the breath cycle, compared with the mid- and end-exhale positions.

CONCLUSIONS

This study provides a quantitative assessment of phase, amplitude, and amplitude-velocity binning for CBME. A joint binning approach should be used to give both the accuracy of amplitude binning, as well as the robustness of phase binning, in areas of limited motion sampling.

Building motion models of lung tumours from cone-beam CT for radiotherapy applications

A method is presented to build a surrogate-driven motion model of a lung tumour from a cone-beam CT scan, which does not require markers. By monitoring an external surrogate in real time, it is envisaged that the motion model be used to drive gated or tracked treatments. The motion model would be built immediately before each fraction of treatment and can account for inter-fraction variation. The method could also provide a better assessment of tumour shape and motion prior to delivery of each fraction of stereotactic ablative radiotherapy. The two-step method involves enhancing the tumour region in the projections, and then fitting the surrogate-driven motion model. On simulated data, the mean absolute error was reduced to 1 mm. For patient data, errors were determined by comparing estimated and clinically identified tumour positions in the projections, scaled to mm at the isocentre. Averaged over all used scans, the mean absolute error was under 2.5 mm in superior-inferior and transverse directions.

External respiratory motion: shape analysis and custom realistic respiratory trace generation

PURPOSE

The authors developed a realistic respiratory trace generating (RTG) tool for use with phantom and simulation studies.

METHODS

The authors analyzed the extent of abdominal wall motion from a real-time position management system database comprised of 125 lung, liver, and abdominal patients to determine the shape and extent of motion. Using Akaike's information criterion (AIC), the authors compared different model types to find the optimal realistic model of respiratory motion.

RESULTS

The authors compared a family of sigmoid curves and determined a four parameter sigmoid fit was optimal for over 98% patient inhale and exhale traces. This fit was also better than sin (2)(x) for 98% of patient exhale and 70% of patient inhale traces and better than sin (x) for 100% of both patient inhale and exhale traces. This analysis also shows that sin (2)(x) is better than sin (x) for over 95% of patient inhale and exhale traces. With results from shape and extent of motion analysis, we developed a realistic respiratory trace generating (RTG) software tool. The software can be run in two modes: population and user defined. In population mode, the RTG draws entirely from the population data including inter- and intra fraction amplitude and period variability and baseline drift. In user-defined mode, the user customizes the respiratory parameters by inputting the peak-to-peak amplitude, period, end exhale position, as well as controls variability in these parameters and baseline drift.

CONCLUSIONS

This work provides a method of generating custom respiratory data that can be used for initial implementation and testing of new technologies.

On-line use of three-dimensional marker trajectory estimation from cone-beam computed tomography projections for precise setup in radiotherapy for targets with respiratory motion

PURPOSE

To develop and evaluate accurate and objective on-line patient setup based on a novel semiautomatic technique in which three-dimensional marker trajectories were estimated from two-dimensional cone-beam computed tomography (CBCT) projections.

METHODS AND MATERIALS

Seven treatment courses of stereotactic body radiotherapy for liver tumors were delivered in 21 fractions in total to 6 patients by a linear accelerator. Each patient had two to three gold markers implanted close to the tumors. Before treatment, a CBCT scan with approximately 675 two-dimensional projections was acquired during a full gantry rotation. The marker positions were segmented in each projection. From this, the three-dimensional marker trajectories were estimated using a probability based method. The required couch shifts for patient setup were calculated from the mean marker positions along the trajectories. A motion phantom moving with known tumor trajectories was used to examine the accuracy of the method. Trajectory-based setup was retrospectively used off-line for the first five treatment courses (15 fractions) and on-line for the last two treatment courses (6 fractions). Automatic marker segmentation was compared with manual segmentation. The trajectory-based setup was compared with setup based on conventional CBCT guidance on the markers (first 15 fractions).

RESULTS

Phantom measurements showed that trajectory-based estimation of the mean marker position was accurate within 0.3 mm. The on-line trajectory-based patient setup was performed within approximately 5 minutes. The automatic marker segmentation agreed with manual segmentation within 0.36 ± 0.50 pixels (mean ± SD; pixel size, 0.26 mm in isocenter). The accuracy of conventional volumetric CBCT guidance was compromised by motion smearing (≤21 mm) that induced an absolute three-dimensional setup error of 1.6 ± 0.9 mm (maximum, 3.2) relative to trajectory-based setup.

CONCLUSIONS

The first on-line clinical use of trajectory estimation from CBCT projections for precise setup in stereotactic body radiotherapy was demonstrated. Uncertainty in the conventional CBCT-based setup procedure was eliminated with the new method.

A Bayesian approach to real-time 3D tumor localization via monoscopic x-ray imaging during treatment delivery

PURPOSE

Monoscopic x-ray imaging with on-board kV devices is an attractive approach for real-time image guidance in modern radiation therapy such as VMAT or IMRT, but it falls short in providing reliable information along the direction of imaging x-ray. By effectively taking consideration of projection data at prior times and/or angles through a Bayesian formalism, the authors develop an algorithm for real-time and full 3D tumor localization with a single x-ray imager during treatment delivery.

METHODS

First, a prior probability density function is constructed using the 2D tumor locations on the projection images acquired during patient setup. Whenever an x-ray image is acquired during the treatment delivery, the corresponding 2D tumor location on the imager is used to update the likelihood function. The unresolved third dimension is obtained by maximizing the posterior probability distribution. The algorithm can also be used in a retrospective fashion when all the projection images during the treatment delivery are used for 3D localization purposes. The algorithm does not involve complex optimization of any model parameter and therefore can be used in a "plug-and-play" fashion. The authors validated the algorithm using (1) simulated 3D linear and elliptic motion and (2) 3D tumor motion trajectories of a lung and a pancreas patient reproduced by a physical phantom. Continuous kV images were acquired over a full gantry rotation with the Varian TrueBeam on-board imaging system. Three scenarios were considered: fluoroscopic setup, cone beam CT setup, and retrospective analysis.

RESULTS

For the simulation study, the RMS 3D localization error is 1.2 and 2.4 mm for the linear and elliptic motions, respectively. For the phantom experiments, the 3D localization error is < 1 mm on average and < 1.5 mm at 95th percentile in the lung and pancreas cases for all three scenarios. The difference in 3D localization error for different scenarios is small and is not statistically significant.

CONCLUSIONS

The proposed algorithm eliminates the need for any population based model parameters in monoscopic image guided radiotherapy and allows accurate and real-time 3D tumor localization on current standard LINACs with a single x-ray imager.