J Maurer

Potential underestimation of the internal target volume (ITV) from free-breathing CBCT.

PURPOSE Localization prior to delivery of SBRT to free-breathing patients is performed by aligning the planning internal target volume (ITV) from 4DCT with an on-board free-breathing cone-beam CT (FB-CBCT) image. The FB-CBCT image is assumed to also generate an ITV that captures the full range of motion, due to the acquisition spanning multiple respiratory cycles. However, the ITV could potentially be underestimated when the ratio of time spent in inspiration versus time spent in expiration (I/E ratio) deviates from unity. Therefore, the aim of this study was to investigate the effect of variable I/E ratios on the FB ITV generated from a FB-CBCT scan. METHODS This study employed both phantom and patient imaging data. For the phantom study, five periodic respiratory cycles were simulated with different I/E ratios. Six patient respiratory cycles with variable I/E ratios were also selected. All profiles were then programmed into a motion phantom for imaging and modified to exhibit three peak-to-peak motion amplitudes (0.5, 1.0, and 2.0 cm). Each profile was imaged using two spherical targets with 1.0 and 3.0 cm diameters. 2D projections were acquired with full gantry rotation of a kiloVoltage (kV) imager mounted onto the gantry of a modem linear accelerator. CBCT images were reconstructed from 2D projections using a standard filtered back-projection reconstruction algorithm. Quantitative analyses for the phantom study included computing the change in contrast along the direction of target motion as well as determining the area (which is proportional to the target volume) inside of the contour extracted using a Canny edge detector. For the patient study, projection data that were previously acquired under an investigational 4D CBCT slow-gantry imaging protocol were used to generate both FB-CBCT and 4D CBCT images. Volumes were then manually contoured from both datasets (using the same window and level) for quantitative comparison. RESULTS The phantom study indicated a reduction in contrast at the inferior edge of the ITV (corresponding to inspiration) as the ratio decreased, for both simulated and patient respiratory cycles. For the simulated phantom respiratory cycles, the contrast reduction of the smallest I/E ratio was 27.6% for the largest target with the smallest amplitude and 89.7% for the smallest target with the largest amplitude. For patient respiratory cycles, these numbers were 22.3% and 94.0%, respectively. The extracted area from inside of the target contours showed a decreasing trend as the I/E ratio decreased. In the patient study, the FB-CBCT ITVs of both lung tumors studied were underestimated when compared with their corresponding 4D CBCT ITV. The underestimations found were 40.1% for the smaller tumor and 24.2% for the larger tumor. CONCLUSIONS The ITV may be underestimated in a FB-CBCT image when a patients respiratory pattern is characterized by a disparate length of time spent in inspiration versus expiration. Missing the full target motion information during on-board verification imaging may result in localization errors.

On-board four-dimensional digital tomosynthesis : First experimental results

The purpose of this study is to propose four-dimensional digital tomosynthesis (4D-DTS) for on-board analysis of motion information in three dimensions. Images of a dynamic motion phantom were reconstructed using acquisition scan angles ranging from 20 degrees (DTS) to full 360 degrees cone-beam computed tomography (CBCT). Projection images were acquired using an on-board imager mounted on a clinical linear accelerator. Three-dimensional (3D) images of the moving target were reconstructed for various scan angles. 3D respiratory correlated phase images were also reconstructed. For phase-based image reconstructions, the trajectory of a radiopaque marker was tracked in projection space and used to retrospectively assign respiratory phases to projections. The projections were then sorted according phase and used to reconstruct motion correlated images. By using two sets of projections centered about anterior-posterior and lateral axes, this study demonstrates how phase resolved coronal and sagittal DTS images can be used to obtain 3D motion information. Motion artifacts in 4D-DTS phase images are compared with those present in four-dimensional CT (4DCT) images. Due to the nature of data acquisition for the two modalities, superior-inferior motion artifacts are suppressed to a greater extent in 4D-DTS images compared with 4DCT. Theoretical derivations and experimental results are presented to demonstrate how optimal selection of image acquisition parameters including the frequency of projection acquisition and the phase window depend on the respiratory period. Two methods for acquiring projections are discussed. Preliminary results indicate that 4D-DTS can be used to acquire valuable kinetic information of internal anatomy just prior to radiation treatment.

Slow gantry rotation acquisition technique for on-board four-dimensional digital tomosynthesis

PURPOSE Four-dimensional cone-beam computed tomography (4D CBCT) has been investigated for motion imaging in the radiotherapy treatment room. The drawbacks of 4D CBCT are long scan times and high imaging doses. The aims of this study were to develop and investigate a slow gantry rotation acquisition protocol for four-dimensional digital tomosynthesis (4D DTS) as a faster, lower dose alternative to 4D CBCT. METHODS This technique was implemented using an On-Board Imager kV imaging system (Varian Medical Systems, Palo Alto, CA) mounted on the gantry of a linear accelerator. The general procedure for 4D DTS imaging using slow gantry rotation acquisition consists of the following steps: (1) acquire projections over a limited gantry rotation angle in a single motion with constant frame rate and gantry rotation speed; (2) generate a respiratory signal and temporally match projection images with appropriate points from the respiratory signal; (3) use the respiratory signal to assign phases to each of the projection images; (4) sort projection images into phase bins; and (5) reconstruct phase images. Phantom studies were conducted to validate theoretically derived relationships between acquisition and respiratory parameters. Optimization of acquisition parameters was then conducted by simulating lung scans using patient data. Lung tumors with approximate volumes ranging from 0.12 to 1.53 cm3 were studied. RESULTS A protocol for slow gantry rotation 4D DTS was presented. Equations were derived to express relationships between acquisition parameters (frame rate, phase window, and angular intervals between projections), respiratory cycle durations, and resulting acquisition times and numbers of projections. The phantom studies validated the relationships, and the patient studies resulted in determinations of appropriate acquisition parameters. The phase window must be set according to clinical goals. For 10% phase windows, we found that appropriate frame rates ranging from 2 to 5 frames/s, gantry rotation speeds ranging from 0.44 to 1.03 degrees/s, and aiming for an approximate maximum angular interval of 3.4 degrees between projections in phase bins were appropriate for dose, scan time, and tumor visibility optimization. Adequate tumor visibility was achieved for coronal 4D DTS images of all three lung tumors with acquisition times ranging from 0.45 to 2.12 min. vs. 1.84 to 4.24 min for 4D CBCT. 4D DTS imaging doses ranged from 0.12 to 0.72 times the dose of a standard CBCT scan vs 0.48 to 1.44 times the dose of a standard CBCT scan for 4D CBCT. CONCLUSIONS A slow gantry rotation acquisition technique for 4D DTS was developed and investigated. Study results indicated that 4D DTS is a feasible technique for imaging lung tumor motion in the treatment room and requires shorter acquisition times and less imaging dose than 4D CBCT for larger tumors that do not require large scan angles for sagittal views and for situations where only coronal views are needed to meet clinical needs.

MO‐D‐BRC‐07: Reducing Artifacts in Cone‐Beam CT Images Caused by the Presence of An Array Used for Tracking Transponders during Radiotherapy

Purpose: To compare two methods for reducing artifacts in cone‐beam CT(CBCT)images caused by the presence of an electromagnetic array used for tracking transponders during radiotherapy. Methods: A pelvis phantom was imaged using a gantry mounted CBCT system at two dose levels typical for prostate imaging. Both dose protocols acquired approximately 670 projections at 125 kVp. Tube currents and pulse widths were 80 mA/13 ms and 63 mA/10 ms for standard and low dose scans. The phantom was imaged with and without the array in position. The two methods used to reduce artifacts were: 1) filtering projection images prior to reconstruction, and 2) halving the dose per projection and doubling the number of projections. The filtering method replaced pixel values with neighboring values when pixels differed from neighboring pixels by a threshold percentage. 18, 22, 26 and 30% thresholds were investigated. Artifact reduction was quantified by computing difference images between the ‘corrected’ reconstructions and the image reconstructions for the phantom without the presence of the array. The filtering method was also applied to a prostate patient data set. Results: Average pixel values of the difference images were reduced by 6.4 and 7.3% for the 18% threshold filtered reconstructions for low and high doses, respectively, compared with difference images between the unaltered reconstructions with and without the panel. Reductions were 4.4 and 4.9% for other thresholds for low and high doseimages. 8.6 and 19.6% reductions were calculated for low and high doseimages using increased projection numbers. Improved soft tissuevisibility was noted in the filtered patient reconstructions. Conclusions: Both filtering and increasing projection numbers reduced artifacts caused by the presence of an electromagnetic array in CBCTimaging. Quantitative and visual analyses suggested that greater artifact reduction was achieved with increased projection numbers.

SU‐DD‐A3‐03: How Accurately Can the Internal Target Volume (ITV) from a Free‐Breathing Cone Beam Computed Tomography (FB‐CBCT) Scan Be Used for Target Verification?

Purpose: An internal target volume (ITV) is often identified by acquiring a free‐breathing cone‐beam computed tomography scan (FB‐CBCT) to verify patient setup of tumors affected by respiratory motion. However, its accuracy is not well understood. The purpose of this study is to characterize the potential underestimation of the ITV when a FB‐CBCT is acquired with irregular respiratory cycles. Methods & Materials: Five patient respiratory profiles were programmed into a 4D Dynamic Thorax (CIRS Model 008) phantom containing a spherical target, 2cm in diameter. The superior‐inferior motion was 3cm with respiratory cycles approximately 5 seconds in length. Five 360° CBCT scans of each profile were acquired using a gantry mounted kV imaging system. A FB‐CBCT, as well as a 4DCBCT of 10 phase bins with 10% phase windows were reconstructed for each profile. Inspiration (0%) and expiration (50%) phase images were compared to FB‐CBCT images. The ITV of each FB‐CBCT displayed a sharp and a smeared region. The percent reduction in contrast‐to‐noise ratio (CNR) between these regions was calculated, as well as the ratio of the average time spent in inspiration versus expiration. The relationship between contrast reduction and the ratio of time spent per phase was investigated. Results: For the five profiles studied, the CNR reductions ranged from 37% to 70%, corresponding to ratios of average time spent in inspiration versus expiration that ranged from 0.42 to 0.12, respectively. As the difference between time spent in expiration versus inspiration increased, the reduction in contrast also generally increased (r2=0.873). Conclusions: The observed loss of contrast in the ITV for irregular respiratory cycles may lead to the potential underestimation of this volume obtained from a FB‐CBCT. Extra caution is required when using an ITV for on‐board target localization when respiratory cycles exhibit large time differences between inspiration and expiration periods.

SU‐FF‐J‐55: Acceptance Testing and Quality‐Assurance Protocols for the Calypso® 4D Localization System™ and Q‐Fix®/Calypso® Couchtop

Purpose: To develop quality‐assurance (QA) protocols for the Calypso® 4D Localization System including acceptance testing, commissioning, daily QA, monthly QA, and annual QA. Method and Materials: The following acceptance testing/commissioning/QA protocols were developed following the implementation of the Calypso System at our institution. Both the Calypso System itself and a new Q‐fix/Calypso couch‐top were tested. The Calypso System tests consisted of (1) transponder imaging and treatment planning, (2) localization and tracking accuracy, (3) accuracy of motion measurements, (4) operational range of the array and system integration, (5) on‐board x‐ray imaging when using the array, (6) safety and collision checks, (7) automatic couch repositioning accuracy and functionality, (8) clinical application protocols and workflow. The Q‐fix/Calypso® couch‐top tests included mechanical tests (accuracy of motion along the three axes, angle indicator accuracy, isocentricity, and sag) as well as measurements of the attenuation through the new couch top and rails. Results:Localization/Tracking Accuracy: The Calypso‐measured displacements agreed with the known displacements within 1mm. Motion: The Calypso System measured the period and amplitude of the longitudinal sinusoidal motion within 0.1s and 0.1cm. Operational Range: We measured the operational range of the Calypso® array to be 14cm lateral, 15.8cm longitudinal, and 28.7cm vertical. Imaging: kV imaging clearly delineates the transponders. The presence of the array introduces significant noise into the CBCT scan, but still allows for localization of the transponders. Safety Checks: The appropriate warning messages were generated when the test patient plans were used. Automatic Couch Repositioning: The ACR correctly repositioned the phantom within 0.05cm of a known 3D shift. Q‐Fix/Calypso® Couch‐top: Recalibrating the vertical couch position was required (3mm change). The maximum couch attenuation was 3.6%, and rail attenuation was 9.9%. Conclusion: The Calypso® 4D Localization System was successfully implemented. Collection of baseline accuracy is important to ensure the continuing functionality of the system.

TH-A-211A-01: Digital Tomosynthesis for Target Localization

Any deviation between treated and planned volumes for 3‐D conformal therapy, such as IMRT, may cause an adverse clinical outcome. It is therefore critical to minimize all potential deviations using an on‐board (or real‐time) imaging procedure immediately prior to radiation delivery. At present, conventional 2‐D radiographicimaging and state‐of‐the‐art 3‐D cone‐beam CT(CBCT) are typically employed for treatment verification. However, 2‐D radiographic verification is mainly based on bony structures and/or implanted fiducials, and is sub‐optimal for soft‐tissue targets. While on‐board CBCT can provide 3‐D soft tissue information, it has three potential limitations: 1) The acquisition time is limited to a minimum of 60 seconds (≈15 breathing cycles) for on‐board CBCT, making single breath‐hold imaging impractical for organs which exhibit respiratory motion; 2) 360° mechanical clearance for CBCT acquisition may limit the use of CBCT for large patients, those with tumors at peripheral locations (e.g. breast), or those with substantial immobilization or support devices; 3) A high radiationdose (2–9 cGy) is delivered to the imaged volume with current imaging techniques, which is undesirable for daily imaging and may be a particular problem for those who are at high risk of developing second malignancies. To overcome these limitations, innovative digital tomosynthesis (DTS) imaging technologies are being developed for 3‐D and 4‐D target localization. Although DTS technology has been used for digital chest and mammography, its use in target localization is new. DTS only requires limited gantry rotation (e.g., a scan angle of 40° or less) to reconstruct 3‐D anatomic information. Thus, imaging time and dose are substantially reduced compared to CBCT, making breath‐hold DTS a simple solution for daily imaging of moving organs. 4‐D DTS can also be generated much faster than 4‐D CBCT, which is desirable for on‐board 4‐D tomographic imaging. Further, the reduced mechanical clearance needed for DTS makes it more widely applicable than CBCT. At present, limited information has been published regarding the fundamentals of DTS for target localization, and/or preliminary clinical DTS localization results. The purpose of this presentation is to provide updated information about these topics, including the following objectives: 1. To understand the technical challenges and clinical potentials of using DTS technologies for target localization in radiation therapy 2. To understand the latest developments in DTS reconstruction and registration methods 3. To understand the clinical feasibility and efficacy of kV DTS compared to kV CBCT 4. To learn about the latest developments in MV DTS and brachytherapy DTS applications Work is partially supported by grants from NIH, Varian Medical Systems, and GE Health Care.

TU‐EE‐A3‐03: On‐Board Four‐Dimensional Digital Tomosynthesis (4D‐DTS): Optimization of Respiratory Motion Dependent Acquisition and Reconstruction Parameters

Purpose: Amplitude and period of respiratory motion vary among patients. For four‐dimensional digital tomosynthesis (4D‐DTS), the frequency of projection acquisition and the phase window must be optimized based on respiratory motion. The purpose of this study was to demonstrate optimization of these parameters. Method and Materials: Experiments were performed to demonstrate optimization of projection acquisition frequency and phase window based on respiratory motion characteristics. Projection images of a CIRS Dynamic Thorax Phantom were acquired using an on‐board imager (OBI) mounted on a clinical accelerator. The trajectory of a radiopaque marker attached to the phantom was monitored in projection space and used to assign phases to and sort projections for 4D‐DTS reconstructions. 4D‐DTS images were reconstructed for motion profiles ranging in superior‐inferior amplitude from 10–40‐mm and ranging in period from 3.5–7‐sec. DTS images for each profile were reconstructed from various sets of projections, simulating different projection acquisition frequencies and phase windows. Results: For a desired phase window, the frequency of projection acquisition must be optimized based on the respiratory period. If projections are acquired at too high of a frequency, many of the projections will not be used in the reconstructions, resulting in unnecessary imaging dose. If projections are acquired at too low of a frequency, 4D‐DTS images can be reconstructed with missing projections or with larger phase windows. The effect of reconstructing with missing projections is minor if the fraction of missing projections is small. As the number of missing projections increases, vertical streaking artifacts appear in the images.Conclusion: This work is part of a feasibility analysis for 4D‐DTS imaging. It establishes relationships between optimal projection acquisition frequency, phase window and respiratory motion characteristics. Projection acquisition frequency based on respiratory period was derived and demonstrated. Conflict of Interest: Research sponsored by Varian.

TU‐FF‐A3‐05: Real‐Time Lung Tumor Motion Prediction Using Neural Network Based Models Constructed From Unsorted Cine CT Images

Purpose: To develop neural network based models for predicting respiratory induced tumor motion from gating signals and unsorted cine CTimages. The ability of the models to predict target motion given real‐time surrogate marker motion was verified using a dynamic phantom with a target moving according to motion profiles derived from patient respiratory data. Methods & Materials:Tumor motion profiles were derived using patient data and imported into CIRS Dynamic Thorax Phantom. The phantom was imaged with a 4‐slice CT scanner in cine mode. Surrogate marker motion from the phantom exterior was recorded simultaneously using Varians RPM system. Neural networks were constructed to model relationships between target presence or absence at each slice location and external marker motion. The combined networks for each of the slice locations were used to predict target motion. To verify the ability of the models to infer target position from surrogate signal, additional scans were performed. Models were used to predict target motion from external marker signals and compared to the motion patterns imported into the phantom. Results: Five preliminary experiments were performed with motion patterns from different patients. The models performed well for two cases, predicting motion with average errors on the order of image resolution, 1.26 mm and 1.39 mm root mean squared errors (RMSEs). For motion patterns with large peak‐to‐peak displacements or highly irregular frequencies / amplitudes, the predictions were less accurate. Conclusion: A neural network based method has been developed for modeling internal target motion using external surrogate signals and unsorted cine CTimages. Preliminary results indicate that the approach may be useful for some cases. Adaptive networks utilizing real‐time image information during the prediction process may be used to improve prediction accuracy for high frequency / amplitude motion patterns. Conflict of Interest: Partially supported by Varian research grant.