Volume 34, Issue 6, June 2007
Index of content:
- Joint Imaging/Therapy Scientific Session: Room M100F
- Localization I
MO‐E‐M100F‐02: Marker‐Less Intra‐Fraction Position Verification of Lung Tumors with An EPID in Cine Mode34(2007); http://dx.doi.org/10.1118/1.2761268View Description Hide Description
Purpose: Patient positioning represents one of the most challenging problems in radiation therapy, especially for some thoracic and abdominal target locations that are under the influence of respiratory motion. To monitor the target during treatment, we have developed an algorithm using a conventional EPID in cine mode and a prior 4DCT scan. A study based on patient data is presented to demonstrate the feasibility of the method. Materials and Methods: Based on the 4DCT recorded prior to treatment, a digitally reconstructed fluoroscopic (DRF) series is produced for each of the treatment fields. During the treatment, a cine EPID acquisition is performed for each field as it is delivered. Post‐treatment, we produce an image mask for the individual EPID and DRF images based on the MLC leaf positions. The masks are spatially registered and the required changes are induced so that the two images have consistent field shape and size. Following the image processing, the two image series are passed through a correlation algorithm that identifies the closest DRF image for each EPIDimage. This enables us to associate a 4DCT predetermined breathing phase to the EPID series of images, hence allowing us to recover the tumor position during the treatment relative to the planned position. Results: Depending on the number of acquired EPIDimages per treatment field (depending on the prescribed treatmentdose per field) we were able to recover between 70% (4 images per field) and 91% (6 images) of the 4DCT prerecorded tumor motion.
Conclusion: Based on a prior 4DCT and a cine EPIDimagesequence taken during the treatment, we have developed a computational algorithm that enabled us to quantify the differences between the tumor motion within the 4DCT that was used to plan the dose distribution on and the actual tumor motion during the treatment.
34(2007); http://dx.doi.org/10.1118/1.2761269View Description Hide Description
Purpose: To test the use of projection point tracking as an alternative to strain gauges for respiratory phase determination. Method and Materials: Projection point tracking uses 2D projection images acquired for the localization MV cone‐beam CT. A point, e.g. the top of the diaphragm under the ipsilateral lung, can be identified in each projection. In views where the top point of a surface is on a plateau, it is difficult to determine such points. By identifying the tracking point in two projections, near full exhale, separated by about 90 degrees, the location of the point in the patient can be calculated and subsequently projected onto every view, providing guide points. The tracking points on full exhale views are the guide points. Since tracking points have minimal motion between adjacent views, the tracking point on a view can be inferred from the one on the previous view and the view's guide point. A plot of the tracking point's coordinates versus view angle provides the projection track. For constant gantry rotation speeds, the projection track provides a plot of the respiratory phase. Results: An examination of projection tracks for a trained patient revealed 1) good agreement with strain gauge measurements, 2) inter‐fraction variation of tumor range of motion, and 3) inter‐fraction reproducibility of phase cycle. Two of the fractions were 180 degrees out of phase, raising the possibility of combining “near‐full‐exhale” projections to reconstruct a “full‐exhale” MV CT. With enough projection sets, it may be possible to reconstruct a 4D MV CT. The projection tracks could be used to select the appropriate views for each phase. Conclusion: Compared to strain gauges that only provide phase, projection point tracking is a more powerful method of characterizing respiration, since the amplitude of motion is also determined.
34(2007); http://dx.doi.org/10.1118/1.2761270View Description Hide Description
Purpose: This study explores the feasibility of using megavoltage fluoroscopy (MVF) for on‐line real‐time verification and guidance for gated/4D treatment delivery. Initial experiences in implementing MVF and methods developed to minimize required dose and optimize the signal‐to‐noise ratio will be presented. Methods and Materials: A Siemens LINAC retrofit with a flat panel imager in the beam direction was used to acquire MVF images with 10242 pixel matrix at 7 frames‐per‐second (fps) using 6 MV photons.Image processing software tools were developed to remove artifacts caused by beam pulsing, dead pixels and stuck bits and to average pixels and video frames. The beam delivery rate was adjusted during MVF acquisition in an attempt to minimize the dose per frame. A phantom designed to test the contrast resolution of kilovoltage fluoroscopy systems was used to assess the MVF image quality at different beam delivery rates. Results: Good quality MVF images with sequences longer than a typical respiratory cycle (5 sec) were obtained with an estimated dose as low as 0.6 cGy. Matching the beam pulse frequency to a multiple of the video frame rate minimized striping artifacts. Image sequences acquired at a frame rate of 3.5 fps and beam delivery rate of 12 MU/min had an SNR of 125; the minimum object resolvable during cine display was 0.063 inch in diameter, and the 0.31 inch diameter object was detectable at 2% contrast.Contrast detectability was improved by averaging video frames, reducing spatial resolution and displaying video in high‐speed cine mode. Conclusions: It is feasible to acquire good quality megavoltage fluoroscopyimages using a low dose‐rate treatment beam, indicating that MVF can be used to verify or guide gated/4D treatment delivery in real time (e.g. prior or during treatment delivery). Conflict of Interest: This work was supported in part by Siemens OCS.
MO‐E‐M100F‐05: Fluoroscopic Tracking of Multiple Implanted Fiducial Markers Using Multiple Object Tracking Algorithm34(2007); http://dx.doi.org/10.1118/1.2761271View Description Hide Description
Purpose: To develop a multiple fiducial marker tracking system providing precise tumor localization for beam gating and tracking. Method and Materials: We developed a multiple marker tracking system allowing us to track implanted fiducial markers for the treatment of mobile tumors. For each patient, LAT and AP fluoroscopic videos were acquired during the treatment. For each fluoroscopic frame, we first applied a prediction module to predict the marker positions and then a detection module to detect markers in the area around the predicted positions. Then, we applied a multiple object tracking algorithm to identify the true markers among all the detected ones. Breathing pattern information was used to improve the tracking performance. Four criteria were used to identify tracking failures, and when tracking failure happened, the system could immediately inform the user and disable the treatment beam. Results: We compared the performance of the proposed system to a conventional tracking system. To test the robustness of the tracking system, artificial markers were added around the true markers to generate false matches. The conventional system easily misses tracking markers in the presence of artificial markers; furthermore, it cannot detect subsequent tracking failures. Our proposed system could track markers well in the presence of artificial markers, and it could also successfully detect tracking failures. Our proposed system achieved a 0% failure rate almost all the time, and it yielded a low e 95 (the maximum marker tracking error at 95% confidence level) — less than 1.5 mm. Conclusion: The proposed marker tracking system can track multiple markers simultaneously without confusion and it is robust enough to continue tracking even when the marker is moving behind bony anatomy. It can also detect tracking failures and inform the user. Conflict of Interest: Research is partially sponsored by an NCI grant (1 R21 CA110177 A 01A1).
MO‐E‐M100F‐06: Intra‐Fraction and Inter‐Fraction Real‐Time Tumor Motion Prediction Using Multiple External Surrogates34(2007); http://dx.doi.org/10.1118/1.2761272View Description Hide Description
Purpose: To determine the feasibility of predicting real‐time tumor positions using multiple external surrogates and Partial‐Least‐Squares (PLS) regression. Methods: Nine patients with lungtumors underwent extracranial stereotactic radiosurgery comprising 3 fractions with the Cyberknife Synchrony™ system. Three LED markers were placed on the patient's abdomen and in close contact with the skin. Synchronous LED marker and tumor marker coordinates were measured periodically using an LEDcamera and stereoscopic x‐ray images, respectively. PLS performs regression on a small number of latent variables (LVs), which represent the underlying factors responsible for the variation in the output response. They are linear combinations of the original independent variables. The number of LVs in the PLS model was determined by performing a cross‐validation using a training data set and minimizing the error. The LV coefficients were then extracted from the cross‐correlation of input and output variables. Tumor position prediction errors were evaluated for the following scenarios: (1) tumor/external surrogate relationship for a given fraction, (2) predictive model developed in one fraction and applied to other fractions. Prediction errors were compared with standard multiple linear regression (MLR). Results: The mean tumor motion displacement was 7.3 mm for the patient data investigated. The average error for intra‐fraction tumor position prediction based on external markers was 1.47±0.90 mm for PLS and 3.07±1.67 mm for MLR when data from a single fraction was used. When the predictive model from one fraction was applied to the remaining two fractions, the average real‐time inter‐fraction error was 1.48±0.57 mm for PLS and 6.69±3.58 mm for MLR. Conclusion: PLS can predict the tumor position using external surrogates with errors of < 2mm and with superior accuracy compared with MLR. Intra‐ and inter‐fraction prediction errors were statistically similar.
MO‐E‐M100F‐07: Measurement of Delays Between Gating Signals and Beam‐On for Imaging, Static and Dynamic Treatments34(2007); http://dx.doi.org/10.1118/1.2761273View Description Hide Description
Purpose: To determine the delays between gating signals and delivery of kV imaging and MV treatment beams for static, enhanced dynamic wedge, and IMRT treatments. Method and Materials: All measurements were taken on a linear accelerator (Trilogy, Varian Medical Systems, Palo Alto, CA), equipped with kV imaging and respiratory gating. A phantom simulating regular motion was used to trigger the gating system. Delivery of both kV imaging and MV treatment beams were monitored using a diode connected to an analog electrometer operating in current mode. The electrometer's voltage output and the gating signal were both monitored by a multichannel oscilloscope. Traces were recorded comparing the diode output to the gate signal. The MV beam was monitored during delivery of static, enhanced dynamic wedge and IMRT fields. For the IMRT field the diode received only scattered radiation resulting in poor signal‐to‐noise but the results were still sufficient for this work. Results: The delay between the gating signal and beam on signal was found to be unmeasurable for the static MV beam, and 30 and 50 msec respectively for the wedged and IMRT beams. The delay for the kV beam was 300 msec. Because tumors can move up to 2 cm in a 5 sec breathing cycle, a 300 msec delay can result in up to a 4 mm position error; however, because gating is generally done at expiration or inspiration where the velocity is at a minimum, the uncertainty should be much lower. Conclusion: There is no appreciable delay between gating signals and MV irradiation. There is a delay between kV imaging and gating signals but the clinical consequences can be minimized by gating where tumor velocity is minimal.
- Localization II
WE‐E‐M100F‐01: 3D Interfraction Position Verification for Patients Undergoing Partial Breast Irradiation: Comparing Digital Tomosynthesis to Cone‐Beam CT34(2007); http://dx.doi.org/10.1118/1.2761583View Description Hide Description
Purpose: This clinical study evaluates digital tomosynthesis (DTS) technology for daily imaging guidance for partial breast irradiation (PBI) and compares its positioning accuracy to the 3D CBCT technique. Compared to CBCT, DTS offers lower imaging dose and the geometrical flexibility that particularly suits PBI. Methods and Material: Ten patients undergoing PBI were imaged using an on‐board‐imager mounted on a Varian 21EX linear accelerator. Following the initials setup using skin markers and 2D KV/MV radiographs, a CBCT scan was acquired to provide 3D positioning guidance. A subset of the CBCT projections were used to reconstruct a stack of DTS image slices using the Feldkamp filtered back‐projection algorithm. To optimize soft tissuecontrast, the DTS images were reconstructed for the 45‐degree oblique view along which the tumor bed, breast tissues, bones, and lung were well separated. Coronal and sagittal DTS views were also reconstructed for comparison. The inter‐fraction position deviations between the 1st fraction and each of the subsequent fractions were measured by coronal‐DTS, sagittal‐DTS, oblique‐DTS and CBCT as four independent technologies. Differences between these technologies and their clinical impact were evaluated. The evaluator was well trained for DTS technology. Results: Eighty‐five imaging datasets were obtained from 10 patients. Surgical clips (when present) were visible in all three DTS views. The tumor bed had the best contrast in the oblique‐DTS. One‐dimensional positioning differences between DTS (averaged over 3 DTS views) and CBCT were ∼0.1 cm when surgical clips were used in the registration and ∼0.2 cm when the tumor bed were used in the registration. Conclusion: DTS is equivalent to CBDT as a 3D imaging technique for daily patient positioning of PBI but with less mechanical constraints and imaging dose.
Partially supported by a research grant from Varian Medical Systems.
WE‐E‐M100F‐02: Interplay Between Image Quality and Temporal Resolution in 4DCT Acquisition Protocols34(2007); http://dx.doi.org/10.1118/1.2761584View Description Hide Description
Purpose: The accuracy and temporal resolution of 4DCT imaging depends on temporal characteristics of the acquisition protocol like gantry rotation speed and image reconstruction interval, in addition to the spatial and temporal characteristics of the motion itself. Therefore, a generic acquisition protocol might not be suitable for all patients. The aim of this study is to evaluate the interplay between the parameters affecting the accuracy and temporal resolution of 4DCT images.Methods and Materials: Several 4DCT images were acquired of cylindrical phantoms under repetitive motion induced by a translation platform. Acquisition settings were varied with respect to image reconstruction interval, gantry rotation speed and motion period of the phantoms. Reconstructed images were sorted into ten phase bins and compared to CTimages of static phantoms at corresponding positions of the respiration phase. Results: Acquisitions with different image reconstruction intervals did not play a significant role in the amount of motion observed in full cycle maximum intensity projection (MIP) images. Single‐phase image integrity was observed to be constant up to a threshold in the value of reconstruction interval, beyond which the image integrity varied somewhat arbitrarily due to the reduced number of images. This threshold is correlated with the number of phase bins and the motion period. Furthermore, image integrity was observed to improve with decreasing gantry rotation periods. Conclusions:Image integrity as evaluated in this study demonstrates relative gains with respect to changes in the temporal resolution parameters of 4DCT acquisition. Due to observed differences, we report that suboptimal settings could result in dramatic target delineation inaccuracies. Optimization of acquisition parameters needs to be performed, assessing the period of motion and limiting factors such as the availability of acquisition settings, image storage and processing power. Conflict of Interest: Sponsored in part by GE Healthcare.
WE‐E‐M100F‐03: Improving Soft Tissue Contrast in Megavoltage Cone‐Beam CT Images for Adaptive Radiotherapy34(2007); http://dx.doi.org/10.1118/1.2761585View Description Hide Description
Purpose: To investigate image quality improvement in Megavoltage Cone‐Beam CT (MV‐CBCT) images using image filtering techniques for adaptive radiotherapy (ART) protocols. MV‐CBCT imaging is often used for daily patient localization. However, soft tissuecontrast in MV‐CBCT images is limited and accurate delineation of targets and organs‐at‐risk is sometimes challenging. Image post‐processing with advanced image filtering techniques can improve the quality without the need to increase the dose exposure for the imaging procedure. Method and Materials: MV‐CBCT images of two image‐quality phantoms and of patients with prostate cancer were post‐processed using noise‐reducing, edge‐preserving image filters. On the phantom images, the contrast‐to‐noise ratio and the spatial resolution before and after filtering were evaluated. The improvement in image quality for the prostate patients was qualitatively judged by physicians based on the ability to delineate the prostate, rectum, bladder and seminal‐vesicle volumes. The optimal combination of MV‐CBCT delivery protocols with different patient doses and filtering techniques was determined for online and offline ART protocols. Results: Using an edge‐preserving noise‐reducing curvature flow image filter, the quality of MV‐CBCT images was improved. The contrast‐to‐noise ratio on the phantoms was improved by up to 30%, while maintaining the spatial resolution. Although the raw cone‐beam image quality of the prostate patients was sufficient for patient treatment localization, the ability to contour anatomical structures was increased on the post‐processed images.Conclusion: Using advanced filtering tools for MV‐CBCT images can improve the image quality, especially soft‐tissue contrast. This allows delineation of organs not clearly visible on the raw images. The tools are potentially beneficial for deformable registration of MV‐CBCT images with planning CTs, for monitoring the delivered dose to targets and organs‐at‐risk, and for ART protocols. Finally using protocols with lower exposure, patient dose from daily imaging procedures can be reduced.
WE‐E‐M100F‐04: Commissioning and Validation of a Beam Model for Calculating Megavoltage CT Dose From Imaging with a Helical Tomotherapy Unit34(2007); http://dx.doi.org/10.1118/1.2761586View Description Hide Description
Purpose: To commission and validate an MVCT beam model that allows for the calculation of dose received by patients due to megavoltage imaging on a helical tomotherapy unit (Tomotherapy, Inc., Madison, WI). Method and Materials: Percent depth dose and profile data were collected in order to commission a new MVCT beam model. The fluence output for the beam model was adjusted to match the measured dose in phantom. The model was then verified through a series of absorbed dose measurements in three phantoms (20‐cm cylindrical phantom, CIRS anthropomorphic phantom, and 30‐cm “cheese” phantom). The multiple scan average dose was recorded for all three phantoms with various changes to CTcollimator pitch and ion chamber location (central versus peripheral points). Results: The delivered doses and the computed doses were on average within 1.5% for all three phantoms, when the ion chamber was centrally located; and within 3.5%, when the chamber was located on the peripheral edge of the phantoms. The measured dose in the anthropomorphic phantom was 2.3 cGy with a pitch of 1.0 (4 mm couch movement per gantry rotation), 1.4 cGy with a pitch 2.0, and 0.90 cGy with a pitch of 3.0, these matched within 1% to the calculated dose. The computed versus measured dose was also within 1% when calculating dose in different tissue densities (lung and bone). Conclusion: This study has shown that with the development of a new MVCT beam model, dose delivered from MVCT imaging can be calculated. Validation measurements, in phantom, have verified that the computed dose can be reported to within 1.5% of the measured dose. The rationale for implementing this MVCT beam model is to provide a future method for calculating patient‐specific MVCT dose. Conflict of Interest: Co‐authors are either funded by a research grant or employed by TomoTherapy, Inc.
WE‐E‐M100F‐05: Evaluation of Respiration‐Correlated Digital Tomosynthesis in the Thorax and Abdomen for Soft Tissue Visualization and Patient Positioning34(2007); http://dx.doi.org/10.1118/1.2761587View Description Hide Description
Purpose: To find optimal parameters for digital tomosynthesis (DTS) image reconstruction, to evaluate ability of respiration correlated DTS to reduce blur caused by respiratory motion, and to assess DTS imaging for soft tissue localization and patient positioning. Methods and Materials:Image acquisition for DTS used a gantry‐mounted kV on‐board imaging system (Varian Medical Systems). We did not acquire DTS separately, but instead simulated DTS acquisition by using projection images acquired for CBCT. DTS reconstruction consisted of backprojection followed by a deblurring operation removing out‐of‐plane objects. For tumors subject to respiratory motion we selected projection images according to an external respiratory monitor signal (Real‐time Position Management System, Varian Medical Systems). Reconstruction and registration of DTS images used vender research software. Results: Optimal DTS quality is achieved with a 6–9 cm long deblurring volume in the direction perpendicular to the image and 20–30° reconstruction arc lengths. Image blur increases with longer arc lengths, while for shorter arcs out of plane objects become more pronounced. RC DTS reconstruction from disjoint arcs containing 2–3 respiratory cycles is feasible, yields images with less motion blur, and allows visualization of tumor movement. Generally, DTS was capable to visualize lungtumors, bronchi, liver, kidneys and abdominal lymph nodes. Estimated 3D registration error of DTS to reference DTS generated from the plan CT DRRs was 3 mm, relative to cone‐beam CT as a standard. Registration of RC DTS of lungtumors was possible to within 2 mm. Conclusions: DTS is capable of soft tissue visualization and patient positioning, including tumors subject to respiratory motion. With several advantages over full rotation CBCT scan, such as much shorter acquisition time, smaller dose and relaxed clearance requirement, DTS can become an efficient imaging modality for image‐guided radiotherapy. Conflict of Interest: Research sponsored by NCI Grant P01‐CA59017 and Varian Medical Systems.
WE‐E‐M100F‐06: Latency Measurements and Demonstration of a 4D Electromagnetic Localization System for LINAC Beam Gating34(2007); http://dx.doi.org/10.1118/1.2761588View Description Hide Description
Purpose: The Calypso® 4D Localization System uses electromagnetic fields to localize and track Beacon® transponder implants. In this study, we used the Calypso® System to perform real‐time beam gating based on transponders mounted to a dynamic phantom. The system's latency was studied in conjunction with a Varian linac to determine feasibility of enabling gated radiation therapydelivery for respiratory applications. Method and Materials: An acrylic transponder holder with three embedded Beacon® transponders was placed 6 cm from the center of a rotating disk on a dynamic phantom. Film measurements were made with a 2 cm by 2 cm field delivered in a static, dynamic‐gated, and dynamic‐non‐gated fashion. A linear gating window with a 2 mm width was used for the film demonstration. Latency measurements compared an in‐volume / out‐of‐volume signal obtained directly from the dynamic phantom to the target‐current signal from the linac. The target signal represents the measured current in the MV electron beam, and is directly correlated to the presence (or absence) of the treatment beam. The investigational prototype gating system does not used predictive modeling; the linac beam is enabled when the transponder is within the gating window based solely on the last position estimate of the transponder. Results: Film measurements demonstrated that signals from a prototype electromagnetic gating system can be used to effectively trigger the beam on/off state of a linac. The combined, motion‐triggered latency of the localization system and linac was 65 msec to disable the beam and 75 msec to enable the beam. Conclusion: Use of wireless electromagnetic implanted transponders has the potential to enable real‐time linac beam gating with the latency and update rates required for respiratory applications without the use of predictive algorithms.
WE‐E‐M100F‐07: Prostate Intrafraction Motion Measurement Using KV Fluoroscopy During Treatment Delivery34(2007); http://dx.doi.org/10.1118/1.2761589View Description Hide Description
Purpose: Margin reduction for prostate radiotherapy is limited by uncertainty in prostate localization during treatment. We investigate the feasibility and accuracy of measuring prostate intrafraction motion using kV fluoroscopy performed simultaneously with radiotherapy.Method and Materials: Three gold coils used for target localization are implanted into the patient's prostate gland before undergoing a hypofractionated online image‐guided step‐and‐shoot IMRT on an Elekta Synergy. At each fraction the patient is aligned using a CBCT, treatment delivery and fluoroscopy are performed simultaneously, and a post‐treatment CBCT is acquired with the patient still on the table. To measure the intrafraction motion we developed an algorithm to register the fluoroscopyimages to the projection images of the post‐treatment CBCT, and we combined information from fluoroscopyimages at different gantry angles to obtain the motion of the coils in 3D. The accuracy and robustness of this technique were evaluated by comparing measured results with those from an independent clinical 3D CBCT registration. Results: The coils can be detected and successfully registered using fluoroscopyimages for all gantry angles, and at some angles they can be tracked with frequency > 1 Hz. Fluoroscopyimages containing MV scatter can be detected and if necessary removed to improve image quality. The mean of the difference between our intrafraction measurement technique and the independent 3D CBCT registration for 40 measurements was 0.04 ± 0.47 mm, −0.07 ± 0.76 mm, and −0.21 ± 0.80 mm in the RL, AP, and SI axes respectively. Conclusion: These results show that measuring prostate intrafraction motion using kV fluoroscopy is feasible and can be performed with sub‐millimeter accuracy and adequate temporal resolution. Future work includes improving 2D registration to account for prostate rotation and deformation, and reporting clinical results. Conflict of Interest: Partially supported by NIH Grant CA118037.
- Margin Assesment
TH‐E‐M100F‐01: Impact of Organ Motion On IMRT Dose Distributions for Patients with Cancer of the Cervix34(2007); http://dx.doi.org/10.1118/1.2761747View Description Hide Description
Purpose: To investigate IMRT for treatment of cervical cancer and the need for adaptive treatment strategies. To use deformable dose registration to assess the effect of organ motion on dose.Method and Materials: Eight women with IB‐IVA cervix cancer were retrospectively selected. All patients had initial CT and MRI prior to treatment, and weekly MRI during external beam radiotherapy. Gross tumor volume, bladder, rectum, sigmoid, cervix, upper vagina, uterus and bilateral parametria regions were contoured on all MR. Clinical target volumes for tumor (HRCTV), pelvic nodes (nodeCTV) and vessels were contoured on the baseline MR. For each patient two IMRTtreatments were planned; Large Margin (LM) using 20mm margin for HRCTV (except inferiorly) and Small Margin (SM) using 5mm. For nodeCTV, 5mm margin was used. Surface‐mesh representations of the ROIs were propagated and conformed to corresponding contours from the weekly datasets. Meshes were exported to a research software, where the effect of organ motion was assessed by deformable dose registration. Results: Simulated treatment delivery, with deformable dose accumulation, showed that five out of eight patients would have achieved clinically approved target coverage from both LM and SM. For three patients, target coverage was unacceptable with SM. For one of those three, LM managed to fulfill target coverage. For all patients, SM showed better OAR protection than LM. Conclusion: The effect of organ motion on delivered dose has been studied by numerical simulation, showing that a large margin of 20mm cannot guarantee target coverage. For a majority of the patients, the smaller margin was enough to achieve full target coverage, this with enhanced OAR protection. This suggests large gain from more advanced patient specific margin recipes and adaptive treatment delivery. Conflict of Interest: This research was sponsored by RaySearch Laboratories, where some of the authors are employees and stock owners.
TH‐E‐M100F‐02: CTV‐To‐PTV Margins for Prostate Irradiations, a Quantitative Assessment Using Novel Voxel Tracking Techniques34(2007); http://dx.doi.org/10.1118/1.2761748View Description Hide Description
Purpose: To quantify adequate anisotropic CTV‐to‐PTV margins for three different setup strategies used during prostate irradiations: (1) no setup corrections, (2) on‐line corrections based on bony anatomy and (3) on‐line corrections based on gold markers. Method and Materials: Three radiation oncologists independently delineated the CTV, the bladder and the rectum in CTimages of 30 prostate patients with implanted gold markers using a 3D model‐based segmentation technique. IMRT plans with zero CTV‐to‐PTV margins were generated for each of the 3 times 30 contoured image datasets. Eight repeat scans were acquired to allow simulation of the delivered dose distributions in changing geometry. Different registration approaches were taken to mimic the different setup strategies. A surface‐model based deformable image registration system was used to warp the delivered dose distributions back to the dose in the planning CT. Based on the geometric extent of underdosed areas in this patient population, a set of anisotropic margins was derived. A second simulation was carried out to assess the target coverage of the derived margins. Results: Without setup correction, margins of 10 mm for the corpus of the prostate and 15 mm for the seminal vesicles were required. These margins could be reduced to 7 mm and 12 mm respectively with on‐line correction using bony anatomy registration methods and to 3 mm and 7 mm using on‐line repositioning based on gold markers. A larger margin at the apex was required to account for the significant observer variability and steep dose gradients at this location (11 mm — skin marker registration, 9 mm — bony anatomy registration, 7 mm — gold marker registration). Conclusion: Novel voxel tracking techniques enable us to calculate accumulated dose distributions and design accurate 3‐D CTV‐to‐PTV margins for prostate irradiations.
34(2007); http://dx.doi.org/10.1118/1.2761749View Description Hide Description
Purpose: When commissioning any component in an integrated radiotherapy system, it is imperative to consider possible compounding of errors through the entire system. Gated imaging and radiotherapy provides a useful example of this. Method and Materials: In gated radiotherapy, the accuracy of treatment delivery is determined by the accuracy with which both the imaging and treatment beams are gated. If the time delays (the time between the IR markers entering/leaving the gated region and the first/last image acquired or treatment beam on/off) for the imaging and treatment systems are not consistent, the required ITV margin may increase above that deduced from the tolerance for either system measured individually. We measured one gating system's time delay on 2 fluoroscopy systems, and two linear accelerator, using a motion phantom of known geometry, varying gating type (amplitude vs. phase), beam energy, dose rate, and period. Results: In the worst case scenario, beam‐off for amplitude‐based gating (3–5s period), the last fluoroscopic image in the gated region was acquired 0.15 ± 0.08 s (1SD) before the IR markers left the amplitude‐gated window, while the treatment beam cut off 0.06 ± 0.02 s after the IR markers had left the same region. For a patient with 1 cm amplitude, 4 s period sine wave breathing, this time delay mismatch increases the ITV margin by 3.7mm. For beam‐on times, the fluoroscopy system was also early, while the linear accelerator beam‐on was late. The images indicate a larger region is treated than is in truth, decreasing the duty cycle and increasing overall treatment time. Dose rate and treatment beam energy had negligible effects. In less‐predictable, physiological breathing motion, these time delays may vary. Conclusion: By following patient flow from simulation through treatment, physicists may more accurately assess accumulated errors in systems including gated radiotherapy.
TH‐E‐M100F‐04: Dosimetric Evaluation of MIP‐Based Patient Aperture and Compensator Designs to Treat Lung Cancer Using Proton Therapy Under Free Breathing Conditions34(2007); http://dx.doi.org/10.1118/1.2761750View Description Hide Description
Purpose: Most proton therapy beam delivery systems still use passive scattering and design patient apertures and compensators under the assumption of static anatomy. Treatment of moving tumors is typically avoided. We investigated the potential of proton therapy to treat lungtumors under free breathing conditions via a MIP‐based design of the aperture and compensator. Dosimetric effect evaluation of this approach was compared to two other strategies based on patient‐specific Internal Target Volumes (ITVs). Method and Materials: A ten phase 4DCT treatment planning study was performed for a 3‐field treatment of a lung patient. GTVs and normal tissue structures were delineated on 4DCT images. MIP images were generated by reassigning each pixel value to the maximum pixel value encountered in all 10 phases. Three sets of treatment plans were generated: Plan ITVEOI, Plan ITVMOE, and Plan ICTVMIP using the correspondingly designed patient aperture and compensator. Patient aperture and compensator were respectively optimized to ITVs derived from end‐of‐inhale or mid‐of‐exhale with 3D motion margins, or internal clinical tumor volume (ICTV) derived from MIP images. DVHs were calculated on ten phases using the same beam, aperture and compensator parameters to verify target dose coverage and dose to normal tissue following a prescribed dose 72Gy to the tumor.Results: Plan ICTVMIP assured the dose to 99% of the CTV (D99) through ten phases (AVG=97.40%, MIN=96.40%, SD=0.5), compared to the results from Plan ITVEOI (AVG= 71.00%, MIN =37.60%, SD=18.9) and Plan ITVMOE (AVG=94.70%, MIN=83.50%, SD=4.0). The average mean lung dose for each strategy was 14.60Gy (Plan ICTVMIP), 14.70Gy (Plan ITVMOE), and 15.20Gy (Plan ITVEOI). Conclusion: The MIP‐based patient aperture and compensator design provides superior tumor coverage and similar dose or lower dose to normal lungtissue compared to the designs based on ITVs derived from end‐of‐inhale or mid‐of‐exhale with 3D motion margins calculated from 4DCT.
34(2007); http://dx.doi.org/10.1118/1.2761751View Description Hide Description
Purpose: Intra‐fraction organ motion may cause significant dosimetric uncertainty with respect to tumor coverage. Accounting for this motion by enlarging the internal margin also increases dose to normal organs near the tumor. Therefore, it is desirable to ensure adequate target dosimetric coverage while maintaining the smallest possible irradiation volume. We present an optimization of the internal margin used for intra‐fraction organ motion and study its dependence on other treatment planning parameters. Methods and Materials: Using a commercial treatment planning system, we created 3D conformal treatment scenarios for spherical targets of varying sizes. The treatment plans were designed to cover target with additional setup uncertainty and internal margins. Target dose was accumulated over 100 fractions using a custom developed software simulating treatment execution uncertainties as well as respiration motion. The internal margin was considered to be optimal if the dose to a moving target was equivalent to dose given to static target planned with no internal margin. Results: The optimum internal amount was found to be much smaller than the full amplitude of the motion. Symmetric margins of 1, 3 and 5 mm were found to be adequate for peak‐to‐peak respiration amplitudes of 10, 15 and 20 mm respectively. The optimal internal margin was also observed to be approximately independent of the margin chosen for setup uncertainties. Optimizations for several patient target volumes also confirmed that margins smaller than full amplitude were adequate. Conclusion: Our findings present significant implications for treatment planning of mobile targets, such as tumors found in the lung and upper abdomen. Using the full motion amplitude for the internal margin is overly conservative, and optimization of the internal margin provides improved sparing of nearby organs at risk without sacrificing dosimetric coverage for the target. Conflict of Interest: Sponsored in part by GE Healthcare.
TH‐E‐M100F‐06: Estimation of the Error in Internal Target Volume (ITV) of Lung Tumor Obtained From Free‐Breathing Cine‐Mode 4DCT: A Simulation and Comparison Study Based On Dynamic MRI34(2007); http://dx.doi.org/10.1118/1.2761752View Description Hide Description
Purpose: To quantitate the error of tumor internal target volume (ITV) as determined from simulated free‐breathing cine‐mode 4DCT using dynamic magnetic resonance imaging (dMRI). Method and Materials: 8 healthy volunteers and 6 lungtumor patients underwent a 5‐minute MRI scan in the sagittal plane to acquire dynamic images of lung motion. A MATLAB program was written to simulate the cine‐mode 4DCT acquisition by segmenting and resorting the MRimages. Maximum intensity projection (MIP) images were generated from both simulated 4DCT (sCT) and dMRI, and the errors in MIP‐based ITV from sCT (ε), comparing to those from dMRI, were determined and correlate to the subjects' respiratory variability (ν). Results: MIP‐based ITVs from sCT were comparatively smaller than those from dMRI in both digital‐phantom studies (ε=−21.64±8.23%) and lungtumor patient studies (ε=−20.31±11.36%). The errors in MIP‐based ITV from sCT linearly correlated (, r2=0.76) with the subjects' respiratory variability.
Conclusions: Because of the low temporal resolution and retrospective resorting, 4DCT may not accurately depict the excursion of a moving tumor. Using 4DCT MIP image to define ITV may therefore cause under‐dosing and increased risk of subsequent treatment failure. Patient‐specific respiratory variability may also be a useful predictor of the 4DCT‐induced error in MIP‐based ITV determination.
TH‐E‐M100F‐07: Amplitude Gated Breath‐Hold Treatment for Upper Abdominal Lesions with On Board Imaging Guidance34(2007); http://dx.doi.org/10.1118/1.2761753View Description Hide Description
Purpose: To apply on‐boardmdash;image (OBI) guided amplitude gating for breath‐hold treatment and to assess treatment margin adequacy by analyzing the isocenter placement based on orthogonal 2D kV radiographs and 3D cone‐beam CT(CBCT) acquired under amplitude‐gated breath‐hold. Method and Materials: 25 patients with liver, pancreas, bile duct, or adrenal lesions were treated by using amplitude‐gated breath‐hold technique during the last 18 months. 2D orthogonal kV and 3D CBCTimages were acquired under amplitude‐gated breath‐hold and matched to the corresponding DRRs and planning CTimages, respectively, for isocenter placement. A total of 438 sets of 2D kV images and 70 sets of CBCTimages were analyzed. The margin of 95% probability, which warrants that the target will be within the treatment fields with a probability of 95%, was used to quantify the margin reduction of using image‐guidance for the breath‐hold treatment. Results: The average isocenter shifts based on 2D kV OBI matching over the breath‐hold laser alignment were 0.31 ± 0.27 cm, 0.28 ± 0.28 cm and 0.29 ± 0.28 cm along the lateral (LR), longitudinal (SI) and vertical (AP) directions, respectively. After the patients were moved to the isocenter based on 2D images, the additional shifts based on 3D CT‐CBCT matching were 0.28 ± 0.28 cm, 0.30 ± 0.37 cm and 0.32 ± 0.27 cm along LR, SI, and AP directions, respectively. The 95% probability margins of daily 2D OBI matching with respect to the 3D CBCT matching were 0.6 cm, 1 cm and 0.9 cm along LR, SI and AP directions. If image guidance technique was not used, the required margins of 95% probability increased to 1.1, 1.3 and 1.2 cm along the LR, SI and AP directions. Conclusion: Amplitude gated breath‐hold technique is useful for the treatment of upper abdominal lesions. Image‐guidance technique substantially reduced treatment margin.