Index of content:
Volume 33, Issue 6, June 2006
- Joint Imaging/Therapy Scientific Session: Valencia A
- Correction Strategies
WE‐E‐ValA‐01: Effect of Action Level and Uncertainties in Daily Imaging and Re‐Positioning On the Distribution of Inter‐Fraction Setup Uncertainty33(2006); http://dx.doi.org/10.1118/1.2241798View Description Hide Description
Purpose: To predict the post‐correction probability distribution functions (PDFs) for inter‐fraction setup uncertainty of individual patients who will undergo daily localization with an action level for setup correction. Method and Materials: An analytical method was developed to derive the PDF at a given action level assuming 1) a Gaussian distribution for the pre‐correction setup uncertainty, and including, 2) uncertainty in online localization, and also 3) the uncertainty with which patients can be repositioned. An interactive spreadsheet was developed to evaluate and graph the PDF, as well as its mean and variance. Plots of the mean and variance of the PDFs predicted at different action thresholds for user specified (or patient‐derived) levels of the three input uncertainties above were used to develop practical action level rules. Results: When the variance of the localization uncertainty is the smallest of the three sources of uncertainty, there is an optimal action level that minimizes post‐correction setup uncertainty. There is quantitative (and graphically demonstrated) improvement when , the sum of the variances of the localization uncertainty and re‐positioning uncertainty, is less than , the variance of the pre‐correction setup uncertainty. A practical rule is to set the action level to σg(σs/σg)0.3 in these situations. The overlap of a resulting PDF with a Gaussian distribution with the same mean and variance is typically well over 90% when the action level is set according to this rule. Conclusion: The analytical method developed here is a useful tool to estimate the post‐correction setup uncertainty at different action levels, and to set rules for clinical specification of the action level in cases where the precisions of localization and setup correction allow an improvement. It also permits evaluation of potential improvements in post‐correction setup uncertainty associated with improved precision in daily localization and/or patient repositioning.
WE‐E‐ValA‐02: Dosimetric Comparison of the No Action Level Alignment Protocol with Daily Alignment Techniques for Prostate Cancer33(2006); http://dx.doi.org/10.1118/1.2241799View Description Hide Description
Purpose: To compare the effectiveness of two off‐line “No‐Action‐Level” (NAL) correction protocols with daily image‐guided alignments using bony registration (simulating electronic portal image alignment), ultrasound, and CT for direct prostate target localization. Method and Materials: Ten prostate patients received 3 CT scans per week using an integrated CTLINAC system immediately prior to radiotherapy (243 CT scans total). A clinical treatment plan was designed on the planning CTimage using current clinical margins and copied onto the daily CTimages. Two NAL protocols, based on CT measurements of the internal prostate shift relative to bony anatomy, were simulated for correcting the predicted internal systematic prostate shifts after 1 week or after 2 weeks of treatment. The NAL protocols were compared to three daily alignment methods, which simulated pelvic bone alignment, ultrasound alignment, and CT alignment. The dosimetric impact on target coverage for each scenario was reported. Reducing the planning margins to 3mm was also evaluated. Results: Daily CT scans are more accurate than daily ultrasound measurements for determining the prostate systematic positional shift, particularly in the anterior/posterior direction. The average minimum prostate dose was greatest with CT alignment (75.8Gy, p<=0.028), then with the two NAL protocols (both 74.Gy, p<=0.017), followed by ultrasound alignment (73.2Gy) and bone alignment (70.2Gy). For plans with 3mm margins, the average minimum dose was greatest with CT alignment (75.1Gy, p<=0.007), then with the two adaptive alignments (71.4Gy and 70.8Gy respectively, p<=0.022), followed by ultrasound alignment (68.4Gy) and bone alignment (63.9Gy). Conclusions: An off‐line NAL correction protocol for reducing systematic internal target shifts proved to be effective when performed after only one treatment week. The target dosimetric coverage from the NAL protocol was as good as daily ultrasound alignments but not as great as daily CT alignments. Using a 3mm planning margin exacerbated the differences in target coverage.
33(2006); http://dx.doi.org/10.1118/1.2241800View Description Hide Description
Purpose: To evaluate the dosimetric consequence of 3D positioning verification for prostate IMRT treatment using CBCT.Method and Materials: Patients in this study were repositioned using 2D orthogonal radiographicimages, taken prior to treatment to match 2D bone structures between the raidiographic and reference images. Following this, CBCTimages were acquired, then the treatment was delivered without an additional shift. A verification plan, CBTreat, was generated based on the CBCT to simulate the actual treatment achieved with positioning verification based on 2D bone structure match. A verification plan, CBBone, was created based on CBCT with the isocenter shifted to match 3D bone structures between CBCT and planning CT. A verification plan, CBSoft, was created based on CBCT with the isocenter shifted to match 3D soft tissues between CBCT and planning CT. These three verification plans were created for 17 patients for the first treatment fraction and compared to the original plans. Results: The average dose coverage of prostate/seminal vesicle (SV) and dose to 30% of bladder/rectum showed very similar results for all three verification plans. Individual dose‐volume histograms (DVH) displayed similar distribution for CBTreat and CBBone of all 17 patients. However, DVHs of CBSoft indicated that the coverage of prostate and SV was improved significantly for a few patients at the cost of increased dose to bladder/rectum. Current patient repositioning is limited to translational couch shift although we observed several patients with variations (e.g. prostate deformation due to rectal gas, bladder filling, volume variation) that could not be resolved by the couch shift. Conclusion:CBCT provides substantial bony and soft tissue information. It also reveals that the prostate is often deformed and simple translation correction will not improve the treatment accuracy. Therefore, customizing margin or adaptive therapy is essential for those patients.
Partially supported by Varian research grant.
33(2006); http://dx.doi.org/10.1118/1.2241801View Description Hide Description
Purpose: To study how important it is to consider organ deformation under volumetric image guidance. Two questions were answered: 1) how much residual misalignment exists after rigid‐body image registration; and 2) what is the dosimetric impact if deformations is ignored by simply shifting the patient. Methods: 10 CTs were acquired on non‐consecutive treatment days for 20 patients receiving radiation therapy of their prostate cancer under an IRB approved protocol. One physician contoured the prostate, rectum and bladder for all scans. To answer the first question, after rigid body registration of each image set acquired during the course of treatment with the planning CT, we measured the distance between the prostate boundaries along the three axes. For the second question, we copied the original plan with the shifts determined by rigid‐body registration and compared with re‐optimized plans based on the images of the day. Plan optimization was performed using the same dose and dose‐volume constraints in the initial planning. Results: 10 patients were analyzed so far. Because these mismatches were measured after rigid‐body registration, they were indicative of the prostate deformations during therapy. The mismatches varied widely among patients. The results were similar to that reported by Deurloo at al . The largest deformation was seen in the A/P direction of patient 3, with a mean of 5.2mm and a standard deviation of 2.3 mm. For most cases with small target deformation, shifting the patient produces similar plans as re‐optimization. In cases where there were substantial organ deformations, the plans resulting from translation were much worse than the re‐optimized plan. Conclusion: Simply shifting the patient can be far from optimal. An IGRT scheme that can handle both translations and deformations is desired.
33(2006); http://dx.doi.org/10.1118/1.2241802View Description Hide Description
Purpose: To determine the feasibility of a moving treatment couch to compensate for real‐time respiration‐induced 3D tumor motion observed in patients. Methods: The couch dynamics were modeled as a critically damped second order system with dead time. The controller was modeled as a first order system to simulate the model dynamics mismatch between the couch and the controller. The feedback system was modeled as a closed‐loop internal modelcontrol system and the parameters to describe the dynamics were obtained from previous feasibility studies. To determine the performance of this system, the average tumor trajectory data derived from 4D CT for 14 patients was considered. To simulate variations in normal intra‐fraction respiration patterns for a given patient, distributions in amplitude and period were modeled and the residual tumor motion determined. The output of the control system was analyzed by evaluating the distribution of residual tumor motion. Furthermore, a detailed analysis of the residual motion as a function of tumor amplitude and velocity was conducted. Results: The mean 3D amplitude of uncompensated tumor motion was 7.1 ± 4.6 mm for 14 patients. Following feedback control, the mean residual tumor motion was 0.35 ± 0.20 mm with a mean respiratory period of 4 s. The residual motion was under 3 mm for all patients, for the range of time constants investigated. The response of the couch correlated linearly with instantaneous tumor velocity for the range of parameters used to describe the system dynamics (R2 = 0.98). Conclusion: The treatment couch can be used to compensate for real‐time tumor motion, given real 3D tumor trajectories. Conflict of Interest: Supported by 3DLine Medical Systems.
33(2006); http://dx.doi.org/10.1118/1.2241803View Description Hide Description
Purpose: We present feasibility studies in support of a real‐time MRI guided external beam radiotherapy delivery system currently under commercial development. Method and Materials: The system, (ViewRay Inc., Renaissance™), combines a low field open MRI scanner and a multi‐headed 60Co γ‐ray IMRT unit equipped with multi‐leaf collimators. It is designed so that the center of the field of view of the MRI and the isocenter of the radiotherapy unit coincide. The inherent compatibility of the units allows for the acquisition of fast ciné MRI simultaneous to radiotherapy delivery to assess intra‐fraction organ motion. Computational feasibility studies were performed to investigate: the compatibility of the MRI and the 60Co γ‐ray IMRT unit; the impact of the MRImagnetic field on the dosimetry; and the feasibility of performing accurate heterogeneity dose computations with MRI data. Results: The 60Co γ‐ray IMRT unit was found not to significantly impact the operation of the MRI; the γ‐ray IMRT unit is capable of producing high quality IMRT treatment plans; the MRImagnetic field eliminates contamination electrons and does not significantly perturb the dose distribution in lung, soft tissue, and bone; and accurate heterogeneity dose computations are possible employing only MRI data. Conclusion: Performing IMRT allows for the seamless integration with, and simultaneous operation of, an open MRI unit. Conflict of Interest: Research sponsored by ViewRay, Inc., Gainesville, Florida USA.
33(2006); http://dx.doi.org/10.1118/1.2241804View Description Hide Description
Purpose: The ability to reconstruct the delivered patient dose can help ensure that the integrated doses to important structures faithfully adhere to prescription. The objective of this work is to develop and test a 3D dose reconstruction procedure based on exit‐dose measurements at treatment time and MV cone‐beam CT.Methods and Materials: The proposed dose reconstruction method uses a Megavoltage cone‐beam computed tomography (MVCBCT) image acquired on the treatment table prior to treatment, 2D portal images taken with an amorphous‐silicon electronic portal imaging device(EPID) during treatment, and an independent validated dose calculation engine. The energy fluence obtained from the EPID is back‐projected through the 3D MVCBCT image. A dose calculation engine based on a collapsed‐cone convolution algorithm subsequently calculates the dose in each voxel. To test the model, a MVCBCT of a cylindrical solid‐water QA phantom was acquired and the MVCBCT numbers mapped to appropriate attenuation coefficients. The phantom was then treated with a 5cmx5cm beam and a portal image acquired. During the treatment, a CC13 ion chamber and MOSFETdetectors were used to measure the dose at 21 points to compare with reconstructed dose. A Pinnacle dose calculation using a conventional CT was also performed for comparison. Results: The mean difference between reconstructed and measured doses was −0.2% (standard deviation = 2.8%). The reconstructed dose in the inner regions of the cylinder differed less than 2% from the measured, although discrepancies of about 10% occurred at one point in the buildup region and at two other peripheral points. In comparison, the mean difference between Pinnacle calculations and measurements was −2.9% (standard deviation =1.6%). Conclusion: Preliminary calculations of reconstructed dose demonstrated good agreement with experiments. Further refinement of the model and its application to clinical conditions are under investigation. Conflict of Interest: Research supported by Siemens.
- Modeling of Intra‐Fraction Organ Motion
33(2006); http://dx.doi.org/10.1118/1.2241691View Description Hide Description
Purpose: To assess the effect of incorrect assignment of respiration phases and irregular breathing on 4DCT image quality. Artifacts are manifested as deformations in the reconstructed images and quantitative effects are measured along with qualitative evaluations. Modifications to the current 4DCT implementation are recommended. Methods and Materials: For the evaluation of image artifacts, we used a motion simulation platform and simulated the respiration patterns of real patients. Artifacts due to inaccurate phase sorting are quantitatively evaluated by comparing differences in the volumes of spherical phantoms with and without the presence of phase assignment problems. Artifacts due to irregular breathing are demonstrated using 4DCT and recommendations are made for modifying the standard acquisition mode to enable gating for motion reproducibility. The advantage of this modified acquisition is proved using an electronic portal imager.Results: Review of clinical 4D scans performed in our clinic showed discrepancies in the phase assignments for about 45% of the cases when compared to our independent check. Significant image artifacts are also observed and measured as a function of the respiration motion amplitude and target size. Volumetric inaccuracies of up to 43% are measured. For the evaluation of irregular breathing, our proposed technique of gated imaging for the reproducibility of the respiration proved to yield superior image integrity when compared to standard acquisition mode. Conclusions: We identified two sources of quality degradation factors associated with 4DCT images and performed quantitative evaluations of associated artifacts. We conclude that for improved image reconstruction, an independent check of the sorting procedure should be performed for each clinical case; also we recommend a modification to the 4DCT acquisition technique to include gating of the x‐rays for the reproducibility of the respiration pattern. Conflict of Interest: Research supported by GE Medical Systems.
WE‐C‐ValA‐02: The Impact of 4D Breathing Motion Effects Versus Tissue Heterogeneity in Lung Cancer Treatment Planning33(2006); http://dx.doi.org/10.1118/1.2241692View Description Hide Description
Purpose: To investigate the relative magnitudes and clinical importances of the dosimetric effects related to 4D breathing motion and tissue heterogeneity for thoracic tumorstreatment planning.Methods: Scans were acquired at normal exhale/inhale breathing phases. The target was the union of the exhale and inhale GTVs, uniformly expanded by 5mm(ITV). Patients were planned with both AP/PA and 3‐D conformal plans using the exhale (“static”) dataset, assuming unit density, for 100±5% ITV dose coverage. Each of these plans was further used to calculate: (a) heterogeneous “static” dose; (b) homogeneous cumulative dose; (c) heterogeneous cumulative dose. The same number of MU were used for each of the calculations and was based on the homogeneous “static” plan. Cumulative dose distributions consisted of a time‐weighted sum of exhale and inhale doses. Doses were calculated using the DPM_MC code which includes secondary electrontransport for the heterogeneous computations. Results: Relative to unit‐density plans, tumor EUD, and lung NTCP increased in the heterogeneity corrected plans; primarily due to the reduced beam attenuation through lungs and the larger than coin‐size tumors investigated. In comparing 4D cumulative dose plans with static plans, clinical EUD and NTCP estimates were relatively unchanged. The insignificant tumor EUD change was a consequence of good target design, while the small lung NTCP change was due to its large volume effect. Accounting for tissue heterogeneity resulted in average changes of 10% in MLD. Accounting for 4D breathing motion effects resulted in <1% changes in MLD from the static value. The magnitude of these effects was not correlated with the dose distribution conformality. Conclusions: In this study we found that tissue heterogeneity effects are likely to have a larger clinical significance on tumor (if ITV is properly designed) and normal lung clinical treatment evaluation metrics than occurs with 4‐D respiratory‐induced changes.
Supported by P01‐CA59827, R01‐CA106770.
WE‐C‐ValA‐03: The Use of CT Density Changes at Internal Tissue Interfaces to Monitor Respiratory Induced Lung Tumor Motion33(2006); http://dx.doi.org/10.1118/1.2241693View Description Hide Description
Purpose: To describe a non‐invasive method to monitor the motion of internal organs affected by respiration without using external markers or spirometry, to apply the method to construct 4D‐CT datasets, to test the correlation with external markers, and to calculate any time shift between the datasets.
Method and Materials: Ten lungcancer patients were CT scanned with a General Electric Fast 4‐Slice CT scanner operating in ciné mode. An external signal was also acquired simultaneously using the Real‐Time Position Management (RPM) Respiratory Gating System (Varian Medical Systems). We retrospectively reconstructed the raw CT data to obtain consecutive 0.5s reconstructions at 0.1s intervals to increase image sampling. We defined regions of interest containing tissue interfaces that move due to breathing on each axial slice and measured the mean CT number as a function of respiratory phase. We constructed 4D‐CT data sets by retrospectively sorting each image set based on the respiratory phase determined by the mean CT number curve. The external marker and tumor motion were directly correlated using the sample coefficient of determination, r 2. Any time shift between the two data sets was calculated by shifting the tumor motion curve until r 2 was maximized. Results: Only three of the ten patients showed correlation higher than r2=0.80 between tumor motion and external marker position. However, after taking into account time shifts (ranging between 0s and 0.4s) between the two data sets, all ten patients showed correlation better than r2=0.8. Conclusions: 4D‐CT acquisition using an internal method improves the temporal registration of CTimages affected by respiratory motion without the need for external markers or spirometry. A non‐invasive method to directly correlate the motion of external markers and internal organs can be used to help guide decisions regarding the validity of the RPM system for respiratory gated radiotherapy on a patient‐specific basis.
WE‐C‐ValA‐04: Derivation of the Tumor Position From External Respiratory Surrogates with Periodical Updating of External/internal Correlation33(2006); http://dx.doi.org/10.1118/1.2241694View Description Hide Description
Purpose: To develop techniques that can derive the tumor position from external respiratory surrogates through periodically updated internal/external correlation. Method and Materials: A simple linear function is used to express the correlation between tumor and surrogate motion. The function parameters are established during patient setup session with both tumor and surrogate positions measured at 30Hz rate. During treatment, the surrogate position, constantly acquired at 30Hz, is used to derive the tumor position. Occasionally, a tumorimage is acquired to enable the updating of the correlation function. Four update methods are investigated. (a) Line shift. (b) Fit model — through point. (c) Fit model — extra weight. (d) Function difference — fit point. Results:Tumor and external surrogate motion demonstrates a high degree of correlation however it dynamically changes over time. Occasionally updating the correlation function leads to more accurate predictions than using external surrogates alone. At the lowest tumorimaging rate tested in this work (0.1Hz) an accuracy improvement of 10% over the prediction by the mere use of external surrogate was observed for the best update method. Update methods (a) and (b) derive the tumor position with larger accuracy than (c) and (d) in case of high imaging rates. The opposite is observed in case of low imaging rates. Conclusion: Occasional calibration of the tumor/external surrogate correlation during treatment substantially increases the accuracy of the tumor localization compared to tumor position derivation by using the external surrogate alone.
This work is partially supported by CenSSIS.
33(2006); http://dx.doi.org/10.1118/1.2241695View Description Hide Description
Purpose: To show the feasibility of using a thermocouple for monitoring patient breathing and to compare its breathing signal with that of a spirometer. Method and Materials: A mask covering the nose‐and‐mouth region was used to channel the subject's breath to a K‐type thermocouple (TC) and a spirometer. Two different placements of the TC were studied: either it was inserted through a side hole of a plexiglass tube connecting the mask and spirometer or in the oxygen inlet of the mask. This setup allowed simultaneous readings of the temperature and air volume in synchrony. The temperature change is a response to the amount of heat deposited to or taken away from the TC junction. Both signals were collected at the sampling rate of 100 KHz using a data acquisition board. The acquisition program was written in LabWindows. Results: The TC temperature and measured air volume are found to be highly correlated at the two locations. When placed in the oxygen inlet away from the direct inhale/exhale air streams, the sensing temperature shows a cubic dependence on the flow rate. When placed in the plexiglass tube where it was more directly exposed to the inhale air, the temperature exhibits some hysteresis on the flow rate. At both locations, the temperature signals show no drift for 10 minutes of breathing. Conclusions: The proposed TC system can be used as an external surrogate to monitor patient breathing at both TC locations. It has the advantages of spirometry in that it directly responds to the lung air flow. The setup is very easy and has no setup error. However, it does not have the drift problem that plagues spirometry. It is also relatively inexpensive. All these qualities make it attractive and suitable for respiratory gating or tracking to deliver external beam radiation therapy.
33(2006); http://dx.doi.org/10.1118/1.2241696View Description Hide Description
Purpose:Lungtumor breathing motion is a function of both the breathing depth (tidal volume) and breathing rate (airflow). Treatment planning for lungtumors will require a patient‐specific tumor motion model. An understanding of the patient's typical breathing cycle will be critical to accurate treatment planning predictions for linear‐accelerator gating or tracking. This work examines patient‐breathing characteristics to determine the patterns and stability of the breathing cycle. Method and Materials: A total of 34 patients (with and without lungcancer) were scanned with a previously described 4DCT protocol under synchronized tidal breathing monitoring. We examined the scatter plot of air flow against tidal volume for the entire scan session. A 2D histogram was then generated by calculating the frequency of data points falling into each tidal volume‐airflow window. The 2D histogram depicted the probability that the patient breathed at certain tidal volume and air flow window. We evaluated the breathing consistency of the three patients with multiple scans. Results: There was no significant differences in the breathing period (P=0.11) or in the peak‐to‐peak amplitude (P=0.22) between patients with lungcancers and patients with upper‐abdomen cancers. The 2D histogram revealed two different breathing patterns: patients who spent more time breathing at the end of exhalation (20 patients), exhibiting a characteristic volume‐flow curve, and patients who spent the majority of time inhaling and exhaling (14 patients). The latter patients did not spend any appreciable amount of time between successive breaths. For the three patients with two sessions, the mean frequency difference between corresponding tidal volume‐airflow windows was <0.3%. Conclusion: We characterized patient breathing by examining the tidal volume‐airflow plot and histogram. The results showed two different breathing patterns suggesting that not all patients are appropriate for gated radiotherapy. All three patients showed high consistency in two scan sessions.
Supported by NIH R01CA96679.
33(2006); http://dx.doi.org/10.1118/1.2241697View Description Hide Description
Purpose: Quantified characterization and better understanding of tumor respiratory motion is valuable for understanding of respiration, motion‐included treatment planning, online prediction, and real‐time control algorithm for dose delivery in image guided radiotherapy. There are two goals of this work: (1) to discover the correlation among various motion variables so that we can understand patient respiratory better and (2) to build an analytical system for online motion modeling and prediction during real‐time treatment delivery. Method and Materials:Statistical analysis of tumor respiratory motion has been performed over 48 real patient data. Quantified information of different motion characteristics, including amplitude, frequency, velocity and the mean positions are computed over different granularities. Sample granularities include a breathing state, a breathing cycle, a treatment session, a patient and the whole patient population. Association rules among different motion characteristics are mined and formulated. Results: We have implemented the software packages for statistical analysis and correlation presentation. Quantified motion information have been computed and displayed. The spatio‐temporal changes of these properties are studied. Knowledge of respiratory motion and the underlying physiological explanation have been exploited. The probability distribution functions of various correlations among different properties have been calculated and visualized. Conclusion: Different statistical analyses over a set of tumor motion characteristics have shown that there are some general rules regarding tumor motion. The analytical results help us to obtain new knowledge and to understand the physiological actions of tumor motion and to treatment moving tumor more efficiently. Conflict of Interest:
WE‐C‐ValA‐08: Development of a Patient Specific Respiratory Motion Model for Predicting Dose to Moving Organs During Radiotherapy33(2006); http://dx.doi.org/10.1118/1.2241698View Description Hide Description
Purpose: Respiratory motion is an important limitation to accurate calculation of dose to organs in the thorax and abdomen during treatment. We describe a model to estimate 3‐dimensional motion in patient CTimages. This model can be parameterized by navigators such as a signal from a respiratory monitor, or diaphragm position measured in fluoroscopy. We evaluate the accuracy of the model in predicting anatomic changes in respiration‐correlated CT (RCCT) images of lung cancer patients. Method and Experiment: Our method makes use of deformable image registration, thereby automating the modelcalibration and providing a complete determination of 3D trajectories for all tissue voxels of the moving organs.Calibration of the model uses a series of RCCT images. Each 3D image in the series is tagged by two navigators: current diaphragm position and precursor position which distinguishes between the inspiration and expiration portion of the respiratory cycle. Nonrigid registration is used to calculate the deformation field that maps each 3D image to a reference 3D image at end expiration. We perform a principle component analysis (PCA) to determine the 3D deformation parameterized by navigators. We evaluate the model by comparing the predicted 3D images with actual 3D images at different phases in the breathing cycle. Results: For RCCT images, predicted images calculated from the first two principal components at different respiration phases are found to accurately reproduce anatomic motion observed in the actual images, indicating the model's ability to predict the voxel trajectory anywhere in the cycle. Furthermore, the model can predict motion induced changes that were not in the original image set, such as deeper and shallower breathing. Conclusion: Preliminary experimental results indicate that the proposed method is a potentially useful tool for treatment planning and evaluation of dose to moving organs when patient breathing is monitored.
33(2006); http://dx.doi.org/10.1118/1.2241699View Description Hide Description
Purpose: We are evaluating the feasibility of a dynamic biological lung phantom for IGRT studies, with the initial goal of developing a reliable phantom suitable for use in validation of deformable registration and volume rendering studies of the lung. The properties of an ideal lung phantom would include complex geometry, anisotropic inflation, and composition, lobar structure and internal airway architecture similar to that of human lung.Method and Materials: Preserved swine lung was obtained and compared to human lung. The prepared lung was statically inflated to different volumes using a regulated nitrogen supply, and can also be dynamically inflated using a medical ventilator. The inflated phantom was imaged on a GE Lightspeed CT scanner. Volume rendering of the CTimage data was performed to visualize and determine coordinates of airway bifurcations.Results: Preserved swine lung was determined to be comparable to human lung in terms of tissue radiological and physical properties, lobar structure, airway architecture, volume and mass. Rendered airway vs. physiologic airway dimensions are undergoing verification by dissection. Analysis of CTimages and volume rendering data demonstrates that the airway architecture may be followed to at least the 5th airway bifurcation, yielding a conservative minimum of 31 reproducible anatomic landmarks evenly distributed throughout the lung. By visual inspection, it is possible to follow the displacement vector of these landmarks in sequential images.Conclusion: Initial analysis shows that a swine lung phantom meets a number of the requirements of a reliable and functional phantom for validation of deformable registration and volume rendering methods. Reference points generated using the CT/volume rendering technique may be useful as a validation tool for both feature‐ and intensity‐based deformable registration techniques. Ongoing study will evaluate the potential of the lung phantom for use in planning, delivering, and validating 4D IGRT.
33(2006); http://dx.doi.org/10.1118/1.2241700View Description Hide Description
Purpose: To develop an artifact‐free four‐dimensional (4D) cone‐beam CT(CBCT)imaging technique for image‐guidedradiotherapy, and to optimize the image quality, scanning time, and patient radiationdose with respect to acquisition parameters such as the number of gantry rotations, gantry‐rotation speed, and x‐ray tube current. Method and Materials: A Varian Trilogy™ system was employed for this study. To sort CBCT projections in terms of patient breathing phase, a CT‐opaque fiducial was adhered to the patient skin and tracked automatically in the projection space, and the phases of the fiducial trajectory were used to tag the projections. Projections after phase‐binning were subsequently reconstructed to yield 4D‐CBCT images. To reduce/eliminate view‐aliasing artifacts due to limited number of projections in each phase, both “multiple‐rotation” and “slow‐gantry‐rotation” strategies were investigated. Quantitative evaluation and comparison were performed with a motion phantom for the following acquisition settings: (i)varying number of gantry rotations (1∼8) with all other scanning parameters kept the same; (ii)varying x‐ray tube current together with the number of rotations while keeping the radiationdose constant, namely, 1‐rotation‐80mA, 2‐rotation‐40mA, 4‐rotation‐20mA, and 8‐rotation‐10mA; (iii)varying gantry‐rotation speed (1deg/sec∼8deg/sec).Three patient cases were used in the study. Results: The image quality, represented by relative error (RE), varied nonlinearly with the number of rotations: RE reduction became less pronounced as the number of rotations increased. After 3∼5 rotations, the benefit resulting from more rotations started saturating. When the dose level was kept the same, the images of 1‐rotation‐80mA and 8‐rotation‐10mA acquisitions had the largest and smallest REs, respectively. Similar RE reduction behaviors were found when slowing down the gantry‐rotation speed. For the patient cases, 4D images were obtained with negligible motion or view‐aliasing artifacts. Conclusion: We have successfully demonstrated that the commercially available OBI system can be utilized to acquire artifacts‐free 4D‐CBCT images without increasing the patient radiationdose.