F Li

FOUR-DIMENSIONAL COMPUTED TOMOGRAPHY-BASED INTERFRACTIONAL REPRODUCIBILITY STUDY OF LUNG TUMOR INTRAFRACTIONAL MOTION

PURPOSE To evaluate the interfractional reproducibility of respiration-induced lung tumors motion, defined by their centroids and the intrafractional target motion range. METHODS AND MATERIALS Twentythree pairs of four-dimensional/computed tomography scans were acquired for 22 patients. Gross tumor volumes were contoured, Clinical target volumes (CTVs) were generated. Geometric data for CTVs and lung volumes were extracted. The motion tracks of CTV centroids, and CTV edges along the cranio-caudal, anterior-posterior, and lateral directions were evaluated. The Pearson correlation coefficient for motion tracks along the cranio-caudal direction was determined for the entire respiratory cycle and for five phases about the end of expiration. RESULTS The largest motion extent was along the cranio-caudal direction. The intrafractional motion extent for five CTVs was <0.5 cm, the largest motion range was 3.59 cm. Three CTVs with respiration-induced displacement >0.5 cm did not exhibit the similarity of motion, and for 16 CTVs with motion >0.5 cm the correlation coefficient was >0.8. The lung volumes in corresponding phases for cases that demonstrated CTVs motion similarity were reproducible. No correlation between tumor size and mobility was found. CONCLUSION Target motion reproducibility seems to be present in 87% of cases in our dataset. Three cases with dissimilar motion indicate that it is advisable to verify target motion during treatment. The adaptive adjustment to compensate the possible interfractional shifts in a target position should be incorporated as a routine policy for lung cancer radiotherapy.

SU‐FF‐J‐56: Patient Dose From Kilo‐Voltage Cone Beam Computed Tomography (kV‐CBCT) Imaging

Purpose: To investigate patient dose from on‐board imager‐based kV‐CBCT. Method and Materials: Radiation doses from kV‐CBCT were measured using TLDs at different locations in three anthropomorphic‐phantoms (H&N, chest and pelvis) and patients retrospectively. kV‐CBCT scans were performed in standard settings (125 kV, 80 mA and 25 ms) using a Varian Trilogy linear accelerator. Both full‐fan (FOV=24 cm) and half‐fan (FOV=40 cm) modes were evaluated for H&N case while only half‐fan (FOV=45 cm) technique was studied for chest and pelvic cases. The skin dose in both phantoms and patients were measured at 4 locations: anterior, posterior, Rt‐Lat, and Lt‐Lat. Doses measured in the phantoms included different critical organs. The dosimeters used were high sensitivity TLD‐100H and only those with standard‐deviations less than 3% and sensitivity within ±3% were selected for this study. Each TLD was individually calibrated using an ion chamber under the irradiation condition. Phantom data was averaged from 3 separate measurements and patient data was averaged from 5 measurements in each category. Results: The skin dose for H&N cases were 9–10cGy for half‐fan mode in both the phantom and patients. The dose for brain and brainstem were 7.1cGy and 7.6cGy, respectively. The doses in same locations were 2–3cGy lower if the full‐fan mode is used. The skin dose for chest cases was 8–10cGy and were same for the phantom and patient measurements. Measured mean lung dose was 8.5cGy and spinal cord dose was 6.2cGy. For pelvis, measured skin dose was 2.9–4.2cGy and the prostate and rectum dose were 2.9cGy. Conclusions: For pelvic cases, kV‐CBCT dose was comparable or less than that from portal imaging. For chest and H&N cases the dose can be two times higher than that for the pelvis cases. Daily CBCT may lead to extra 400cGy to skin and 250cGy to spinal cord in 40 fractions.

SU‐GG‐T‐260: An Experimental and Monte Carlo Study of Output Factors for Small Radiosurgery Beams

Purpose: To accurately determine output factors(Sc,p) for small radiosurgery beams through experimental measurements using multiple detectors and Monte Carlo(MC) simulation. Method and Materials: A pinpoint chamber(PTW 31016), two solid state detectors(Scanditronix PFD and SFD) and radiographic film(Kodak EDR2) were used to measure Sc,p for twelve cones(φ5mm∼φ30mm) for a 6MV photon beam on a Varian Trilogy linear accelerator. All measurements were made at a depth of 1.5cm with SAD 100cm. MC models based on BEAM and XVMC codes were developed to simulate the treatment head and detectors to obtain theoretical values of Sc,p and correction factors for the detectors. The experimental data were compared with the MC values. The impacts of detector perturbation and energy dependence on the detectors behavior were also investigated. Results: For the cones larger than 20mm, the differences between the measurements and MC calculations are within 1% for all the detectors except SFD. As the cone size decreases, the differences between the chamber measured and MC calculated values increase significantly. For example, the chamber underestimates Sc,p by almost 23% for the 5mm cone. Sc,p values obtained with films and PFD give most consistent results for all cones. Although the SFD detector has a very high spatial resolution(0.6mm), the data obtained with the SFD detector is found to be less than the theoretical values by about 5%. MC simulations show that the mean photon energy is higher in small beams. The response of SFD detector is larger for low energy beams than that for high energy beams. We contribute 5% underestimation to the energy dependence of SFD detector.Conclusion: It is found that films and PFD diode provide reliable Sc,p results for small beams. MC simulation is a useful tool, which provides benchmark and correction factors for detectors used for determination of output factors in small radiosurgical beams.

SU‐GG‐T‐460: Multiple Gating for Lung Stereotactic Body Radiotherapy Treatment

Purpose: An inherent problem in treatment of lungcancers with stereotactic body radiotherapy(SBRT) is target motion. Multiple‐phase‐based gating refers to treating the target at two or more phase windows with plans optimized for each window individually. It closely approximates the four‐dimensional (4D) tracking technique when large number of gating windows is used. Compared to conventional single gating window technique, multiple‐phase‐based gating can improve the normal tissue sparing by taking advantage of the location change of the target. In addition, the duty cycle may be further improved. Compared to 4D tracking, the dose conformity of the multiple‐gating technique may be inferior. However, its implementation may be more robust because of the allowance of residual error inside each gating window. In this study, we evaluate the dose performances for different techniques including static, single gated, dynamic 4D tracking, and two‐window‐gated SBRT plans. Method and Materials: A motion phantom was scanned with 4DCT and planned using the above techniques. The motion was 1D sinusoidal with amplitude of 1cm. For any optimized plan, 95% of the PTV was covered by the prescription dose of 60Gy. Film dosimetry was performed to compare the doses delivered in each plans. The above techniques were also compared for a lung patient who was undergoing SBRT treatment, prescribed to 60Gy in 3 fractions. Results: Both the phantom and the patient studies showed similar results. If normalized to the static plan, the V20s were decreased by 56%,22%,16%, and the maximum doses in “lung” were changed to 102%,58%,45%, for single window gating, two‐window gating, and 4D tracking, respectively. Conclusion: Multiple‐gating technique has significant improvement in lungdose sparing compared with static or single‐gating techniques. Its difference from 4D plan is relatively small. The dose performance also highly depends on the size of tumor and the extent of target motion.

SU‐DD‐A4‐01: Model‐Based Deconvolution for 4D PET

Purpose: An inherent problem in four‐dimensional (4D) PETimaging is the poor statistics in each phase, because the total coincidence events are divided into several phase bins during image acquisition. We propose in this work a simple but efficient image post‐processing method to improve the 4DPET image quality. Method and Materials: In this method, the entire acquired coincidence events are used for each individual phase to enhance their signal‐to‐noise ratios (SNRs). A summed 3D PETimage is first obtained from the noisy 4DPET images, which represents the maximum SNR achievable. By deconvolving the 3D image with a deformable model derived from 4DCT, an improved 4DPET phase series can be obtained. For the best image quality, voice coaching was used to assure a regular and consistent breathing pattern during the course of PET and CT scans. The method was quantitatively evaluated with numerical and physical phantom experiments. Three clinical studies of pancreatic, lung and liver cancer patients were then carried out. Results: Numerical simulations showed that the model‐based 4D‐PET deconvolution method converged monotonically to the “ground truth” within a few iterations, and the SNR of the physical phantom images showed an increase of 83% over the conventional 4D PET without sacrificing spatial resolution. Similar performance was also observed for the patient study. Conclusion: We have developed a new method for improved 4D‐PET imaging. A salient feature of the method is that the coincidence events acquired at different time points are considered simultaneously when reconstructing each phase‐resolved image, leading to substantially improved 4D‐PET images.

SU‐FF‐J‐65: Feasibility Study of Management of Respiration Induced Target Motion for the Radiotherapy Treatment of Lung Cancer Patients in the Absence of a 4DCT Simulator

Purpose: Varians RPM™ system for respiration induced tumor motion management allows acquisition of CTimages and gated treatments under free breathing. 4DCT may not be possible because of lack availability of appropriate CT hardware or software. This study evaluated whether a breath hold CT scanning technique can be used as a substitute for a 4DCT scan. Materials and Methods: A 4DCT scan is obtained on a 4 slice GE Lightspeed™ scanner with the patient breathing freely and the respiratory period regulated using audiovisual cues from RPM™. Additional helical scans are obtained using an end inhalation or exhalation breath hold modified gating method (MGM). The PTV is drawn on the MGM scan(s) and for each of phase of the 4DCT scan. Comparison of target volume, centroid and extent of target volume is made between the MGM scan and the corresponding phase of the 4DCT scan. A treatment plan is developed using the MGM scan. Dose is recalculated using the 4DCT scan with the beams isocenter and apertures obtained from the MGM scan. DVH comparison is made. Results: 20 patients had both a 4DCT scan and at least one MGM scan. 8 patients exhibited respiration induced target motion of >5 mm during free breathing. Maximum target motion observed was 25 mm. For 14 end inhalation scans, 9 passed, 3 passed marginally, and 2 failed the equivalency tests to the corresponding 4DCT scan. For 18 end exhalation scans, 14 passed, 4 passed marginally, and 0 failed the equivalency tests to the corresponding 4DCT scan. Conclusion: All end exhalation breath hold scans are suitable substitutes for the corresponding phase 50 4DCT scan. However only 6/18 patients exhibited sufficient (>5 mm) respiration induced target motion on which to base any conclusions about the suitability of MGM. Conflict of Interest: Research supported by Varian Medical Systems.

SU‐FF‐T‐422: The Influences of Detector Energy Dependence and Perturbation On the Determination of Small Field Output Factors

Purpose: In radiotherapy, accurate dose determination requires accurate measurements of output factors. The employment of small field size cones in radiosurgery imposes significant challenges for such accurate measurements, since the detector perturbation and its energy dependence, along with the detector size, may introduce significant errors in the output factor determination. This study investigated the impacts on the output factors introduced by the detector perturbation and energy dependence. Method and Materials: Output factors were measured on a Varian Trilogy™ linear accelerator for various collimator defined small square fields and twelve radiosurgery cones that were shipped with the machine. The measurements were performed by using four different detectors, namely, a pinpoint ion chamber, Koda EDR2 films, a stereotactic diode detector, and a pinpoint diode detector. The energy dependence was also evaluated for the ion chamber and diode detectors. Results: Both perturbation and energy dependence present non‐trivial effects in the determination of the output factors. The impact of detector perturbation was more pronounced for the small fields, while that of energy dependence was evident for all the field sizes investigated. For a cone of 5 mm diameter, the perturbation introduced by using a PinPoint ionization chamber could result in almost 30% reduction in the output factor value. The deviations caused by the energy dependence varied from 3% to 6% depending on field sizes.Conclusions: The output factor measurements of small fields must not only account for the size of detector but also its perturbation and energy dependence to ensure the measurement accuracy for stereotactic radiosurgery.

SU‐FF‐T‐156: Multi‐Institutional Retrospective Analysis of IMRT QA Measurements

Purpose: To review IMRT QA measurements from several of the 50+ institutions for which we provide IMRT treatment plans and determine if institutional, anatomic site, or measurement biases exist. Method and Materials: For each patient receiving IMRT, the treatment plan is delivered to a solid water phantom and the dose measured using a small volume ion chamber and with a single EDR film placed 1 cm above the chamber plane. Of the almost 3000 IMRT treatment plans calculated and delivered in 2004, more than 1000 random, de‐identified plans were reviewed. Ratios of chamber/calculated and film‐center/calculated doses were tabulated for six anatomic sites (breast, prostate, pelvis, head & neck, brain, and other). Film dose distributions were compared to calculations using one of several commercially available QA packages. Results: The institutions with the best results had average errors of less than ±0.5% (i.e. randomly distributed about zero) with standard deviations of 1.25–1.50%. A few centers had average errors and standard deviations approaching 3%, indicating a bias in which a systematic dose measurement error was found. Agreement between chamber and film center dose was also institution specific with the best results found for those centers that had the lowest errors compared to calculation. One institution had excellent agreement between chamber and calculation (−0.2±1.7%), but 2–3% lower film dose. Although exceptions were found, little variation in the agreement between chamber measurement and calculation occurred as a function of anatomical site. Conclusion: Since all treatment plans were calculated in one central location and many centers had excellent agreement between measurement and calculation, it is likely that the higher errors were due to measurement technique rather than errors in the dose calculation. Error was not anatomic site dependent possibly due to the purposeful placement of the ion chamber in a region of relatively uniform dose.

SU-FF-T-32: Imaged-Based Simulation Technique To Determine Stepping Source Dwell Position For MammoSite(r) Brachytherapy Procedures

Purpose: Incorrect dwell position for the stepping source in MammoSite® radiation therapy system would result in severe dose error to the treated volume. In many centers, CT‐simulators have replaced the fluoroscopic simulators. An alternative method must be developed for this purpose. This project evaluates the feasibility of CT‐based simulation to determine the dwell position for the stepping source of the Nucletron® High‐Dose Rate (HDR) unit. Method and Materials: A MammoSite® balloon applicator is placed in the surgical cavity intraoperatively at the time of segmented mastectomy for breast cancer. The balloon is inflated to near spherical shape with saline solution mixed with a small amount of radiographic contrast to aid in visualization. After recovery, the patient is brought to the radiation oncology facility to determine the quality of the implant and also to determine the stepping source dwell position. A dummy source train is initially inserted in the applicator and pushed to the distal end. The distance is measured using the Nucletron® measuring tool. CT scans of the breast was taken with 1 mm slice thickness. After the images have been acquired, a virtual 3‐dimensional breast is generated. Based on the virtual breast, the path of the dummy source train inside the applicator is assessed. Results: A digitally reconstructed radiography(DRR) that maximizes the projection of the pathway is created. A method is formulated to determine the center of the sphere and marks on the source pathway. The dwell position is determined by subtracting the difference of distance between the distal seed and center of the sphere from the maximum source distance as set on the HDR unit. Conclusion: For institutions where the fluoroscopic simulator has been replaced by a CT‐simulator, imaged‐based simulation allows an effective method of determining the stepping source dwell position for MammoSite® brachytherapy procedures.

SU‐FF‐T‐169: The Use of Diode in In‐Vivo Dosimetry Quality Assurance in IMRT

Purpose: In IMRTtreatments, the ultimate QA procedure is to carry out in‐vivo dosimetry measurement to ensure the accuracies of both patient setup and beam delivery. This study was designed to explore the use of in‐vivo diode dosimetry measurement for QA of IMRTtreatments. Method and Materials:IMRT plans were generated based on a set of CT scans of a head & neck anthropomorphic phantom. Corresponding IMRT QA verification plans were also generated. Diode calibration readings (Rc) were obtained for each beam during the routine dose verification QA process. During verification, a diode was placed along the beam central axis on the surface of a flat QA phantom at the SSD specified in the QA plan. Radiation was delivered dynamically using the same dynamic MLC files that were to be used for the patient treatment. For in‐vivo measurements, the anthropomorphic phantom was setup according to the treatment plan. For each beam, a diode was placed along the central axis at the beam surface entry point. Radiation was then delivered according to the plan and the diode reading (Ri) was recorded. If both the setup and the beam delivery were correct, Ri should be in agreement with quantity Rc*fSSD within certain uncertainty (fSSD is SSD correction factor); otherwise, it would be an indication of incorrect patient setup or incorrect beam delivery. Results: It was found that the calibration diode readings followed the SSD inverse square law within an uncertainty of 0.4%. f ssD = ( SSS c SSD i ) 2 . The derived in‐vivo diode readings (Rc*fSSD) were in agreement with those measured ones within 3.6% for three beams at different gantry angles, with an average difference of 1.8%. Conclusion: With a proper calibration method, diode verification can be used relatively accurately for in‐vivo measurements to check on the accuracies of patient setup and beam delivery for IMRTtreatments.