Proceedings of the 50th Annual ASTRO Meeting S543

produce comparable results with better treatment efficiency. We compared dose distributions generated with VMAT, conformal3D-RT and step-and-shoot IMRT for typical anal cancer planning target volumes (PTV).

Materials/Methods: 8 typical patients with anal cancer treated previously at our department were chosen. A typical PTV includingthe primary tumor, pelvic and inguinal lymph nodes was planned to receive a mean dose of 36 Gy. VMAT plans were generatedwith ERGO++ 1.7 (Elekta) version based on a recently implemented modified Bortfeld-algorithm. This technique with 2 rotationswas compared to a 9 beam step-and-shoot IMRT (Corvus 6.3, Nomos) and the routinely applied refined 3D-RT (Masterplan, Thera-nostic). All three planning systems used the same structure set for OAR and PTV on identical CT datasets. These 3 treatment tech-niques were compared using dose-volume histograms (DVH) of the OARs and PTV. In addition, the conformality index (CI) andthe homogeneity index (HI) as described in the RTOG guidelines and the total treatment time (TTT) were analyzed. All data arepresented as mean values ± standard deviation (SD).

Results: In the high dose region the VMAT showed the best coverage of the PTV described by the following values (isodose aspercentage of prescription dose [PD] encompassing 95% of the PTV / percentage of tissue outside the PTV encompassed by 95% ofPD). For VMAT the values are (92 ± 1% / 2.2 ± 0.9%), for IMRT (89.8 ± 1% / 0.7 ± 2.6%) and for 3D-RT (94.7 ± 0.7% / 5.4 ±1.5%). With values of 1.33 ± 0.21 and 1.39 ± 0.08, the CI’s for IMRT and VMAT are better than for 3D-RT with 1.71 ± 0.11. TheHIs for the prescribed dose (HI36) for 3D-RT and VMAT with values of 1.06 ± 0.01 and 1.11 ± 0.02, respectively, are slightlybetter than for IMRT with 1.15 ± 0.02. Dose to OAR for 3D-RT and VMAT were similar with the exception of the bladder (min-imal dose [99.0% of PD] to bladder: VMAT 21 Gy ± 9.3; 3D-RT 30 Gy ± 4.8). The TTT considers the beam-on-time (BOT) and thegantry movements. The mean TTT for 3D-RT is 220 seconds and for IMRT 557 seconds. The TTT for VMAT for one single arc isestimated with 90 seconds.

Conclusions: VMAT produces high quality treatment plans with excellent conformality and homogeneity. The main advantages ofthis new approach are the shorter treatment delivery time and a lower number of monitor units.

Author Disclosure: F. Stieler, None; D. Wolff, None; F. Lohr, None; V. Steil, None; Y. Abo-Madyan, None; F. Wenz, None; S.Mai, None

2837 Analysis of Respiratory Induced Liver Motion using Fiducial Marker in Stereotactic Radiosurgery for

Liver Tumor

H. Je, E. Choi, S. Ahn, S. Lee, S. Park, J. Kim

ASAN Medical Center, Seoul, Republic of Korea

Purpose/Objective(s): To validate image guided radiation therapy in stereotactic radiosurgery(SRS) for liver tumor by comparingmotions of fiducial markers on simulation computerized tomography (CT) with these on cone-beam CT.

Materials/Methods: Fourteen hepatocellular carcinoma and two hepatic metastasis patients who had received SRS in Asan med-ical center were enrolled for this study. Three gold markers were implanted near the tumor, and then 4-dimensional gated CT sim-ulation was done. Marker motions on maximal intensity projection (MIP) image were compared with motions on CBCT. Andmarker migration was validated with analysis of change in inter-marker distance.

Results: One standard deviation (SD) of the marker coordinates between two observer was less than 0.7 mm in end-expiratoryphase and 0.8 mm in end-inspiratory phase. And one SD of components of the motion vectors was 1.8 mm or less in cranio-caudal(C-C) direction. Respiratory induced motion of the markers in C-C direction was greater than other directions, that was 15.4�16.9mm and 15.2�17.5 mm on MIP and CBCT image respectively.

The only one marker showed discrepancy of motion on MIP and CBCT image, mean difference was 1.3 mm in C-C direction. Onthe other hand there was 1.5�3.3 mm mean difference in all markers, when the marker tracks on CBCT were compared with markercontours on MIP image.

Liver volume on simulation CT was 7.8% larger than volume on pre-treatment CT image, and the volume on last follow-up CTdecreased by 11.5% compared with simulation CT image. Change of liver volume in cirrhotic liver was larger than non-cirrhoticliver. All inter-marker distances decreased on last follow-up CT image compared with simulation CT image, the deduction wasfrom 2.0 ± 3.6 mm to 5.0 ± 2.5 mm. And besides deduction of inter-marker distance was larger in whom time intervals betweentwo CT scans were more than 2 months.

Conclusions: There was significant discrepancy in C-C direction between marker tracks on CBCT and marker contours on MIPimage in all markers. So marker tracks on MIP image rather than contours are more similar to marker tracks on CBCT in imageguide radiotherapy.

All three inter-marker distances decreased but there was no clarified conclusion whether marker is migrated or not, because livervolume also decreased in follow-up CT image. So it is necessary to investigate the relationship between marker position and changeof liver volume.

Author Disclosure: H. Je, None; E. Choi, None; S. Ahn, None; S. Lee, None; S. Park, None; J. Kim, None.

2838 Electronic Portal Cine Imaging of Implanted Marker Seeds during Respiratory Gated Treatment to

Evaluate Interfraction Movement for Stereotactic Body Radiotherapy of Liver Metastases

C. F. Serago, L. A. Vallow, R. Paz-Fumagalli, S. Kim, A. A. Gale, W. L. Magalhaes

Mayo Clinic, Jacksonville, FL

Purpose/Objective(s): To investigate the feasibility of the use of a respiratory-gated technique for stereotactic body radiotherapyof liver metastases with cine verification of implanted fiducial marker seeds during treatment.

Materials/Methods: Patients in this study were treated under a phase I dose finding pilot study of stereotactic body for the treat-ment of liver metastases. Patients had gold fiducial seeds implanted in or near the liver metastases. Each patient had a 4D-CT sim-ulation with respiratory motion data acquired simultaneously. The maximum displacements of the fiducial seeds during normalrespiration were determined from the 4D-CT. Prior to treatment, patients were positioned with the assistance of gated kV X-ray

S544 I. J. Radiation Oncology d Biology d Physics Volume 72, Number 1, Supplement, 2008

image guidance. Digital reconstructed radiographs (DRRs) were created for each static treatment field showing the expected po-sition of the fiducial seeds within the treatment field. The treatment with 6 MV photons was conducted with normal patient breath-ing and respiratory gating. During treatment, continuous cine images of the treatment field were acquired. The positions of thefiducial seeds on the cine images were compared to their expected DRR positions to evaluate positioning accuracy and interfractionmovement.

Results: For the 4 patients studied, the average maximum displacement of the fiducial seeds during the CT simulation withoutgating was 11.1 mm (range, 8.8 to 13.2 mm). During respiratory-gated treatment, the average maximum displacement of the fidu-cial seeds was 5.9 mm (range, 4.5 to 9.4 mm). The mean displacement of the fiducial seeds during treatment was 2.2 mm (standarddeviation ± 1.4 mm).

Conclusions: Cine imaging during treatment is an effective tool to evaluate the interfraction position of fiducial seeds for livermetastases. This preliminary study suggests that with respiratory gating the treatment margin to account for the interfraction move-ment of the liver may be reduced.

Author Disclosure: C.F. Serago, None; L.A. Vallow, None; R. Paz-Fumagalli, None; S. Kim, None; A.A. Gale, None; W.L. Mag-alhaes, None.

2839 Selection of Breathing Management versus Margins for Intrahepatic Targets

I. H. Lee, T. M. Williams, D. S. Tatro, M. Rosu, R. K. Ten Haken, J. Balter

University of Michigan, Ann Arbor, MI

Purpose/Objective(s): Attempts to account for breathing motion in patients undergoing conformal radiotherapy to intrahepatictargets include use of a margin (ITV), minimizing/suspending breathing motion, or gating the radiation treatment. Eliminationof margins for breathing has been estimated to reduce the effective volume (Veff) of irradiated liver by 5%. At our institution,we employ an active breathing control (ABC) device for any patient who can tolerate it; however it’s use increases treatmenttime and can lead to patient discomfort. Therefore, it is of interest to determine how the dosimetric benefit of ABC varies withpatient breathing amplitude and to thus identify which patients could be treated to equivalent clinical outcome without thedevice.

Materials/Methods: The effects of breathing motion and random setup error were modeled in the cranial-caudal direction byconvolving a planned dose distribution with a Gaussian kernel for setup variation and a previously described probability den-sity function derived from a simple breathing model. Representative cranial-caudal dose profiles were obtained via 3D-dosecalculations for opposing axial fields of different sizes. For each dose profile, we determined the width of the 95% isodoselevel (IDL) and the shift in the center of the 95% isodose surface. The impact of varying expansions on the effective volume(Veff) of normal liver was calculated for 5 previously treated patients using the DVH reduction method of Kutcher andBurman.

Results: Displacement of the center of the 95% IDL varies linearly with breathing motion amplitude and does not depend onfield size or setup uncertainty. The width of the 95% IDL decreases linearly as a function of increasing breathing amplitude andsetup error, and varies only slightly with field size and different parameters for the breathing model. For Veff, the size of thePTV and the relationship to Veff appear linearly related. In the absence of ABC, the required increase in the PTV margin iswell-approximated by a linear function of breathing amplitude and setup uncertainty, and relatively independent of breathingasymmetry. Expansions for random setup error and breathing motion can essentially be added linearly but require a PTV ex-pansion that is equal to the observed breathing amplitude. If random setup errors have a standard deviation of 0.5 cm, then inpatients with breathing amplitudes of 1 cm or less, it may be more worthwhile to improve on setup error before eliminatingbreathing motion.

Conclusions: For larger breathing amplitudes typical of patients undergoing liver radiation, it is important to take into account theVeff associated with the individual patient. Current data suggests that those with Veff between 35-60% are most likely to benefitfrom elimination of breathing motion during treatment.

Author Disclosure: I.H. Lee, None; T.M. Williams, None; D.S. Tatro, None; M. Rosu, None; R.K. Ten Haken, None; J. Balter,None.

2840 Impact of Localization Technique on the Accuracy of Daily Repositioning of Prostatic Isocenter during

Radiotherapy

T. He, D. G. Kaurin, J. Tanyi, J. Wu, W. D’Souza, M. Fuss, A. Hung

Oregon Health Sciences University, Portland, OR

Purpose/Objective(s): In prostate cancer radiotherapy, it is unknown how large the PTV margins must be to account for the iso-center correction tolerance and intrafraction motion. The risk of geographic miss can be minimized by the placement of fiducialmarkers in the prostate gland for daily pretreatment localization and adjustment of patient position if necessary. In this study,we assess the magnitude of interfraction isocenter displacement using three setup techniques: skin mark, bony landmarks, and elec-tromagnetic transponder fiducials.

Materials/Methods: Fifteen supine prostate IMRT patients with three implanted transponders each were studied. Initial daily lo-calization was based on three laser and skin marks. Daily localization error distribution was determined from offsets between theinitial setup position and that determined by Calypso electromagnetic transponder system. Post setup with the Calypso system,isocenter localization was immediately independently verified by acquiring a CBCT of the radio-opaque transponders using theVarian On-Board Imaging System (OBI). Both Calypso and CBCT localization techniques produced lateral, longitudinal, and ver-tical target offsets from machine isocenter. Organ motion or patient movement during treatment was continuously monitored by theCalypso system at a 4-mm threshold. A retrospective analysis was performed CBCT imagesets to determine the accuracy of CBCT-based bony alignment. The CBCT dataset and the reference planning CT were imported into Corvus 6.3 (NOMOS) treatment