S440 I. J. Radiation Oncology d Biology d Physics Volume 72, Number 1, Supplement, 2008

imaging with a patient-specific threshold; and 4) imaging every other day, with non-imaged fractions corrected using the mean shiftof the first five imaged fractions. For each imaging scenario, the overall mean (M), systematic setup error (S), random setup error(s) and population margins were calculated, assuming daily CBCT to be the gold standard.

Results: The systematic setup error (S) was reduced with increasing frequency of imaging: from 3.4 mm for no imaging down to1.0 mm for imaging every other day. The random setup error (s), however, varied little regardless of the frequency of imaging. Therandom error calculated for the three-dimensional vector was 2.9 mm, 3.0 mm, 3.4 mm, and 3.2 mm for no imaging, weekly im-aging, first five then weekly imaging, and every other day imaging, respectively. Because the random error was relatively unaf-fected by any imaging frequency scenario, the residual error remained substantial. Even when imaging was performed everyother day, residual errors of greater than 5 mm occurred in 24% of all fractions. Population margins were calculated to be approx-imately 1-1.6 cm for no imaging, 4-6 mm for weekly imaging, 4-6 mm for the first five then weekly imaging, and 4-5 mm for im-aging every other day. Since the residual error of daily CBCT was not included in this analysis, these margins would compare toa margin of zero for daily CBCT.

Conclusions: Daily image guidance is ideal in conventionally fractionated radiotherapy for lung tumors as the setup margin can bereduced by approximately 5mm versus a non-daily imaging scenario. However, if daily image guidance is not possible, then there islittle benefit in imaging more often than once a week.

Author Disclosure: K. Olivier, None; J. Li, None; W. Shi, None; H. Newlin, None; A. Chvetsov, None; C. Liu, None; J. Palta,None; A.R. Yeung, None.

2606 Four-dimensional CT (4D-CT) Image Guidance for Stereotactic Body Radiation Therapy (SBRT)

M. Fuss1, A. Srisuthep2, J. Kalpathy-Cramer1, W. D’Souza1

1Oregon Health & Science University, Portland, OR, 2Bhumibol Adulyadej Hospital, Bangkok, Thailand

Purpose/Objective(s): To investigate changes in target volume (gross tumor volume, GTV; planning target volume, PTV), res-piration target motion trajectory and motion envelope over 3 to 5 fraction SBRT for primary lung and liver cancer.

Materials/Methods: Seventeen patients underwent 4D-CT simulation, and repeat 4D-CT imaging for SBRT image-guidance (64re-4D-CT; 1-5 re-4D-CT/patient, median 5). Target volumes were individually delineated in the following CT data sets: free-breathing (FB), 10 respiratory phases (0-90%), maximum intensity projection (MIP), minimum intensity projection (MinIP),and average intensity projection (AveIP). We analyzed the pattern of target motion and motion changes over time. We also ana-lyzed target volume changes over time, and assessed if typical SBRT planning target volumes (PTVFB: GTVFB + 5 mm axial, 10mm cranio-caudal; PTVMIP: ITVMIP + 5 mm) would encompass the motion envelope during the SBRT course.

Results: GTVs were stable (+/� 20%) over a course of SBRT in 13 patients; 2 tumors did shrink by up to 39%, while 2 tumorsincreased in size by up to 31%. Shape of the motion trajectory was characteristic for individual patients over the SBRT course in 15/17 patients, while maximum respiration target centroid displacement was more variable. Motion trajectories approximated a motionplane in 13/17 patients; 2 cases studied showed a complex three-dimensional motion profile, while 2 cases demonstrated somewhaterratic motion. Inter-individually, motion trajectories varied widely. The simulation derived PTVMIP and PTVFB were larger thanrespective volume calculations in control 4D-CT studies in 69, and 68.4%, respectively, indicating that the target motion envelopeat simulation was larger than during the SBRT course. However, in 21.4, and 29.9% of controls, the PTVFB and PTVMIP derivedfrom simulation 4D-CT studies were smaller than control CT derived PTVs by up to 75%, with individually associated loss ofmotion envelope coverage. The respective motion envelope coverage by the clinically utilized PTVMIP was better than 90%. Whilethe PTVFB provided similar coverage, the respective volumes differed from PTVMIP by up to 70% (mean 106%).

Conclusions: In this preliminary analysis, 4D-CT image-guidance assessed changes in target motion over time were observed inthe majority of patients. Individually, significant changes in respiratory target motion were observed that prompted adaptive re-planning to minimize the risk of target miss (n = 2) or allowed for significant normal tissue sparing (n = 2). Changes in targetmotion were randomly observed early and late in the SBRT course, and were individually not predictive of motion during the fol-lowing fraction.

Author Disclosure: M. Fuss, Philips Medical Systems, GE Healthcare, C. Other Research Support; Philips Medical Systems, GEHealthcare, D. Speakers Bureau/Honoraria; GE Healthcare, F. Consultant/Advisory Board; A. Srisuthep, None; J. Kalpathy-Cramer, None; W. D’Souza, Philips Medical Systems, C. Other Research Support.

2607 Results of a Prospective Study on Brachytherapy in Lung Cancer in Combination with other Modalities

R. Sur1, B. Donde2, S. Puksa1, A. McLellan1, T. Farrell1, J. Hayward1, R. Hunter1, T. Corbett1, G. Okawara1, J. Wright1

1McMaster University, Hamilton, ON, Canada, 2University of the Witwatersrand, Johannesburg, South Africa

Purpose/Objective(s): High Dose Rate Intraluminal brachytherapy (HDRILBT) is an effective form of treatment for lung cancerpatients with intraluminal disease causing hemoptysis, shortness of breath and cough. This report describes our experience withHDRILBT in 98 patients who have completed one year follow-up.

Materials/Methods: There were 98 patients with non-small cell lung cancer who were treated with HDRILBT and had completed1-year follow-up were included in this analysis (70- untreated; 28 previous treatment). Following treatment, patients were assessedusing the EORTC QOL questionnaires for hemoptysis, shortness of breath and cough regularly.

Results: The Mean Hemoptysis Free survival (HFS) for all patients was 232.29 days; mean Cough Free survival (CFS) 140.32days, and mean Dyspnea free survival (DFS) 173.45 days. There was no impact on any of the 10 prognostic markers (gender,age, smoking history, performance status, presence of co-morbidities, tumor site, brachytherapy dose, additional treatment postbrachytherapy, previous treatment and treatment intent) on HFS, CFS and DFS on univariate and multivariate analysis. Additionalexternal radiation, chemotherapy or both did not improve HFS, CFS and DFS. Previous treatment prior to HDRILBT also did notimpact QOL (p . 0.05). Overall survival (OS) for the whole group was 13.43% at 12 months (mean- 192 days). When the impactof the 10 prognostic markers were analyzed, on multivariate analysis, performance status (p = 0.0407), presence of cardiac