43 Reproducibility of Organ Position with Respiratory Gating for Liver Tumors: Use in Dose-Escalation R. Wagman, E. Yorke, P. Giraud, E. Ford, K. Sidhu, G. Mageras, B. Minsky, K. Rosenzweig Radiation Oncology, MSKCC, New York, NY Purpose: To evaluate the clinical utility, accuracy, and benefit of a respiratory gating system for the treatment of liver tumors. Materials and Methods: Six patients with liver tumors were selected for evaluation of the Varian Real-Time Position Monitor (RPM) respiratory gating system. The RPM system is a passive system that uses an infrared-sensitive camera to track the motion of reflective markers mounted on the abdomen. When the marker positions fall within a user-selected gate, the system permits acquisition of breathing synchronized CT images. In treatment mode, the system allows the linear accelerator to “beam-on”. At simulation, a fluoroscopic movie and breathing trace were acquired in treatment position using the RPM system. The gate-width was set around end-expiration. Each patient underwent two end-expiration-synchronized CT scans for treatment planning and to assess gating system reproducibility. Two patients also underwent an end-inspiration-synchronized scan. On each CT, a physician contoured liver, kidneys, tumor, spleen, and diaphragms. Fixed bony anatomy was used to register different scan sets. Each organ’s positional change between two scan sets was quantified by calculation of: 1) the shift in the center of mass and, 2) the ratio of the volume common to the two versions of the organ to the volume included in at least one (intersection/union). Recorded instructions were used at all sessions in order to regularize breathing patterns. Normal tissue complication probability (NTCP) calculations and dose-volume histograms (DVH) assisted in determining the treatment dose. During treatment, extra portal images were obtained to monitor gated treatment reproducibility and assess daily variation in diaphragm position. Results: One hundred and three treatments were performed in 5 patients using RPM gating and static conformal beams; all patients tolerated treatment well. Initial fluoroscopic data showed that average superior-to-inferior (SI) diaphragm motion was reduced from 22.7 mm without gating to 4.8 mm with gating. Comparing the end-inspiration to end-expiration (E-I) CT scans, the average SI movement of right and left diaphragm was 11.5 mm and 21.4 mm, respectively. In contrast, the average SI movement of right and left diaphragm between the two end-expiration (E-E) scans was 2.2 mm and 3.8 mm, respectively. Improvement in reproducibility is also shown by center of mass shifts, which are smaller between E-E than E-I scans; for all organs, average E-I SI motion was 12.8 mm (range: 8.7-19 mm). In contrast, average E-E SI organ motion was 2.0 mm (range: 1.6-2.5 mm). Intersection/union volume ratios were closer to 1.0 for E-E than for E-I scans, indicating reproducibility of gating. The average ratio of E-I studies was 0.62 (range: 0.48-0.74). In contrast, the average ratio of E-E studies was 0.83 (range: 0.77-0.87). Portal images confirmed reproducibility of gating technique during treatment. The patient-averaged change in SI position of the diaphragm apex from AP DRR to AP portal image was 3.9 mm (range: 0.8-6 mm). Treatment time was prolonged by less than ten minutes with the use of gating. The reproducible decrease in organ motion with gating enabled the standard 2 cm margin of expansion from GTV to PTV to be halved (1 cm with RPM). Margin reduction allowed dose increases of 7-27% (median 21.3%) with stable liver NTCP values. Toxicities included grade 2 nausea (1 patient) and grade 3 thrombocytopenia (1 patient). No radiation-induced liver damage was seen (median follow-up 5 months). Conclusion: The Varian RPM system allows accurate, reproducible and clinically efficient treatment of liver tumors at end-expiration, making it possible to safely reduce margins on tumor volume. Target dose can thus be increased beyond that allowed by conventional radiation treatments. 44 Application of Real-Time Tracking Radiation Therapy (RTRT) System for the Treatment of Spinal and Paraspinal Diseases A. Yonesaka 1 , R. Onimaru 1 , H. Shirato 1 , K. Kitamura 1 , H. Aoyama 1 , T. Seki 2 , K. Hida 2 , Y. Iwasaki 2 , K. Miyasaka 1 1 Department of Radiology, Hokkaido University, School of Medicine, Sapporo, Japan, 2 Department of Neurosurgery, Hokkaido University, School of Medicine, Sapporo, Japan Purpose: Fluoroscopic real-time radiation therapy (RTRT) has been shown to be useful to reduce set-up error considerably in the treatment of extracranial diseases. We have used this system for stereotactic irradiation of spinal and paraspinal lesions. Materials and Methods: The subjects were two patients with eccentric intramedullary arteriovenous malformation (AVM) located at the cervical and thoracic spinal cord and three patients with spinal schwannoma located at the paraspinal region at a minimum distance of 1 to 5 mm from the spinal cord, treated in 1999 and 2000. Patients with AVM were treated using one internal marker, and patients with schwannoma were treated recently using 3 markers. Internal fiducial markers were inserted into paraspinal muscle near the lesion under local anesthesia. During each treatment day, translational set-up error was corrected by adjusting the actual position of one of the internal fiducial markers, as detected by two fluoroscopies in the treatment room, to its planned position, which was transferred from the three-dimensional radiation treatment planning (3DRTP) system. The rotational set-up error was calculated using two additional markers in the 3 patients with schwannoma. All patients were treated with 3D conformal narrow beam radiotherapy. The fraction size (Gy) and number (F) was 22Gy/4F and 20Gy/4F for AVM, 35Gy/8F for 1 schwannoma, and 50Gy/25F for 2 schwannomas. A phantom was used to assess the accuracy of the RTRT system for the calculation of rotational set-up error. Results: In the phantom study, the translational set-up error was detected with an accuracy of 0.2  0.3 mm. The rotation of the phantom was detected with a median accuracy of -0.4 (range: -2.1 to 0.8 degree), 1.0 (-1.2 to 1.4), and -0.4 (-0.4 to 2.4) degrees around the lateral axis, cranio-caudal axis, and ventro-dorsal axis, respectively. The mean  standard deviation (SD) in dislocation after manual set-up in 5 patients was 0.1  5.0mm, 2.8  6.0mm, and -0.1  3.8mm in the lateral, cranio-caudal, and ventro-dorsal directions, respectively. The dislocation after RTRT set-up was 0.4  0.8 mm, -0.1  0.7 mm, and 0.2  0.8 mm in the three directions, respectively. There was a significant reduction in translational set-up error as a result of using RTRT set-up (p0001, Mann-Whitney’s U test). Rotational set-up error after manual set-up was -2  5 degrees, 3  5 degrees, and 6  8 degrees around each axis, respectively. The rotational set-up error did not change as a result of the translational RTRT set-up, as expected. The dose to a part of the spinal cord in the same transaxial plane was reduced to 50% of the prescribed dose to the AVM. The doses to the edge of the spinal cord were 2.5Gy/5F (5%), 28Gy/8F (80%) and 35Gy/25F (70%), and 28 I. J. Radiation Oncology ● Biology ● Physics Volume 51, Number 3, Supplement 1, 2001