S 566
SRS-SBRT
1519 poster A PLAN ROBUST ANALYSIS FOR AVM TREATMENTS USING A TOMOTHERAPY UNIT G. Guidi1 , E. Cenacchi1 , F. Bertoni2 , T. Costi1 1 A Z .O SPEDALIERO -U NIVERSITARIA DI M ODENA, Medical Physics, Modena, Italy 2 A Z .O SPEDALIERO -U NIVERSITARIA DI M ODENA, Radiation Oncology, Modena, Italy Purpose: The aim is to verify the feasibility of the AVM (Arteriovenous Malformation) radiosurgery using a Tomotherapy unit without a head invasive frame. Using a plan robust analysis, we have evaluated few cases, to compare the lower, mean and upper limit of doses delivered. We have simulated an helical treatment and a potential setup error due to the immobilization device. We have evaluate the risks of dose missing in case of contours using MRI (AVM) and orthogonal images (AngioROI) acquired by the Angio-Rx exams. Materials: Previous AVMs treated using a LINAC with microMLC and head invasive frame, are re-planned using a Tomotherapy unit. 20Gy are prescribed to 95% of the AVMs volume and 18Gy is the minimum accepted. The AVMs volume are contoured using the co-registration of the MRI and CT exams. The AngioROI are contoured using planar images obtained by an Angio-Rx exams and superimposed to the CT projection. Using CERR (WUSTL University) and a plan robust analysis, we have simulated the stability of the treatment assuming a mean setup error (-1.2mm,-0.1mm,1.7mm) due to the non use of invasive head frame. The mean setup error is been calculated by previous work of the author based on 21 radiosurgery recorded using a MVCT co-registration and delivered using a Tomotherapy Unit and a thermoplastic mask as immobilization devices. Results: The contoured volume are similar and no difference are evident as reported into the conclusion. The table show the results of the plan robust analysis using a single fraction and 1000 trials of simulation.
The data of the robust analysis show discrepancies only for the min dose, but no incongruity are highlighted for upper, mean and lower bound. Considering the 18Gy as a minimum dose request, all the volume are covered adequately. Conclusions: The AngioROIs contoured using orthogonal planar images are smaller than the AVMs contoured using the MRI images but equal considering the resolution of the systems used to calculate and delivery the doses. Only few voxel of discrepancy (<0.5cc) are highlighted by the different method of contouring and are smaller than the grid dose dimension (2.15x2.15x2mm). Anyway the helical delivery can guarantee to delivery the dose prescription and the minimum dose of 18Gy to both the volume without geographic missing, even considering possible setup error. The robust analysis has showed the possibility to treat the patients without the use of invasive head frame and increase the comfort of the patient, maintaining accuracy and safety procedures 1520 poster COMPARATIVE DOSE-VOLUME ANALYSIS FOR FRACTIONATED RADIOTHERAPY OF CENTRAL BRAIN TUMORS IN PEDIATRIC PATIENTS USING CYBERKINFE, IMRT, AND BRAINLAB J. Heinzerling1 , C. Timmerman1 , Z. Hong2 , E. Ramirez1 , C. Ding1 , R. Foster1 , T. Solberg1 , R. Timmerman1 1 U NIVERSITY OF T EXAS S OUTHWESTERN M EDICAL C ENTER, Radiation Oncology - Cancer Care Pavillion, Dallas, USA 2 U NIVERSITY OF T EXAS S OUTHWESTERN M EDICAL C ENTER, Biostatistics, Dallas, USA Purpose: Pediatric patients (pts) with low grade central brain tumors are typically treated with fractionated IMRT (FIMRT). The purpose of this study was to identify potential dosimetric advantages in target coverage and organ at risk (OAR) dose in the setting of image-guided, fractionated stereotactic radiosurgery (FSRS) with reduced margins compared to FIMRT. Materials: 10 pts were selected with central brain tumors previously treated to a dose of 45-50.4 Gy in 25-28 fractions. Pts were re-contoured to include unique organs at risk. GTV and OAR volumes were delineated utilizing MRICT fusion. Three methods of treatment planning were carried out for each pt: for FSRS, pts were planned utilizing both Cyberknife Multiplan (CK) and BrainLab IMRT with iPlan (BL), and were compared to FIMRT planned with Pinnacle. The GTV was expanded by 5 mm for FIMRT planning, but because of the availability of daily image guidance, no PTV margin was added for either CK or BL planning (i.e. GTV=PTV) in order to simulate actual treatment
conditions. FIMRT plans were generated using 6 non-coplanar beams with Pinnacle version 8.0m. Inverse planning was used and plans were optimized to achieve optimal PTV coverage while minimizing OAR doses to at least those consistent with institutional dose constraints, and plans were typically prescribed to 95-100%. For CK, one to two collimators were used to provide highly conformal dose distributions, and plans were typically prescribed to the 50-70% isodose line. For BL, 7 noncoplanar beams were generated and plans were typically prescribed to 95-100%. All plans were prescribed to a dose of 45 Gy in 25 fractions. Dose volume analysis for various endpoints was performed and comparative analysis was done using SAS (Cary, NC). Results: Significant improvement was observed for FSRS compared to FIMRT for the following OAR parameters: mean dose, minimum dose, and dose to 2 cc of brainstem; mean dose and dose to 2 cc of caudate; max, mean, minimum, and dose to 5 cc of cerebellum; mean dose, dose to 10 cc of cerebrum, and volume of cerebrum receiving 12 Gy; minimum dose to the optic chiasm; max dose, mean dose, and dose to 0.5 cc of the hippocampi, max, mean, and minimum dose to the bilateral optic nerves and tracts; minimum dose to the pineal gland, mean and minimum dose to the pituitary gland, mean dose and dose to 2 cc of putamen; mean dose and dose to 2cc of thalamus. For the two FSRS planning methods, CK resulted in significant reduction in dose to 10 cc of cerebrum and volume of cerebrum receiving 12 Gy. Conclusions: FSRS with image guidance allows reduction of margins and results not only in reduction of dose to adjacent OAR, but also to novel structures away from the target compared to FIMRT planning for treatment of central brain tumors in pediatric patients. Less OAR dose may translate into decreases in late toxicity in pediatric patients and will be evaluated in a prospective clinical trial at our institution. 1521 poster COMPARISON OF CONVENTIONAL VS. SOFTWARE-BASED METHODS FOR RAPID-ARC BASED SRS QA A. Etemadnia1 , H. Yampolsky2 1 S AINT J OHN ‘ S H OSPITAL, Physics, Santa Monica, USA 2 VALLEY R ADIOTHERAPY A SSOCIATES M EDICAL G ROUP, I NC., Center for Radiation Therapy of Beverly Hills, El Segundo, California, USA Purpose: Rapid Arc is a volumetric modulated are therapy delivery system which delivers dose through upto 360 degree rotation using 177 control points per arc with five degrees of freedom. These are modulations of dose rate, gantry angles, field sizes, gantry speed & the multi-leaf collimator. As a result, it presents a unique set of challenges when it comes to performing standard quality assurances (QA). This task spans further in the case of stereotactic radiosurgery (SRS) using the rapid arc delivery method due to higher dose per fraction, steep dose gradients & smaller field sizes. From QA perspective, this translates into methods & equipment that could ensure adequate QA which is critical to patient safety. In this study, we have evaluated methods of RapidArc SRS QA performing Absolute Dosimetry and Relative Dosimetry. Materials: We have performed absolute measurements using Standard Imaging‘s Exradin A16 thimble chamber with a collective volume of 0.007 cc. The relative dosimetry has been performed using Kodak EDR2 film in conjunction with the RIT software. These results have been compared to a portal dosimetry based collection method using the Dosimetry Check (DC) software. QA was performed on 18 treatment plans using chamber & film. The same plans were delivered & analyzed using the DC software. Dose & gamma index (3mm/3%) deviation from the expected values were evaluated for each method. Results: The absolute dosimetry method using the chamber resulted in an average difference of 1.2% with a SD of 2.1% vs. the absolute dosimetry method using the DC with an average of -0.2% difference with a SD of 2.8% from the predicted values. The relative dosimetry method using the film resulted in an average of -0.8% difference with a SD of 1.8% vs. the relative dosimetry method using the DC with an average of 0.9% difference & a SD of 1.0% from the predicted values. Conclusions: DC successfully performs QA demonstrating comparable results when compared to conventional methods of QA. The advantage mainly lies in logistic factors such as single exposure absolute & relative results as well as systematic advantages such as detector dimensions (averaging effects) & their spread (detectors/mm), film saturation, processing complexities, cable noises & performing manual shifts on multiple targets with single isocentre treatments which would not be of concern using DC. In addition, DC reconstructs dose on original CTs as well as a DVH comparison of organs contoured using the same measured data. 1522 poster COMPREHENSIVE VERIFICATION OF THE LINAC ISOCENTRE FOR STEREOTACTIC RADIOSURGERY USING CINE-EPID AND ARC