ESTRO 37 Abstract book

S1212

ESTRO 37

2 Royal Marsden NHS Trust, Clinical Oncology, London, United Kingdom Purpose or Objective Elements (Brainlab) is a new TPS with an automatic planning algorithm generating plans for the treatment of up to 15 brain metastases using a conventional linac with a single isocenter and non-coplanar dynamic conformal arcs. We performed a planning study comparing the quality and delivery of plans generated by Multiplan (Accuray) for robotic radiosurgery treatment on Cyberknife, and by Elements for treatment on a linac coupled with the robotic image-guidance system Exactrac. Material and Methods Fourteen patients with multiple brain metastases were previously treated in our institution using Cyberknife. Radiosurgery treatments originally planned with the Multiplan package were replanned using the Multiple Metastases module, within the Elements package, for delivery on a linac with 5mm MLC. Either 4 or 5 couch angles were permitted; 1 or 2 half arcs could be delivered at each couch angle. Plans were assessed according to target coverage and ensuring that the local dose constraints to organs at risk were met. Results Between 2 and 12 lesions (median of 6 lesions) were treated, the volumes ranged from 0.015cc to 6.6cc. The prescription dose was 20Gy or 24Gy depending on lesion volume, unless a lower dose had been used at the time of the Cyberknife radiosurgery treatment. All plans were delivered in a single fraction. The mean coverage by the prescription isodose was 98% of the GTV volume for plans generated on both Multiplan and Elements. The dose constraints to the local normal brain (V12Gy≤10cc) and to the brainstem (Dmax≤12.5Gy) were met by all plans delivered on Cyberknife, and by all but one plan obtained with Elements: for this plan the local V12Gy was 11.8cc near the largest GTV of the 14 cases (V = 6.6cc). The local V12Gy correlated with the volume of the lesions (fig.1), and the total normal brain V12Gy was slightly higher for linac treatments (mean of 9.0cc) compared to Cyberknife treatments (6.6cc).

conformal than the Cyberknife treatments due to a combination of the 5mm MLC and more limited range of beam directions available on the linac. Treatments were quicker to plan and to deliver using the Elements linac system compared to equivalent Cyberknife treatments. These improvements in treatment planning and delivery times have implications for radiotherapy departmental efficiency and for treatment tolerability for patients with multiple brain metastases. EP-2189 Action levels of robotic small animal irradiator performance derived from 4 years of QA measurements P. Brodin 1 , W. Koba 2 , C. Guha 1 , W. Tomé 1 1 Montefiore Medical Center/Albert Einstein College of Medicine, Radiation Oncology, Bronx, USA 2 Montefiore Medical Center/Albert Einstein College of Medicine, Department of Radiology Nuclear Medicine, Bronx, USA Purpose or Objective Modern small animal irradiators are capable of delivering precision radiation therapy through image-guidance and accurate dosimetry, with the goal of improving the translation of research findings into clinical practice. This places a requirement on performance similar to that of clinical linear accelerators. Here we present action levels for the performance of robotic image-guided small animal irradiators, derived from our long-term experience with a monthly quality management program (QMP). Material and Methods The QMP was developed in-house and employed monthly from August 2013 to September 2017 to assess dose output consistency (derived from AAPM TG-61 protocol guidelines), image quality, image-guided target localization (imaging vs. radiation isocenter agreement) and dose calculation consistency. Here we focus on the dose output and imaging vs. radiation isocenter measurements. Dose output was measured based on temperature- and pressure-corrected charge collection using a parallel-plate ion chamber placed at isocenter distance. The imaging vs. radiation isocenter agreement was measured by localizing the center of a radio-opaque 2mm diameter ball bearing (BB) on cone-beam computed tomography (CBCT) imaging. The robotic stage was then moved placing the BB at the imaging isocenter. Images were then acquired with the gantry at 90° and 0°, respectively, with a collimator insert producing a 5x5mm 2 irradiation field. Images were analyzed using a semi- automatic MATLAB script determining the center-of-mass of the BB in relation to the 5x5mm 2 radiation field. Perfect imaging/radiation isocenter agreement would be represented by the BB perfectly centered within the 5x5mm 2 square. Results The 4-year results of dose output consistency, reported as the dose rate at D max , and imaging vs. radiation isocenter agreement are shown in the figure below.

The use of Elements considerably decreased both the planning time (several hours with Multiplan for an experienced planner compared to a few minutes with Elements) and the mean treatment delivery time (1h 40min on Cyberknife compared to 40min on the linac, which includes imaging at all couch angles with For the planning of multiple brain lesions, the Brainlab Elements TPS generated treatments that were clinically acceptable for lesions smaller than 6.6cc. For larger lesions, the linac based Elements treatments were less ExacTrac). Conclusion