ESTRO 2021 Abstract Book

S612

ESTRO 2021

pelvic lymph nodes with the Cybeknife system. Materials and Methods

Daily CBCT images from 16 pelvic lymph node (75 images) and 14 para-aortic lymph node (59 images) SBRT linac-based treatments were employed in this analysis. Patients were immobilized during CT simulation and treatment with an alpha cradle Daily corrections on each treatment fraction based on node positions were compared with displacements proposed by the closest vertebral bodies, as it would have been done by the XSight Spine tracking mode employed in Cyberknife. Mean (μ), Systematic (Σ) and random (σ) residual errors were obtained for each treatment location. Margins needed to obtain a correct coverage of the patient population were calculated with Van Herk´s formula assuming an isodose level of 85% and a penumbra σ p =2mm. Results Population based results and treatment margins for both locations are shown in Table 1, while histograms of residual errors are shown in Figure 1. Mean error (μ) was negligible for para-aortic lymph nodes, while values of Σ and σ below 1mm allowed a population coverage over 90% to be obtained with applied margins of around 2mm.

On the other hand, a mean displacement of 2mm was obtained in the AP direction for pelvic lymph nodes, probably due to pelvic rotations with respect to vertebral bodies. In case these displacement could be corrected, margins of 4mm and 5mm would still be needed in LAT and AP directions, while only a margin of 2mm would be necessary in the SI direction.

Conclusion Fiducial-less of para-aortic lymph node treatment by means of spine-based tracking is feasible given the residual errors that are present. Pelvic lymph nodes present larger uncertainties and systematic errors that would recommend the use of fiducial markers to guide the treatment. Proper immobilization of pelvis with respect to spine is key to allow for accurate fiducial-less treatment of pelvic lymph nodes with Cyberknife. PD-0781 Validation of optical surface/thermal imaging and X-ray positioning accuracy for SRS treatments V. da Silva Mendes 1 , M. Reiner 1 , S. Corradini 1 , M. Niyazi 1 , C. Belka 1 , G. Landry 1 , P. Freislederer 1 1 University Hospital, LMU Munich, Department of Radiation Oncology, Munich, Germany Purpose or Objective The novel Exactrac Dynamic (Brainlab AG, Germany) provides real-time 3D surface imaging by combining structured light with thermal information, and is complemented by an in-room kV X-ray imaging system. The aim of the study was to compare the positioning accuracy of the combined surface/thermal imaging system and the gold-standard, stereoscopic X-rays, for stereotactic radiosurgery (SRS) treatments. In order to provide a more realistic scenario, a head phantom prototype with a specific heat signature profile was used. Phantom specifications regarding surface temperature stability were also investigated. Materials and Methods An anthropomorphic 3D-printed head phantom (Prime, RTsafe, Greece) with bone equivalent material, 3 embedded ball bearings and can be filled with water up to 45°C was used. It was fixed to the table using a 4Pi open face mask (Brainlab AG, Germany). To investigate the phantom’s surface temperature stability, the phantom was filled with warm water (41°C) and surface temperature was measured with an infrared thermometer at 7 locations within the area of the face opening, app. every 10 min, for 65 min. The surface/thermal imaging positioning (D S+T ) was measured and compared to X-ray based positioning (D X-RAY ) at an Elekta Versa HD linac. The couch (at 0°) was displaced in all 3 directions (lateral, longitudinal and vertical) from the planned position (10 random translations, max. translation 2.5 mm), using the HexaPOD TM evo RT