ESTRO 2020 Abstract Book

S875 ESTRO 2020

1 Hospital de la Santa Creu i Sant Pau, Radiofísica y Protección Radiológica, Barcelona, Spain Purpose or Objective The daily position of the shoulders during treatment might have an impact in cases where, in order to deliver the prescribed dose to the PTV, the beam could pass through them. During the treatment course of a Head and Neck (HN) patient, we co-registered a CBCT/CT focusing on the relevant PTV area. Thus, the daily position of the shoulders may differ from the position in the CT plan (Fig.1).

The main component is a sensitive respiratory and motion compensation movement, which was programmed based on the Robot API from KUKA. The desired contact force was set to 5 N. The movement was tested on subjects for different sound positions and breathing patterns. MITK has been extended with filters for the spatial assignment of coordinates and the display of ultrasound images, ea. for creating 3D ultrasound volume by means of scanning movements of a 2D ultrasound head. Results Image acquisition was continuous due to respiratory motion compensation and the transducer could be repositioned and positioned by remote control. The applied and resulting measured force at the transducer was constant. The control concept and image representation are combined in a GUI, whereby the developed filter pipeline allows the simultaneous display of registered ultrasound images and planning data in real time (Fig. 2).

In our institution, all patients are treated with a VMAT technique. Optimization is performed using an avoidance strcuture, enclosing the shoulders, to decreased beam shoulder penetration. This work investigates the use of PerFRACTION TM (a 3D in vivo dosimetry software) as a tool to asses our immobilization technique as well as our VMAT planning strategy, and to evaluate the dosimetric impact caused in the PTV by the shoulder positioning. Material and Methods We collected all the images performed for 13 HN patients during their course of treatment. Only sessions having a CBCT and an EPID-integrated image were selected. In our institutions, HN patients are immobilized with a 3-points mask while grabbing special straps at a fixed position that help them push their shoulders down. To assess the daily variation of the shoulder position, we edited manually the co-registered CBCT/CT focusing on the shoulder area and measuring shoulder displacement. The effect of shoulder variation was also evaluated with PerFRACTION TM using a 3%-3mm gamma analysis between the expected and delivered dose fluences images for each treatment arc. To account for the dosimetric impact, we determined PTV daily doses with the reconstructed 3D dose distribution calculated by PerFRACTION TM in the co- registered CBCT/CT. We extracted the PTV mean dose and V95 as the dosimetric estimators of the whole 3D dose distribution. Results We examine a total of 13 patients and 30 sessions with co- registered CBCT/CT and EPID-integrated images. The average shoulder shift was 4.2 mm. 63.3% of the observed shifts (19 out of 30) were less than 5 mm (range between 0.0 mm and 17.4 mm). Per-Fraction detected 4 sessions with a poor 3%-3mm gamma performance: 92.3%, 87.41%, 89.26% and 95.9%. The rest of the sessions performed with gamma passing rates higher than 98%.

Conclusion A reliable, universal method for motion compensation, under low force constraints, in robot-guided ultrasound imaging has been developed and successfully tested. This is a first step towards a safe, autonomous robot positioning on the patient. In addition, the positional image data could be integrated into a system, providing the basis for image registration for ultrasound-based radiotherapy. [1] Nolden, M., e. al.: The medical imaging interaction toolkit: challenges and advances. International Journal of Computer Assisted Radiology and Surgery 8(4), 607–620 (2013). DOI 10.1007/s11548-013-0840-8 PO-1608 Assessment of shoulder position variation and its impact on VMAT doses using an EPID-based software I . Valverde Pascual 1 , P. Gallego 1 , J. Perez-Alija 1 , N. Jornet 1 , A. Latorre-Musoll 1 , M. Barcelo-Pages 1 , N. Garcia- Apellaniz 1 , P. Carrasco de Fez 1 , M. Lizondo 1 , P. Simon Garcia 1 , P. Delgado-Tapia 1 , M. Adria Mora 1 , A. Ruiz Martinez 1 , M. Ribas 1