Abstract Book

ESTRO 37

S561

PO-1006 Geometric acceptance testing for a hybrid MRI-Linac system J.W.H. Wolthaus 1 , J.J. Bluemink 1 , S.J. Woodings 1 , J.H.W. Vries- de 1 , T. Soest- van 1 , J.G.M. Kok 1 , H.M. Zijp- van 1 , S.L. Hackett 1 , B. Asselen- van 1 , B.W. Raaymakers 1 1 UMC Utrecht, Department of Radiation Oncology, Utrecht, The Netherlands Purpose or Objective In our department an Elekta Unity MRI-Linac, combining a 7 MV FFF linac with a 1.5T diagnostic MRI, was installed. Machine acceptance and commissioning comprise geometrical beam (alignment) verification, reference dosimetry, beam characterisation and MR imaging. We have followed international guidelines on reference dosimetry and on linac QA to comply with clinical standards. However, specific tests have been adapted to incorporate the magnetic field effects on delivered dose and measurement. The presented work will focus on QA of beam alignment. Material and Methods No light field and laser system are available on the MRI- Linac for alignment. Therefore, we use the on-board EPID, rigidly mounted on the gantry opposite to the beam generating system, for alignment of the measurement equipment (e.g. instruments, detectors, watertank, film). On the EPID images the isocenter pixel is identified from multiple projection images of a ball bearing. Since no collimator rotation is possible, beam centres cannot be derived directly from field edges. Therefore, beam alignment has to be determined fr om multiple tests. (1) Beam orthogonality is verifie d using the alignment of profiles at different depths measured in a watertank. (2) Focal spot alignment in the lateral direction is determined using two longitudinal opposing half fields with opposing gantry angles irradiated onto a film (fig 1a). The film is placed between copper plates to mitigate the effect of electron deflection due to the Lorenz force. The difference in centre position (defined by the edges) of the two profiles extracted through each of the two half fields is related to the misalignment of the focal spot. (3) Alignment of the MLC bank in the lateral direction is determined by comparing the profile peak position of a 30 x 20 cm 2 field to the centre defined by the field edges. (4) MLC leaf position accuracy is determined using a stripe/picket- fence-test in which the film is registered to simultaneously acquired EPID images to define the isocenter in the film (see fig 2). (5) Gantry isocenter accuracy is determined using a spoke-test on a film, sandwiched between copper-rings (fig 1b). (6) A general check of the beam alignment and field-size is performed by evaluating the congruence of two opposed fields of different sizes for 0/180° and 90/270° (fig 1c).

Netherlands 2 Academic Medical Center, Radiation Oncology, Amsterdam, The Netherlands 3 Catharina Hospital Eindhoven, Radiation Oncology, Eindhoven, The Netherlands 4 Radiotherapygroup - treatment location Deventer, Radiation Oncology, Deventer, The Netherlands Purpose or Objective To obtain optimal radiotherapy treatment within the Netherlands, patients should be treated with plans of equal high quality independent of the institute they are being treated at. The aim of the current pilot study was to gain insight in the different plan evaluation criteria used in 4 Dutch radiotherapy institutes. An earlier survey showed that a wide variation is present in radiation treatment techniques for breast cancer between radiation oncology departments in the Netherlands. To decide on which technique may be the optimal one, consensus is needed on plan evaluation criteria. Material and Methods The 4 participating institutes represent a mixture of characteristic hospitals within the Netherlands (academic, general and specialized radiotherapy hospitals). Each institute provided data of 10 to 15 patients for 3 different treatments: left breast radiotherapy without boost, left breast radiotherapy with integrated boost and left breast radiotherapy including lymph node levels I-IV. For all patients, deep-inspiration breathhold was used. Treatment technique, delineation and DVH parameters were assessed. The following DVH parameters were evaluated: PTV V95%, Dmean, D2%, D98%, Paddick CI, MHD, Heart V5Gy, Heart V30Gy, MLD, Lungs V5Gy and V20Gy. Results All institutes used a tangential IMRT technique and added IMRT and/or VMAT beams in case of a boost or lymph node irradiation. Delineation varied a lot: two centers delineated the CTV according to the ESTRO atlas, whereas two centers based the CTV mainly on markers put in place before CT-scanning and also included some non-breast tissue. Consequently, DVH parameters with respect to target coverage also varied. Our data showed that for all treatment groups, PTV V95% as well as Dmean differed on average by < 2.5 % (p-values 0.001-0.6) between the institutes. MHD and MLD fluctuated typically by about 1 Gy (p values 0.001-0.9) and < 1.5 Gy (p-values 0.001-0.8), respectively. For almost all patients, MHD < 3 Gy and MLD < 4/5/7 Gy for group 1, 2, 3 respectively. CI ranged between 0.67 and 0.74 for breast and breast with integrated boost (p-values 0.001-0.03). For breast plus lymph node levels I-IV, CI varied between 0.41 and 0.69 (p-values 0.00-0.03). Although the differences are relatively small for most cases, some were statistically significant and may also be clinically relevant. Conclusion Based on the results of this 4-institute study, it can be concluded that the differences in radiation technique do not seem to lead to large variations in target volume coverage and specified doses to heart and lung. However, quite some large variation was seen in CI, suggesting that the institutes take different considerations into account when evaluating target dose conformity versus more dose in the OAR. The next step will be to organize a workshop with different institutes to discuss the observed differences, aiming to reach consensus on the tissues to be spared and on accepted dose-volume levels.