ESTRO 37 Abstract book

S152

ESTRO 37

Purpose or Objective To investigate the feasibility of hypoxia-PET based dose escalation (DE) in head and neck cancer (HN C) patients with a high-field MR-linac and to compare treatment plan quality of DE applied with the MR-linac

and a standard linac. Material and Methods Ten patients with locally advanced HNC who

received dynamic FMISO PET/CT scans and a treatment planning CT prior to definitive radio-chemotherapy in a phase II dose escalation trial were retrospectively included in this study. Two treatment plans were created for each patient with Monaco Research 5.19.03 (Elekta AB, Stockholm Sweden) using the same patient model and fractionation scheme with 30×2 Gy and 30×1.8 Gy in the PTV 60 and PTV 50 , respectively and a sequential boost of 5×2 Gy in the PTV 70 . Dose was escalated to 2.2 Gy for all fractions in hypoxic sub-volumes (HV) of the PTV 70 identified by dynamic FMISO PET. Delineations of HVs were available from the original treatment plan. For the standard linac, VMAT plans with two full arcs were created. For the MR-linac (Elekta AB, Stockholm Sweden), Step-and-Shoot (SNS) plans with nine beams were created with automatically optimized beam angles based on the angular MU distributions in the corresponding VMAT plan. Specifications of both treatment units are given in Tab. 1. Treatment plans were optimized to show similar mean doses in the escalation volume and PTV 70 . Typical dose parameters were compared between the two plans using a Wilcoxon signed rank test in R 3.2.3.

Conclusion Treatment planning for definitive radio-chemotherapy of locally advanced head and neck carcinoma patients including a 10% dose escalation in hypoxic tumour sub- volumes is feasible with the current high-field MR-linac technology. MR-linac plans are slightly less conformal and show slightly inferior OAR sparing, which may in the future be compensated by margin reduction due to MR- based daily adapted radiotherapy and implementation of VMAT. OC-0297 Dosimetric feasibility of on-body receive array placement to enhance image quality of the MR- linac S.E. Zijlema 1 , L. Van Dijk 1 , S.L. Hackett 1 , J.J.W. Lagendijk 1 , H.N. Tijssen 1 , C.A.T. Van den Berg 1 1 UMC Utrecht, Radiotherapy, Utrecht, The Netherlands Purpose or Objective The current MR-linac anterior and posterior receive arrays contain four channels each. The anterior array is elevated above the patient using a coil bridge. To increase the image quality and speed in the MR-linac, we are developing a new 64-channel receive array, which will be placed directly onto the patient. Consequently, more secondary electrons that originate from the copper and plastic will reach the skin. In this work we assess the dosimetric impact of various coil materials and optimize the radiolucent design of the receive array in the MR- linac. Material and Methods Prototypes were created with five layers: leather, plastic support, copper loops, foam, and leather again. The low (electron) density foam serves as a spacer between the material with high electron density (i.e., copper and plastic) and the patient surface to reduce skin dose. There is, however, a tradeoff: a larger foam thickness results in larger coil-to-body distance at the expense of imaging performance. Optimal foam thickness: multiple 1×2 element array prototypes with foam thicknesses ranging from 4 to 15 mm were irradiated with a 250 MU, 22×56 cm² field on an MR-linac (Figure 1a,b). Dosimetry was performed using GAFCHROMIC (Ashland, Bridgewater, NJ) EBT3 films. Doses were measured without the coils ( D_open ) and with the coils placed directly onto the film ( D_coil ).

Results For both techniques, it was generally possible to create plans adhering to our internal guidelines, though compromises were slightly more frequent for MR-linac plans, e.g. with respect to D max in the mandible. Hypoxia dose escalation as per study protocol was possible for MR- linac plans, with a median D mean of 75.6 Gy in the escalation volume. MR- and standard linac plans showed similar coverage of the escalation volume with a median difference of 0.24 Gy in D mean (p=0.32). For PTV 70 a small, but significant, dose increase (0.22 Gy, p=0.02) was found in MR-linac plans. This trend persists in PTV 60 (0.85 Gy, p=0.01) and PTV 54 (0.66 Gy, p=0.01). The reason is a lower conformity of the MR-linac dose distributions. Only for one major organ at risk parameter a significant difference was found, an increased D mean in the right parotids in MR-linac plans (Fig. 1 shows positive differences for an increased dose in MR-linac plans). Further differences were observed in OARs with lower priority, e.g. a median increase in mandible D max of 0.96 Gy (p=0.04).