ESTRO 2024 - Abstract Book

S3883

Physics - Image acquisition and processing

ESTRO 2024

2020

Digital Poster

Comparison of receive coil setups for brain MR imaging with radiotherapy immobilization devices

Bertrand Pouymayou, Riccardo Dal Bello, Ya Wang, Alessandro Mencarelli, Nicolaus Andratschke, Michael Mayinger, Matthias Guckenberger, Stephanie Tanadini-Lang

University Hospital Zürich, Department of Radiation Oncology, Zürich, Switzerland

Purpose/Objective:

Brain MRI in radiotherapy (RT) treatment position facilitates CT registration 1 , offers multiple contrasts, and enables MR-only RT workflow. However, immobilization devices introduce further constraints regarding coil positioning compared to diagnostic setups. In this project we compare 3 surface coil arrangements compatible with commercial immobilization systems to a diagnostic head and neck coil on a 1.5T scanner (Sola, Siemens). We report detailed signal-to-noise ratio (SNR) profiles in the three orthogonal planes and demonstrate the benefice of integrating receive surface coils within immobilization devices.

Material/Methods:

Three standard vendor coils were used in the different configurations: i) an Ultraflex 18 channels large ii) an Ultraflex 18 channels small iii) a 11cm diameter loop coil. The coil arrangements and immobilization devices are described in table 1. Setup a) is recommended by the vendor and inspired by (Mengling et al 2 ) the small flex coil lies under the table top. Setup b) uses a custom carved headrest (Civco Posifix) to fit a 11cm loop coil between the patient head and the treatment table top. Setup c) uses a custom 3D-printed support exploiting the space under the SRS Brainlab mask to fit the small Ultraflex coil between the posterior part of the patient head and the table top. Those three setups were compared to a Siemens 20 channels head and neck coil (setup d) as a diagnostic reference. The latter does not allow the use of immobilization devices. In this work we focus on a T1 sequence as a standard in brain tumor delineation (Paulson et al 3 ). A 16.5cm diameter spherical phantom (1,25g/l NiSO 4 ) was scanned with a T1 MPRAGE sequence (resolution 0.49x0.49x1mm 3 , FOV: 249x249x192 mm 3 , acceleration factor 2, normalization: prescan). The coil combination method was set to adaptive combine. The images were acquired twice with and without RF power voltage to obtain an image and the corresponding noise measurement. Those images were combined to obtain an SNR measurement as defined by the NEMA MS 6-2008 standard:

SNR=0.655*(mean ROI Intensity)/(standard deviation ROI noise)

We automatically detected the phantom shape (Matlab) in each slice and we defined the region of interest (ROI) as a disk corresponding to the detected diameter minus 2cm and with a minimal area of 10cm 2 .