ESTRO 2024 - Abstract Book

S3954

Physics - Image acquisition and processing

ESTRO 2024

Figure 2. Gamma analysis comparing dose matrices from s-CT and p-CT def using different criteria (global and local 1%/1mm, and local 2%/2mm) with varying low dose thresholds (10%, 40%, and 90%).

Conclusion:

Dosimetric calculation on the s-CT generated by MRCAT Pelvis algorithm does not significantly differ from that performed in p-CT def as long as there is no much presence of air in abdominal cavity. Therefore, the clinical implementation of an MR-only workflow for prostate is considered safe and feasible.

Keywords: MRCAT, MRI-only workflow, synthetic CT

2859

Digital Poster

Deep Learning reconstructed MR images: impact on manual and automatic contours

Jonathan J Wyatt, April Eassom, Serena West, Samantha Warren, Rachel A Pearson

Newcastle upon Tyne Hospitals NHS Foundation Trust, Northern Centre for Cancer Care, Newcastle, United Kingdom

Purpose/Objective:

MR is the preferred imaging modality for contouring in prostate radiotherapy due to its superior soft-tissue contrast to CT(1). MR images acquired for radiotherapy planning need to be acquired in the radiotherapy position using a flat couch top and coil bridge(2). This significantly reduces the MR signal, resulting in images with high noise levels(3). Recently, Deep Learning (DL) methods of reconstructing MR images have been introduced, which reduce image noise and improve resolution(4). The aim of this study was to investigate the impact of the improved MR image quality on manual and automatic contours for prostate radiotherapy.

Material/Methods: