Authors (including presenting author) :
Wu WS(1), Yip WLW(1), Cheng NS(1), Leung KC(1), Lee VWY(1), Ngar YK(1), Wong CS(1)
Affiliation :
(1) Department of Clinical Oncology, Tuen Mun Hospital
Introduction :
In radiotherapy, magnetic resonance imaging (MRI) is mostly necessary for tumor delineation due to its superior soft tissue contrast. However, its utilization in treatment planning has been constrained by the lack of electron density information, necessitating the addition of computed tomography simulator (CT-Sim) for dose calculations. The introduction of the synthetic CT, Magnetic Resonance for Calculating Attenuation (MRCAT), can potentially reduce CT/MRI fusion error, avoid patient multiple attendances and alleviate the workload on CT-Sim. However, MRCAT is new to Hospital Authority (HA) and has not been widely adopted in daily practice.
Objectives :
This study aims to establish the MR-based workflow and test the feasibility of application of MRCAT, which generates CT-like density maps from a single MRI sequence (3D-FFE mDIXON), in an oncology centre of HA with MR simulator (MR-Sim).
Methodology :
To ensure the safe implementation of MR-based workflow, a two-phase study was conducted. Phase I involved a retrospective analysis of dose accuracy using MRCAT in 20 patients who underwent CT-based planning. Phase II concentrated on prospectively analyzing the accuracy of MRCAT-based planning compared to a backup CT in 7 patients.
Result & Outcome :
MRCAT was successfully generated for 40 out of 42 cervical cancer patients (95.2%) treated by pelvic radiotherapy from 2023 to 2024. Failure modes (4.8%), such as uncorrected positioning, angulated femur, and motion artifacts, were identified by the software's sanity check. In Phase I, CT-based treatment plans of 20 patients were recalculated retrospectively using MRCAT, revealing an average dose difference to the tumor of 0.25±0.32%, with localization translational differences of -0.20±0.31mm, 0.48±0.44mm, and 0.31±0.75mm, respectively. In Phase II, dose accuracy of 7 patients underwent MRCAT-based planning in 2024 was validated with a backup CT, yielding an average dose difference to the tumor of 0.22±0.30%. The simulation time of MR-Sim plus CT-Sim compared with MR-Sim alone is 1.5 hours and 1 hour, respectively.
The MRCAT-based planning showed clinically acceptable dose accuracy compared to MRCAT-CT-based planning for cervical cancer patient, and hence MR-only workflow and MRCAT can be safely adopted with omission of CT simulation. It can reduce simulation time of 33%, and potentially minimize patient attendances and decrease the CT-Sim’s workload.
