Click here to


Are you sure ?

Yes, do it No, cancel

Clinical Optimization of Diffusion Weighted Imaging On An MR-Linac

H Nusrat1,2*, R Chan3, A Elzibak1, B Chugh1,2, A Lau3, A Sahgal2, B Keller1,2, (1) Department of Medical Physics, Odette Cancer Centre, Sunnybrook Health Sciences Centre, Toronto, ON, CA, (2) Department of Radiation Oncology, University of Toronto, Toronto, ON, CA, (3) Sunnybrook Research Institute, Toronto, ON, CA


(Tuesday, 7/14/2020) 3:30 PM - 5:30 PM [Eastern Time (GMT-4)]

Room: Track 2

Purpose: With the advent MR-linac technology, diffusion weighted imaging (DWI) during radiotherapy (RT) can be used to monitor treatment via functional changes in the tumor. Despite DWI being common in diagnostic imaging, the optimal scan parameters for DWI in an MR-linac are unknown and will be different due to its specialized hardware constraints. Several scan parameters require optimization along with validation and testing. In this work, DWI was examined in an MR-linac using a standardized phantom in order to examine and optimize DWI prior to clinical implementation.

Methods: Using the NIST-QIBA Phantom (High Precision Devices, Boulder, USA), DWI scans were run on a 1.5T Elekta Unity MR-linac (Elekta AB, Stockholm, Sweden) as well as a Philips Ingenia 1.5T MR system (Philips, Amsterdam, Netherlands) and scanner-reconstructed ADC maps were used for analysis. Twenty-four hours prior to measurement, the phantom was prepared with ice and water, ensuring a homogeneous temperature of 0°C(+/-0.3°). Both MR-linac and MR-simulator scans were validated by comparing the measured ADC values from the manufacturer recommended scan to their standard data. Four sequences were used on each machine: phantom (b-values: 4; range: 0-2000), brain (6; 0-1000), prostate (12; 0-1000), and head and neck (12; 0-800). Scans were conducted monthly; during each measurement, scans were repeated (n=3) to account for scanner and set-up variation.

Results: For the brain, head and neck, and phantom sequences, ADC values agreed within 72×10?6 mm²/s. For the prostate sequence, MR-linac consistently measured lower values ((933+/-5)×10?6 mm²/s vs. (1187+/-10)×10?6 mm²/s in MR-sim for vials with highest ADC – 0%PVP).

Conclusion: DWI in the MR-linac was examined using several sequences. For certain sequences, ADC measurements were biased due to noise at high b-values. Results suggest that rigorous QA must be done when an MR-simulator image is used as a baseline for DWI-based plan adaptation on an MR-linac.

Funding Support, Disclosures, and Conflict of Interest: FedDev ON, Canada


MRI, Diffusion, Image-guided Therapy


IM/TH- MRI in Radiation Therapy: MRI protocols for therapy

Contact Email