Click here to


Are you sure ?

Yes, do it No, cancel

Multicenter Characterization of MRIs Used for Radiation Therapy: The Collaborative Quality Assurance Program

L Conroy1*, C Foottit2, B Zhang1,3, A Elzibak4, R Hunter5, P Rapley6, J Kraus Himmelman7, E Gutierrez7, T Tadic1, A McNiven1, D Letourneau1, (1) The Princess Margaret Cancer Centre, University Health Network, Toronto, ON, CA, (2) The Ottawa Hospital, Ottawa, ON, CA, (3) Stronach Regional Cancer Centre, Newmarket, ON, CA (4) Sunnybrook Health Sciences Centre, Odette Cancer Centre, Toronto, ON, CA, (5) Juravinski Cancer Centre, Hamilton, ON, CA, (6) Thunder Bay Regional Health Sciences Center, Thunder Bay, ON, CA, (7) Ontario Health (Cancer Care Ontario), Toronto, ON, CA


(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: The Collaborative Quality Assurance (CQA) Program is an initiative to assess the performance of radiation therapy infrastructure across multiple centers in a single jurisdiction. The current objective is performance characterization of magnetic resonance imaging systems used for RT (MRI-for-RT). We present survey results from 14 centers on MRI-for-RT infrastructure and access, and interim imaging performance results of seven systems from five centers.

Methods: An MRI access survey was sent to 14 cancer programs. Phantoms were shipped to participating centers with vendor-specific setup instructions. Image quality was assessed using the ACR head phantom and geometric distortion was characterized using the Modus QUASAR MRID3D Phantom. The ACR phantom was scanned according to ACR test guidance documentation and analyzed using an in-house automated program (geometric accuracy, high-contrast spatial resolution, slice thickness, slice position, percent-signal ghosting) and manually (image uniformity, low-contrast object detectability). The MRID3D phantom was scanned using a 3D T1-weighted acquisition with geometric distortion correction applied. Geometric distortions were analyzed in a 368mm diameter, 321mm long region using phantom-specific software.

Results: Survey results identified a range of infrastructure (nine models, three vendors) and MRI ownership (dedicated MR-SIM vs. hospital diagnostic MRI), with access from easy/fairly-easy/with notice(n=6) to limited/very-limited(n=8). Of the systems evaluated for imaging performance, 5/7 met all ACR-recommended tolerances. One system failed slice positioning accuracy and low-contrast detectability, another failed geometric accuracy and vertical spatial resolution. The mean(95% confidence interval) gradient distortion vectors with 3D-distortion correction(n=5) ranged from 0.84mm(1.10mm) to 1.10mm(1.40mm); with 2D-distortion correction(n=2) from 2.09mm(3.72mm) to 2.67mm(4.22mm).

Conclusion: We found variable imaging performance results across several MRI models and considerable institutional differences in MRI infrastructure and access. Upon program completion, results from 17 scanners across 11 institutions will facilitate vendor-specific comparisons. Findings will be used to guide and harmonize QA for MRI-for-RT as use of this emerging technology expands.

Funding Support, Disclosures, and Conflict of Interest: T. Tadic holds a patent and licensing agreement with Modus Medical Devices Inc. related to the MRID3D phantom and software used in this work.


MRI, Image Analysis, Quality Assurance


IM/TH- MRI in Radiation Therapy: MRI for treatment planning

Contact Email