A modular phantom and software to characterize 3D geometric distortion in MRI

Area of Science:

• Medical Physics
• Radiotherapy Imaging
• Image Quality Assessment

Background:

• Magnetic Resonance Imaging (MRI) provides excellent soft tissue contrast crucial for radiation therapy planning.
• Spatial distortions in MRI can lead to systematic errors in treatment delivery.
• Accurate geometric calibration is essential for online adaptive image-guided radiation therapy.

Purpose of the Study:

• To develop and validate a modular phantom and software for characterizing geometric distortion (GD) in large field-of-view MRI images.
• To assess the impact of acquisition parameters on GD in MRI for radiation therapy simulation.
• To ensure the reliability and transportability of MRI distortion assessment tools.

Main Methods:

• A modular phantom with fiducial markers was designed for transportability and adaptability to different MRI systems.
• The phantom was evaluated on a 1.5 T MR-guided linear accelerator (MR-Linac) and 1.5 T/3.0 T diagnostic scanners.
• Image acquisition involved varying pulse sequences (T1/T2 weighted), correction algorithms (2D/3D), and receiver bandwidth (BW).

Main Results:

• Average GD was quantified across different MRI systems: 0.94 ± 0.58 mm (MR-Linac), 0.90 ± 0.53 mm (1.5 T), and 1.15 ± 0.62 mm (3.0 T) within a 400 mm diameter.
• Geometric distortion increased with decreasing receiver bandwidth and with 2D versus 3D correction algorithms.
• Phantom set-up reproducibility and software validation demonstrated high accuracy (GD differences ≤ 0.13 mm and ≤ 0.07 mm, respectively).

Conclusions:

• A novel modular phantom and software effectively characterize geometric distortions in MRI for radiation therapy.
• The developed system ensures accurate spatial fidelity of MRI data, crucial for precise treatment planning and delivery.
• This tool enhances the reliability of MRI in image-guided radiation therapy workflows.