Motion modeling from 4D MR images of liver simulating phantom

Area of Science:

• Medical Imaging
• Radiotherapy Physics
• Computational Anatomy

Background:

• Developing patient-specific models for radiotherapy planning (RTP) requires accurate liver motion data.
• Four-dimensional magnetic resonance imaging (4D-MRI) offers a potential method for capturing this motion retrospectively.

Purpose of the Study:

• To develop and validate a novel 4D-MRI method for retrospective liver modeling.
• To generate a subject-specific deformable liver motion model for RTP.

Main Methods:

• A deformable phantom was imaged using 1.5 T MRI and CT scanners.
• Image registration techniques and displacement vector fields (DVFs) were used to develop a motion model from 4D-MRI data.
• Extensive phantom tests were conducted to evaluate the accuracy and geometric distortion of the 4D-MRI method.

Main Results:

• The 4D-MRI method achieved an accuracy of 2 mm compared to 4D-CT.
• Mean accuracy between coordinates and DVFs from 4D images was less than 2 mm.
• Geometric accuracy tests showed mean distortions of 0.9 mm and 2.2 mm at different distances from the isocenter.

Conclusions:

• The 4D-MRI method has been successfully validated through phantom testing.
• This validated method is crucial for generating patient-specific deformable liver models for RTP.
• The developed motion model will be integrated into the patient-specific liver modeling process.