Time-resolved volumetric MRI in MRI-guided radiotherapy: an in silico comparative analysis

Area of Science:

• Medical Physics
• Radiotherapy Technology
• Medical Imaging

Background:

• 2D cine-MRI in radiotherapy aids motion detection but lacks 3D volumetric data due to spatio-temporal limitations.
• Existing methods to derive time-resolved 3D MRI information require rigorous comparative analysis.

Purpose of the Study:

• To quantitatively investigate and compare five published strategies for deriving time-resolved volumetric MRI in MRI-guided radiotherapy.
• To evaluate the performance of these strategies using a digital phantom and patient-derived motion data.

Main Methods:

• Utilized an MRI digital phantom with patient-derived motion signals and tumor shapes.
• Generated average 4D cycles and 2D cine-MRI data with 3D ground truth.
• Quantitatively analyzed five strategies: Propagation, out-of-plane motion compensation, Fayad model, ROI-based model, and Stemkens model.

Main Results:

• Propagation and ROI-based models demonstrated superior performance with median tumor center of mass 3D distances of 1.1 mm and 1.3 mm, respectively.
• Diaphragm motion was accurately captured with distances below 1.6 mm for these top models.
• Other methods exhibited higher errors, particularly with irregular breathing patterns and in-plane variations.

Conclusions:

• The Propagation and ROI-based models are most effective for time-resolved 3D MRI in radiotherapy.
• The developed in silico framework facilitates the development and comparison of strategies for improved MRI-guided radiotherapy.