A 2D MTF approach to evaluate and guide dynamic imaging developments.

Area of Science:

• Medical Imaging
• Image Reconstruction
• Dynamic Imaging

Background:

• Partially sampled dynamic imaging methods are increasingly complex.
• Reliable performance evaluation strategies are crucial for these advanced techniques.
• Existing methods may lack robustness or require internal knowledge of reconstruction algorithms.

Purpose of the Study:

• To present an analytical framework for evaluating dynamic imaging reconstruction methods.
• To develop a robust evaluation scheme independent of reconstruction method specifics.
• To introduce methods for assessing noise and artifact properties across frequencies.

Main Methods:

• Developed a perturbation algorithm for robust evaluation.
• Utilized a two-dimensional modulation transfer function (MTF) for spatial-temporal frequency analysis.
• Tested the framework on phantom and cardiac cine datasets with subsampling (R=4, 8).
• Proposed and evaluated a hybrid reconstruction method combining advantageous features.

Main Results:

• The framework successfully evaluated different reconstruction methods.
• The 2D MTF provided an interpretable assessment of spatial and temporal frequency capture.
• The proposed hybrid method demonstrated effective performance in cardiac cine imaging at acceleration factors of 3.5 and 6.3.

Conclusions:

• The presented analytical framework offers a reliable and robust method for evaluating dynamic imaging reconstruction.
• The hybrid approach effectively integrates beneficial features from various reconstruction techniques.
• This work provides essential tools for advancing the field of accelerated dynamic MRI.