Movement correction in DCE-MRI through windowed and reconstruction dynamic mode decomposition.

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Radiology

Background:

• Respiration induces complex organ motion during dynamic contrast-enhanced magnetic resonance renography (DCE-MRR).
• This motion creates artifacts that impede accurate clinical assessment of kidney function.
• Conventional registration techniques often fail due to rapid contrast agent changes in DCE-MR sequences.

Purpose of the Study:

• To develop an automated, registration-free method for correcting respiratory motion artifacts in kidney DCE-MRR.
• To overcome the limitations of semi-automated approaches, such as inter-observer variability and time-consuming manual inspection.

Main Methods:

• Implementation of a novel automated movement correction approach using windowed and reconstruction variants of dynamic mode decomposition (WR-DMD).
• Validation of the WR-DMD method on DCE-MRI data sets from ten healthy volunteers.
• Evaluation using block-matching-block analysis of the image sequence generated by WR-DMD.

Main Results:

• The WR-DMD method successfully eliminated a significant mean motion magnitude compared to original data.
• Demonstrated the elimination of $$ of mean motion magnitude.
• The results confirm the viability of WR-DMD for automatic movement correction in kidney DCE-MRI.

Conclusions:

• WR-DMD offers a robust, automated solution for motion artifact correction in kidney DCE-MRR.
• This technique enhances the reliability and efficiency of kidney function assessment from DCE-MRI data.
• The registration-free nature of WR-DMD addresses key limitations of existing methods.