An augmented hybrid multibaseline and referenceless MR thermometry motion compensation algorithm for MRgHIFU...

Area of Science:

• Medical Physics
• Biomedical Engineering
• Radiology

Background:

• MR thermometry is crucial for monitoring hyperthermia treatments.
• Motion artifacts can compromise temperature accuracy during MR thermometry.
• Existing motion compensation algorithms struggle with sporadic movements.

Purpose of the Study:

• To optimize and validate a hybrid principal component analysis and projection onto dipole fields (PCA-PDF) algorithm for MR thermometry motion compensation.
• To enhance the PCA-PDF algorithm using atlas image augmentation for improved performance, particularly with sporadic motion.

Main Methods:

• Experiments utilized a 3T Philips MRI and a high-intensity focused ultrasound system.
• An MR-compatible robot induced periodic and sporadic motion in custom gelatin phantoms during hyperthermia.
• The PCA-PDF algorithm was augmented with a predictive atlas to address larger, sporadic motion artifacts.

Main Results:

• The original PCA-PDF algorithm improved temperature standard deviation during periodic motion.
• For large sporadic motion, the augmented PCA-PDF algorithm reduced temperature standard deviation from $$ to $$ C.
• The augmented algorithm significantly enhanced motion compensation compared to the original PCA-PDF.

Conclusions:

• The PCA-PDF algorithm achieves temperature accuracy below 1°C during periodic motion.
• Augmenting the PCA-PDF algorithm with a predictive atlas is essential for accurate thermometry during sporadic motion.
• The enhanced PCA-PDF algorithm substantially improves temperature accuracy in the presence of significant motion during hyperthermia treatments.