Regularized B-spline deformable registration for respiratory motion correction in PET images

Area of Science:

• Medical imaging
• Radiology
• Image processing

Background:

• Respiratory motion is a significant challenge in thoracic Positron Emission Tomography (PET) imaging.
• This motion degrades image quality, potentially impacting diagnosis and patient management.
• Current methods may not fully address motion artifacts or may increase radiation exposure.

Purpose of the Study:

• To present a novel deformable registration algorithm for respiratory gated PET images.
• To address the challenge of respiratory motion artifacts in thoracic PET imaging.
• To improve PET image quality without increasing radiation burden.

Main Methods:

• A deformable registration algorithm was developed specifically for respiratory gated PET images.
• The registration process relies solely on PET image data, avoiding additional radiation exposure.
• A Markov random field regularizer was incorporated to penalize noisy deformation fields during registration.

Main Results:

• The regularized registration algorithm effectively suppressed noise in the PET images.
• Satisfactory deformation fields were achieved, accurately reflecting respiratory motion.
• Motion correction using the developed algorithm led to a significant improvement in overall PET image quality.

Conclusions:

• The proposed deformable registration algorithm with Markov random field regularization is effective for motion correction in thoracic PET imaging.
• This approach enhances image quality by reducing artifacts caused by respiratory motion.
• The method offers a non-invasive solution to improve diagnostic accuracy in PET scans.