Characterization of a new physical phantom for testing rigid and deformable image registration

Area of Science:

• Medical imaging
• Image processing
• Biomedical engineering

Background:

• Accurate medical image registration is crucial for various clinical applications, including image-guided surgery and radiation therapy.
• Existing phantoms for testing image registration systems can be complex and expensive, limiting their widespread use.

Purpose of the Study:

• To introduce a novel, user-friendly, and low-cost phantom for evaluating the accuracy of rigid and deformable image registration (DIR) systems.
• To demonstrate the functional efficacy of this new phantom in simulating anatomical variations and deformations.

Main Methods:

• The phantom was constructed from acrylic with interchangeable inserts to mimic diverse tissue properties and shapes.
• CT scans of the phantom were acquired and used to test the Velocity software's rigid and DIR accuracy across multiple rotation, translation, and deformation scenarios.

Main Results:

• Rigid registration accuracy achieved mean errors of 0.50° (axial) and 0.13° (coronal) for rotation, and 1 mm for translation in soft tissue and bone.
• Deformable image registration (DIR) accuracy yielded a mean Dice similarity coefficient of 0.93 for soft tissue and 0.79 mm mean distance for bone.

Conclusions:

• The developed phantom offers a simple, cost-effective solution for verifying DIR system performance.
• This phantom facilitates efficient and effective validation of image registration technologies, enhancing diagnostic and therapeutic accuracy.