Volume registration based on 3-D phase correlation for tumor motion estimation in 4-D CT

Area of Science:

• Medical Imaging
• Radiation Oncology
• Image Registration

Background:

• Respiration-induced tumor motion in four-dimensional (4-D) computed tomography (CT) poses challenges in radiation therapy.
• Accurate estimation of tumor motion is crucial for effective radiation targeting.

Purpose of the Study:

• To present a novel three-dimensional (3-D) volume registration method for estimating respiration-induced tumor motion.
• To enhance the accuracy of tumor motion estimation in 4-D CT for radiation therapy.

Main Methods:

• The study extends the 2-D phase correlation method to 3-D volume registration.
• 3-D phase correlation is computed using the 3-D inverse Fourier transform of a normalized cross power spectrum.
• Tumor motion is modeled as a translational shift and estimated by identifying the peak in the 3-D phase correlation, with sub-voxel accuracy improvements.

Main Results:

• The proposed 3-D phase correlation method effectively estimates tumor motion along three directions.
• The method achieves sub-voxel accuracy for improved motion estimation.
• Experimental results show the superiority of the 3-D method over conventional 2-D phase correlation techniques.

Conclusions:

• The developed 3-D phase correlation method provides an accurate and effective approach for estimating respiration-induced tumor motion in 4-D CT.
• This technique has the potential to improve the precision and efficacy of radiation therapy by accounting for tumor movement.