Raj Shekhar PhD, Presenter: Nothing to Disclose

Vivek Walimbe, Abstract Co-Author: Nothing to Disclose

Shanker Raja MD, Abstract Co-Author: Nothing to Disclose

Mangesh H. Kanvinde MD, PhD, Abstract Co-Author: Nothing to Disclose

Guiyun Wu, Abstract Co-Author: Nothing to Disclose

Bohdan Bybel MD, Abstract Co-Author: Nothing to Disclose

Combined PET/CT scanners have greatly facilitated the registration and fusion of whole-body PET and CT images by removing common sources of misalignment: differences in body positions and orientations, and couch shapes. Breathing-related nonrigid mismatches, however, still persist, correcting for which is the goal of the present study.

We have developed a new, fully automated mutual information-based three-dimensional elastic (nonrigid) image registration technique that can accurately align whole-body PET and CT images. The algorithm morphs the PET image to align spatially with the CT image by generating an elastic transformation field by interpolating quaternions (rotations) and translations from multiple six-parameter rigid-body registrations, each obtained for hierarchically subdivided image subvolumes. Five whole-body PET/CT image pairs acquired on a combined scanner were registered. The cases were selected randomly without any specific inclusion or exclusion criteria. A rigorous quantitative validation was performed by evaluating algorithm performance in the context of variability among three clinical experts in the identification of up to 32 homologous anatomic landmarks.

Visual inspection indicated improved matching of homologous structures in all five cases. The mean registration accuracy of 5.9 mm was found comparable to the mean inter-expert difference of 6.6 ± 3.4 mm in landmark identification. The variability in landmark identification did not show statistically significant changes when any expert was replaced by the algorithm. The average execution time was 45 minutes.

We have presented a new method based on automated elastic registration to correct for residual nonrigid and mostly breathing-related misalignment artifacts in whole-body PET and CT images acquired on a combined scanner. The algorithm performance was on par with the average opinion of three experts.

Shekhar, R, Walimbe, V, Raja, S, Kanvinde, M, Wu, G, Bybel, B, Improving Registration of Combined Whole-body PET/CT.  Radiological Society of North America 2005 Scientific Assembly and Annual Meeting, November 27 - December 2, 2005 ,Chicago IL. http://archive.rsna.org/2005/4406698.html