An interactive technique for three-dimensional image registration: validation for PET, SPECT, MRI and CT brain studies

Abstract

A multipurpose three-dimensional registration technique was validated with PET, SPECT, CT and MRI scans, which had been obtained under normal clinical conditions. In contrast to fully automated procedures, this coregistration method is highly interactive, which has the advantage that it does not impose rigid restrictions by data type and by alterations in normal anatomy or brain function resulting from disease.

Methods: Basically, a computer program provides a variety of tools to examine the accuracy of coregistration visually and to specify necessary translations and rotations in all three dimensions. Tools and criteria to accept coregistration were applied according to a standardized protocol. Reproducibility was assessed with five independent users on nine pairs of image sets. In two pairs of these image sets, coregistration was repeated three times by each user.

Results: Depending on the resolution of the images involved, the reproducibility of translation distances ranged from 0.32 to 2.22 mm (s.d.) and of rotation angles from 0.32 to 1.70 degrees. It was always much smaller than the point-spread full-width half maximum of the device with the lower resolution. The accuracy of coregistration was examined using two arbitrarily misplaced image sets. Interindividual and intraindividual variance were similar, which suggested that the influence of subjectivity was not significant. Average displacements after coregistration were 0.43 and 0.29 mm or less for PET and MRI data, respectively, which indicated the absence of a systematic bias.

Conclusion: The results indicate the high reproducibility and accuracy of this three-dimensional coregistration technique, which is comparable or superior to those of automated techniques and methods based on external artificial landmarks.