Comparison of phantom and registration scaling corrections using the ADNI cohort

Abstract

Rates of brain atrophy derived from serial magnetic resonance (MR) studies may be used to assess therapies for Alzheimer's disease (AD). These measures may be confounded by changes in scanner voxel sizes. For this reason, the Alzheimer's Disease Neuroimaging Initiative (ADNI) included the imaging of a geometric phantom with every scan. This study compares voxel scaling correction using a phantom with correction using a 9 degrees of freedom (9DOF) registration algorithm. We took 129 pairs of baseline and 1-year repeat scans, and calculated the volume scaling correction, previously measured using the phantom. We used the registration algorithm to quantify any residual scaling errors, and found the algorithm to be unbiased, with no significant (p=0.97) difference between control (n=79) and AD subjects (n=50), but with a mean (SD) absolute volume change of 0.20 (0.20) % due to linear scalings. 9DOF registration was shown to be comparable to geometric phantom correction in terms of the effect on atrophy measurement and unbiased with respect to disease status. These results suggest that the additional expense and logistic effort of scanning a phantom with every patient scan can be avoided by registration-based scaling correction. Furthermore, based upon the atrophy rates in the AD subjects in this study, sample size requirements would be approximately 10-12% lower with (either) correction for voxel scaling than if no correction was used.

Fig. 1

The baseline percentage volume correction (X-axis) against the corresponding repeat image percentage volume correction (Y-axis), calculated using the phantom correction procedure. Any points that lie off the y = x line represent where the phantom correction differed between baseline and follow-up—which should indicate the presence of a scaling change on the scanner between the two MRI exams.

Fig. 2

The distribution of the absolute repeat scan to baseline scan phantom corrected volume scale factors as a percentage. 97 of the 129 cases have a volume scale correction of <0.5%, and 115 examples have a volume scale correction of <0.75%.

Fig. 3

A Bland–Altman plot showing the mean and difference between the percentage volume scale correction calculated using the phantom procedure, and the percentage volume scale correction calculated using the 9DOF registration on the uncorrected images.

Fig. 4

Example images, to demonstrate the scaling errors visually inspected in the Experiments section. Image (a) is a baseline image, coronal view, and figure (b) is the 6DOF registered repeat image. The red bar is the same size in all images, and yet the enlarged section shows clear differences between (a) and (b), most visible at the end of the bar. The difference is a superior–inferior scaling issue. Images (c) and (d) show the same subject as (a) and (b) but in saggital view. Image (c) is the baseline, and (d) is the 6DOF registered repeat image.