Partial-volume effect correction improves quantitative analysis of 18F-florbetaben β-amyloid PET scans

Abstract

Neocortical atrophy reduces PET signal intensity, potentially affecting the diagnostic efficacy of ?-amyloid (A?) brain PET imaging. This study investigated whether partial-volume effect correction (PVEC), adjusting for this atrophy bias, improves the accuracy of 18F-florbetaben A? PET. Methods: We analyzed 18F-florbetaben PET and MRI data obtained from 3 cohorts. The first was 10 patients with probable Alzheimer disease (AD) and 10 age-matched healthy controls (HCs), the second was 31 subjects who underwent in vivo imaging and postmortem histopathology for A? plaques, and the third was 5 subjects who underwent PET and MRI at baseline and 1 y later. The imaging data were coregistered and segmented. PVEC was performed using the voxel-based modified Müller-Gärtner method (PVELab, SPM8). From the PET data, regional and composite SUV ratios (SUVRs) with and without PVEC were obtained. In the MRI data, mesial temporal lobe atrophy was determined by the Scheltens mesial temporal atrophy scale and gray matter volumes by voxel-based morphometry. Results: In cohort 1, PVEC increased the effect on AD-versus-HC discrimination from a Cohen d value of 1.68 to 2.0 for composite SUVRs and from 0.04 to 1.04 for mesial temporal cortex SUVRs. The PVEC-related increase in mesial temporal cortex SUVR correlated with the Scheltens score (r 5 0.84, P , 0.001), and that of composite SUVR correlated with the composite gray matter volume (r 5 ?0.75, P , 0.001). In cohort 2, PVEC increased the correlation coefficient between mesial temporal cortex SUVR and histopathology score for A? plaque load from 0.28 (P 5 0.09) to 0.37 (P 5 0.03). In cohort 3, PVEC did not affect the composite SUVR dynamics over time for the A?-negative subject. This finDing was in contrast to the 4 A?-positive subjects, in 2 of whom PVEC changed the composite SUVR dynamics. Conclusion: The influence of PVEC on 18F-florbetaben PET data is associated with the degree of brain atrophy. Thus, PVEC increases the ability of 18F-florbetaben PET to discriminate between AD patients and HCs, to detect A? plaques in the atrophic mesial temporal cortex, and potentially to evaluate changes in brain A? load over time. As such, the use of PVEC should be considered for quantitative 18F-florbetaben PET scans, especially in assessing patients with brain atrophy.