Neuropathologic heterogeneity does not impair florbetapir-positron emission tomography postmortem correlates

Abstract

Neuropathologic heterogeneity is often present among Alzheimer disease (AD) patients. We sought to determine whether amyloid imaging measures of AD are affected by concurrent pathologies. Thirty-eight clinically and pathologically defined AD and 17 nondemented patients with quantitative florbetapir F-18 (F-AV-45) positron emission tomography (PET) imaging during life and postmortem histological β-amyloid quantification and neuropathologic examination were assessed. AD patients were divided on the basis of concurrent pathologies, including those with Lewy bodies (LBs) (n = 21), white matter rarefaction (n = 27), severe cerebral amyloid angiopathy (n = 11), argyrophilic grains (n = 5), and TAR DNA binding protein-43 inclusions (n = 18). Many patients exhibited more than 1 type of concurrent pathology. The ratio of cortical to cerebellar amyloid imaging signal (SUVr) and immunohistochemical β-amyloid load were analyzed in 6 cortical regions of interest. All AD subgroups had strong and significant correlations between SUVr and histological β-amyloid measures (p μ 0.001). All AD subgroups had significantly greater amyloid measures versus nondemented patients, and mean amyloid measures did not significantly differ between AD subgroups. When comparing AD cases with and without each pathology, AD cases with LBs had significantly lower SUVr measures versus AD cases without LBs (p = 0.002); there were no other paired comparison differences. These findings indicate that florbetapir-PET imaging is not confounded by neuropathological heterogeneity within AD.

Figure 1

Examples of severity scores given to plaque densities (top) and white matter rarefaction (WMR- bottom). Top: 80 µm sections of the superior frontal gyrus stained with the Campbell Switzer enhanced silver stain -from left to right: 0 - none, 1 - mild, 2 - moderate, and 3 - frequent plaque densities. Bottom: macro view of 80 µm sections of the frontal lobe stained with hematoxylin and eosin, WMR was scored as from left to right: 0 - none, 1 - mild, 2 - moderate, and 3 - severe. In this study, a case was defined as having significant WMR if it had a score of 2 or higher in one or more of the following lobes: frontal, parietal, temporal and occipital.

Figure 2

Measures of average cortical amyloid load in AD subgroups with different concurrent pathologies. Left: Graphs demonstrating the variability of in vivo amyloid imaging measures for average cortical amyloid load, using the mean standard uptake value ratios (SUVr) for each AD subgroup. The thick black line on all graphs represents the mean of all AD cases, while the dashed and lighter lines represent one and two standard deviations from the mean, respectively. The red line at an SUVr value of 1.1. represents an adopted SUVr cut off between amyloid positive and negative (32). The x-axis numbers are the individual case numbers as listed in Table 2. Right: photos taken at 10× of the respective pathologies. Panel A: AD with white matter rarefaction (WMR) – photo from an 80 µm section of parietal cortex white matter stained with hematoxylin and eosin. Panel B: AD with Lewy bodies (LBs-grey circles indicate clinicopathological diagnosis of DLB), photo from a 5 µm paraffin amygdala section stained immunohistochemically for phosphorylated α-synuclein (black) and counterstained with Neutral Red. Panel C: AD with argyrophilic grains (Arg), photo from an 80 µm section of amygdala stained with the Gallyas silver method. Panel D: AD with TDP-43 inclusions (white diamond indicates case with a clinicopathological diagnosis of FTLD-TDP-43 in addition to AD) ; photo taken from a 40 µm section of middle frontal gyrus that was immunohistochemically stained for phosphorylated TDP-43 (black) and counterstained with Neutral Red. Panel E: AD with severe cerebral amyloid angiopathy (CAA); photo from an 80 µm Thioflavin-S stained section of frontal cortex.

Fig. 3

Box plots of median, 25^th and 75^th percentile of in vivo amyloid imaging measures for average cortical amyloid load of Alzheimer’s disease cases using the standard uptake value ratios (SUVr) in subjects with presence (gray boxes) or absence (white boxes) of concurrent pathologies: (CAA; with N = 11; without N = 27), TDP-43 (with N = 18; without N = 20), LBs (with n = 21, without N = 17), WMR (with N = 27, without N = 11), and Arg (with N = 5, without N = 33). Whiskers above and below the box indicate the 90th and 10th percentiles. Utilizing the Mann-Whitney U-test, the only significantly different pairwise comparison was that comparing AD subjects with and without LBs (p = 0.002).