Amyloid PET in European and North American cohorts; and exploring age as a limit to clinical use of amyloid imaging

Abstract

Purpose: Several radiotracers that bind to fibrillar amyloid-beta in the brain have been developed and used in various patient cohorts. This study aimed to investigate the comparability of two amyloid positron emission tomography (PET) tracers as well as examine how age affects the discriminative properties of amyloid PET imaging.

Methods: Fifty-one healthy controls (HCs), 72 patients with mild cognitive impairment (MCI) and 90 patients with Alzheimer's disease (AD) from a European cohort were scanned with [11C]Pittsburgh compound-B (PIB) and compared with an age-, sex- and disease severity-matched population of 51 HC, 72 MCI and 84 AD patients from a North American cohort who were scanned with [18F]Florbetapir. An additional North American population of 246 HC, 342 MCI and 138 AD patients with a Florbetapir scan was split by age (55-75 vs 76-93 y) into groups matched for gender and disease severity. PET template-based analyses were used to quantify regional tracer uptake.

Results: The mean regional uptake patterns were similar and strong correlations were found between the two tracers across the regions of interest in HC (ρ = 0.671, p = 0.02), amyloid-positive MCI (ρ = 0.902, p < 0.001) and AD patients (ρ = 0.853, p < 0.001). The application of the Florbetapir cut-off point resulted in a higher proportion of amyloid-positive HC and a lower proportion of amyloid-positive AD patients in the older group (28 and 30 %, respectively) than in the younger group (19 and 20 %, respectively).

Conclusions: These results illustrate the comparability of Florbetapir and PIB in unrelated but matched patient populations. The role of amyloid PET imaging becomes increasingly important with increasing age in the diagnostic assessment of clinically impaired patients.

Fig. 1

Typical examples of individual SUVR images from HC, and MCI amyloid-negative (MCI-), MCI amyloid-positive (MCI+) and AD patients matched for age, gender and MMSE. [11C]PIB scans are on the left and [18F]Florbetapir scans are on the right. Amyloid positivity has been defined as a CCTXR value above the cut-off points of 1.42 for [11C]PIB and 1.34 for [18F]Florbetapir

Fig. 2

Histogram representing the CCTXR values in HC, MCI and AD patients for [11C]PIB and [18F]Florbetapir. The dotted line in each plot represents the calculated [11C]PIB and [18F]Florbetapir cut-off points (1.42 and 1.34, respectively)

Fig. 3

Means ± SDs of tracers’ SUVR values for the investigated ROIs in the respective diagnostic groups; HC, and MCI amyloid-negative (MCI-), MCI amyloid-positive (MCI+) and AD patients. Amyloid positivity has been defined as a CCTXR value above the cut-off points of 1.42 for [11C]PIB and 1.34 for [18F]Florbetapir. The data represented graphically here are tabulated (means ± SDs) in Supplementary Table 2. ([18F]Florbetapir = coral; [11C]PIB = turquoise)

Fig. 4

Scatterplots representing the mean uptake ratios (SUVRs) of each ROI for [11C]PIB versus [18F]Florbetapir in the respective groups; (a) HC (amyloid-positive and amyloid-negative individuals), and MCI amyloid-negative (MCI-), MCI amyloid-positive (MCI+) and AD patients; and (b) the same graph individually for the HC group (amyloid-positive and amyloid-negative individuals). The linear equation derived from the correlation in the HC group (b) was used to convert the [11C]PIB cut-off point (1.42) to a [18F]Florbetapir cut-off point (1.32). Amyloid positivity was defined as a CCTXR value above the cut-off points of 1.42 for [11C]PIB and 1.34 for [18F]Florbetapir. Every data point represents the mean value for a bilateral ROI and not an individual participant. †: Caudate nucleus; ¥: Occipital cortex

Fig. 5

Density plots of the [18F]Florbetapir CCTXR values of younger (y; 55–75 y) and older (o; 76–93 y) age groups in (a) HCs, (b) MCI patients and (c) AD patients. Mixture model analysis using the expectation-maximisation algorithm allowed investigation of the possible underlying components of the distributions. The dotted line in each plot represents the calculated [18F]Florbetapir cut-off point (1.34)