Reliable quantification of 18F-GE-180 PET neuroinflammation studies using an individually scaled population-based input function or late tissue-to-blood ratio

Ralph Buchert,Christian Schütze,Florian Wilke,Lilli Geworski,Martin Mamach,András Polyák ,Meike Dirks,Linda Hamann,Mario Lukačević ,Henning Pflugrad,Tobias L Roß ,Christian H Wetzel ,Laura Langer ,Mariella Kessler,Rainer Rupprecht,Frank M Bengel,Karin Weißenborn ,Carlotta Petrusch,Ann-Katrin Wirries,Georg Berding

doi:10.1007/s00259-020-04810-1

Abstract

PurposeTracer kinetic modeling of tissue time activity curves and the individual input function based on arterial blood sampling and metabolite correction is the gold standard for quantitative characterization of microglia activation by PET with the translocator protein (TSPO) ligand 18F-GE-180. This study tested simplified methods for quantification of 18F-GE-180 PET.MethodsDynamic 18F-GE-180 PET with arterial blood sampling and metabolite correction was performed in five healthy volunteers and 20 liver-transplanted patients. Population-based input function templates were generated by averaging individual input functions normalized to the total area under the input function using a leave-one-out approach. Individual population-based input functions were obtained by scaling the input function template with the individual parent activity concentration of 18F-GE-180 in arterial plasma in a blood sample drawn at 27.5 min or by the individual administered tracer activity, respectively. The total 18F-GE-180 distribution volume (VT) was estimated in 12 regions-of-interest (ROIs) by the invasive Logan plot using the measured or the population-based input functions. Late ROI-to-whole-blood and ROI-to-cerebellum ratio were also computed.ResultsCorrelation with the reference VT (with individually measured input function) was very high for VT with the population-based input function scaled with the blood sample and for the ROI-to-whole-blood ratio (Pearson correlation coefficient = 0.989 ± 0.006 and 0.970 ± 0.005). The correlation was only moderate for VT with the population-based input function scaled with tracer activity dose and for the ROI-to-cerebellum ratio (0.653 ± 0.074 and 0.384 ± 0.177). Reference VT, population-based VT with scaling by the blood sample, and ROI-to-whole-blood ratio were sensitive to the TSPO gene polymorphism. Population-based VT with scaling to the administered tracer activity and the ROI-to-cerebellum ratio failed to detect a polymorphism effect.ConclusionThese results support the use of a population-based input function scaled with a single blood sample or the ROI-to-whole-blood ratio at a late time point for simplified quantitative analysis of 18F-GE-180 PET.

Highlights

There is increasing evidence that chronic neuroinflammation caused by cells of the innate neuroimmune system after activation by danger-associated molecular patterns such as misfolded proteins contributes to the pathogenesis of neurodegenerative diseases [1, 2]
Input functions and whole-blood time activity curves (TAC) of 18F-GE-180 fulfill this requirement to good approximation (Fig. 1), there was an effect of the translocator protein (TSPO) genotype: on average, there was an underestimation of the area under the input function in HAB subjects and an overestimation in MAB subjects by the population-based method PB1
The primary finding of this study was the very strong correlation of the Logan VT estimated with population-based blood curves scaled by a single blood sample (PB1 and PB2) as well as the ROI-to-whole-blood ratio at a late time point with the reference VT across all brain regions (Table 2, Fig. 1 The top row shows the measured input functions (MIF) before and after scaling to the area under the individual input function

Summary

Introduction

There is increasing evidence that chronic neuroinflammation caused by cells of the innate neuroimmune system after activation by danger-associated molecular patterns such as misfolded proteins contributes to the pathogenesis of neurodegenerative diseases [1, 2]. Chronic pro-inflammatory reactions of the neuroimmune system most likely play a prominent role in various other neurological and psychiatric diseases including stroke [3], multiple sclerosis [4], brain tumors [5], hepatic encephalopathy [6], and major depression [7]. In patients with mild cognitive impairment or mild dementia of the Alzheimer type, for example, PET with the first-generation TSPO ligand [11C](R)-PK11195 [13] demonstrated increased tracer binding in the brain regions with the most prominent synaptic dysfunction/degeneration in Alzheimer’s disease, suggesting microglial activation at early clinical stages of the disease [14]

Objectives

Methods

Results

Conclusion