Optimization of early‐phase florbetapir as a surrogate of FDG‐PET in ageing and Alzheimer's clinical syndrome: Neuroimaging / New imaging methods. (7th December 2020)

• Record Type: Journal Article
• Authors: Vanhoutte, Matthieu; Landeau, Brigitte; Sherif, Siya; la Sayette, Vincent De; Dautricourt, Sophie; Abbas, Ahmed; Manrique, Alain; Chocat, Anne; Chetelat, Gael
• Is Part Of: Alzheimer's & dementia. Volume 16(2020)Supplement 4
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-12-07
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.040232
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: FDG‐PET is a validated and widely used sensitive biomarker for neurodegeneration (N) in ageing and Alzheimer's clinical syndrome as highlighted in the β‐amyloid (A)/tau (T)/N scheme (Jack et al ., 2016, 2018). Early‐phase Florbetapir (eAV‐45) appears as a promising proxy for FDG‐PET but needs further validation. Optimization of the time window and preprocessing methods of eAV‐45 are scarce, and no previous study assessed the sensitivity and the reproducibility of this measure in individual cases. Method: We included 191 participants including young to elderly volunteers and Alzheimer's clinical syndrome patients, all of whom had both FDG‐PET and dual‐phase Florbetapir acquired on the same scanner, along with structural MRI and neuropsychological examinations. The time window was first optimized at the vertex level by correlation maximization between eAV‐45 and FDG‐PET, and correlation minimization between eAV‐45 and late AV‐45 (lAV‐45). Then, the six reference regions used for PET intensity scaling were compared by computing metrics of: (i) vertex‐wise pattern overlap between eAV‐45 and FDG‐PET (derived from patients versus controls comparisons and correlation with cognitive scores), and (ii) machine learning classification performance reflecting the discriminative power of both eAV‐45 and FDG‐PET at the individual level. Result: The time interval from 0 to 4 minutes gave optimal vertex‐wise within‐ and inter‐subject correlations of both eAV‐45 versus FDG‐PET (maximal) and eAV‐45 versus lAV‐45 (minimal) (Figure 1). Balanced accuracy of pattern overlap was globally maximal with pons scaling (Figure 2), whereas classification performance between patients' subgroups and controls were similar across scaling regions for both eAV‐45 and FDG‐PET (Figure 3A). Finally, classification performance was significantly superior for combined early plus late AV‐45 compared to FDG‐PET alone (Figure 3B). Conclusion: Results show that eAV‐45 from 0 to 4 minutes with pons scaling is an optimal surrogate of FDG‐PET in ageing and Alzheimer's clinical syndrome. This study highlights the strong potential of optimized dual‐phase AV‐45, allowing to outperform FDG‐PET discriminative power when combining both early plus late AV‐45 information obtained from a single PET‐tracer injection. Interestingly, the use of dual‐phase AV‐45 instead of FDG‐PET plus lAV‐45 will reduce the radiation dose, total time, number of visits and costs.