Impact of reduced injected dose on semi‐quantification of in vivo amyloid‐β pathology using [18F]flutemetamol

Abstract

AbstractBackgroundReduced injected doses in PET examinations result in lower radiation exposure, reduced acquisition times, reduced costs for institutions and ethically allow for additional scans within the same subject. Previous research has demonstrated that the injected dose of several different PET tracers can be reduced without substantial effects on quantitative outcome measures. Here, we studied the effect of reduced injected doses on signal‐to‐noise ratios (SNRs) and standardised uptake value ratios (SUVRs) for the FDA‐approved amyloid‐β (Aβ) PET tracer [18F]flutemetamol.MethodThirty [18F]flutemetamol PET scans from previously classified Aβ‐positive and Aβ‐negative individuals were included in this study (Table 1). Amyloid positivity was derived using a validated centiloid pipeline. Subsets of the PET data containing 70%, 50%, 40% 30% and 20% of the total events was extracted randomly from the list‐mode data to generate sinograms and reconstructed using e7 tool. SUVRs were computed for a whole brain region‐of‐interest from an intracranial volume mask using whole cerebellum as a reference region. Reference and target ROIs were derived in MNI space. Next, group separation between Aβ‐positive (A+) and Aβ‐negative (A‐) using mean SNRs was computed using formula 1. Additionally, the differences in SUVR and SNR for true and the respective reduced injected dose were calculated.Group Separation= (meanA+ ‐ meanA‐)/(0.5√((SDA+)2+(SDA‐)2) (1)ResultInitial results (Table 2) show an absolute change in group separation from 0.65 to 0.62 (4.8% change) for original injected dose and a reduction of injected dose to 20% respectively. Whole cortical SUVRs changed by a maximum of 0.01 (<1% change) in both Aβ‐positive and ‐negative individuals. Figure 1 shows scans of a representative Aβ‐positive individual derived from true and different simulated reduced injected doses.ConclusionPreliminary findings in our pilot sample suggest that lower injected tracer‐doses for [18F]flutemetamol yield robust SUVRs at injected doses as low as 20% of the original dose. We will report further validation in a larger dataset. In addition, a whole cortex region was used for SUVR derivation, and investigations of smaller ROIs will be added. The effect of scanner and reconstruction settings as well as the effect of reduced injected doses on separating Aβ‐positive from ‐negative individuals will be investigated.