Plausibility of early and short acquisition windows for 18 F‐NAV4694 and 18 F‐florbetapir PET to expand clinical throughput

Résumé fourni par la source

Abstract Background Monoclonal antibody treatments may lead to increased demand for diagnostic Aβ PET. Simultaneously, high costs and limited reimbursement are pushing clinicians to alternatives such as plasma biomarkers, despite their inability to quantify amyloid burden. Seeking to reduce cost and improve productivity, we assessed the impact of early and short 18 F‐NAV4694 (NAV) and 18 F‐Florbetapir (FBP) PET scan protocols on Centiloid quantification. Method 38 dynamic NAV scans (200±10% MBq) on a Philips Allegro from the AIBL study and 454 dynamic FBP scans (370±10% MBq) on various scanners from the OASIS‐3 study were collected. Centiloid was quantified using the CapAIBL PET‐only software, using the standard 50‐70 minute window and three abbreviated windows of 50‐60, 40‐50 and 30‐40 minutes. A linear association was assumed and estimated between shortened and standard windows, validated with Bland‐Altman analysis. After applying the estimated linear correction, Centiloid error was assessed by mean absolute error (MAE) and 95% range of error (1.96σ). Result Figures 1 and 2 (top) depict NAV and FBP, respectively, with fit parameters and metrics summarised in Table 1. Shortened windows correlated strongly with the standard windows (R 2 >0.95), with correlation decreasing for earlier windows. Bland‐Altman analyses revealed residual error was uncorrelated with Amyloid burden and consistent with a linear association for both NAV and FBP (Figures 1 and 2, bottom). NAV at 30‐40 minutes introduced a mean error of 2.7 CL (95% <7.4 CL), or a relative error in SUVR units of 2.1%, comparable to published test‐retest results available for florbetapir (2.4%). This NAV window allows a 50% scan‐time and 40% uptake‐time reduction. For FBP, the 30‐40 minute window yielded higher error (mean 5.4 CL, 95% <14.1 CL), so a 40‐50 minute window (mean 3.3 CL error, 95% <8.5 CL) may be more appropriate. Conclusion Acquiring NAV at 30‐40 minutes or FBP at 40‐50 minutes, with linear correction, is unlikely to change clinical management. Thus, 50% scan time reduction and 40% (NAV) or 20% (FBP) uptake time reduction is achievable. Importantly, the tracers have not reached steady‐state at early windows, so accurate timing of the acquisition window is essential.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.