Optimisation of Q.Clear image reconstruction for [¹⁸F]flutemetamol in the brain using a PET-MR scanner

Abstract

Objective: This study evaluates the benefits of Q.Clear reconstruction in the amyloid application and investigates the ability of Q.Clear to reduce the count data of [¹⁸F]FMM using clinical and phantom datasets in a PET-MR scanner. Materials and methods: Esser PET phantom and [¹⁸F]FMM of seven brain datasets for elderly healthy participants were reconstructed using OSEM, OSEM-TOF, OSEEM-TOF-PSF, and Q.Clear algorithms with β values of 100, 150, 200, 250, and 300 over three frame durations consisting of twenty, five and one minute. The Standardised Uptake Value Ratio (SUVR) of the cingulate (posterior and interior), hippocampus, and the whole cerebellum (as reference) were measured and identified on co-registered MRI. The Contrast Recovery Coefficient (CRC), Background Variability (BV), and Image Roughness (IR) were calculated for the phantom quantitation. Results: Visually, Q.Clear with a higher β value (300 ≥ ) was found to be more effective. Quantitatively, Q.Clear with a lower (100 ≥ ) value was found to be preferable for reducing the spilling effect. Furthermore, it was identified that Q.Clear reconstruction may help to reduce the required dose. Conclusion: The Q.Clear reconstruction algorithm was found to perform more effectively than conventional reconstruction. Furthermore, the Q.Clear reconstruction may also help to reduce the radiation dose in the amyloid application, although this will require additional investigation.