Non-invasive modelling and parametric methods for quantification of MAO-B activity using [¹¹C]L-deprenyl-D2 PET

The PET radioligand [¹¹C]L-deprenyl-D2 binds to monoamine oxidase B (MAO-B), enabling studies of neuroinflammation in various neurological disorders. This study aimed to evaluate non-invasive and parametric methods for quantifying [¹¹C]L-deprenyl-D2 PET. Twenty-eight subjects underwent 60 min dynamic [¹¹C]L-deprenyl-D2 PET. The irreversible binding parameter λk₃ and the net influx rate relative to the non-displaceable volume of distribution K_ND were estimated with an irreversible two-tissue plasma-input compartment model (2T3k). Non-invasive quantification of K_ND was performed using cerebellum corrected for specific binding with previously suggested corrections as reference tissue. These corrections were further optimised by comparison to the 2T3k non-displaceable signal in cerebellum. Parametric images were computed using a basis function implementation of the 2T3k model, reference Patlak methods and the reduced reference tissue model. Additionally, a semi-quantitative specific binding index (SBI) was calculated as the ratio of the mean activity concentration at 30-40 or 50-60 min p.i. in target tissues to the peak cerebellum uptake. Parametric λk₃ images agreed well with VOI-based nonlinear regression. Modified cerebellar reference models and SBI demonstrated poor-to-moderate correlations with arterial input-based analysis and yielded false positive differences in voxel-wise group comparisons. Full kinetic analysis with arterial input is required for accurate assessment of MAO-B activity with [¹¹C]L-deprenyl-D2.