摘要
Background: We previously developed an automated ensemble SSM/PCA framework for differential diagnosis of parkinsonism, combining patterns of dopamine transporter (DAT) availability, measured as specific binding ratio (SBR), and relative cerebral blood flow (R1) from dynamic ¹¹C-PE2I PET. How SBR and R1 pattern expression relate to each other across diagnoses, and to regional striatal DAT loss, remains unknown. We examined whether the SBR and R1 relationship differs by diagnosis, and how both expression scores relate to caudate and putamen SBR, their ratio, and striatal asymmetry.
Methods: Patterns for Parkinson’s disease (PD) and dementia with Lewy bodies (DLB) versus healthy controls (HC) were derived across 100 Monte Carlo seeds from 289 patients and 47 HCs. In validation subjects, SBR and R1 expression of their own disease pattern were correlated with each other and with regional SBR measures. Age-adjusted correlation coefficients were averaged across seeds using Fisher z-transformation; mean r is reported.
Results: SBR and R1 expression were negatively correlated in PD (r=-0.27), but positively correlated in DLB (r=0.28). In PD, SBR and R1 expression were related to the caudate/putamen ratio in opposite directions (r=0.59 and r=-0.29, respectively). Higher SBR expression was associated with lower putamen SBR (r=-0.35) and greater asymmetry (r=0.28), whereas higher R1 expression was associated with lower caudate SBR (r=-0.28). Post hoc adjustment for the ratio weakened the PD SBR-R1 correlation to r=-0.15. In DLB, higher SBR and R1 expression were both associated with lower caudate SBR (r=-0.49 and r=-0.42, respectively). Figure 1 shows distributions and IQRs.
Conclusions: In PD, SBR expression reflected putamen-predominant, asymmetric DAT loss, whereas R1 expression reflected caudate involvement. The caudate/putamen ratio partly explains their negative correlation, which may reflect disease stage or a caudate-involved phenotype. In DLB, both scores tracked caudate DAT loss, consistent with greater caudate involvement reported in DLB. Medication effects cannot be excluded.
