Four year Longitudinal Assessment of DAT Imaging Biomarkers in a Progressing Parkinson Disease Cohort: Analysis Strategies and Implications for Treatment Trial Design

Objectives: Imaging measures of presynaptic dopaminergic function have been proposed as Parkinson Disease (PD) progression biomarkers offering a more objective measure of disease status in assessment of disease modifying therapies. The Parkinson Progression Marker Initiative (PPMI) study is an international, multicenter longitudinal evaluation of biomarkers in a progressing early Parkinson's disease cohort. The aims of this study are to evaluate within subject changes in dopamine (DAT) using 123-I Ioflupane SPECT in PD participants studied over four years employing different analysis strategies for enhancing signal:noise (S:N) of the change signal in different striatal subregions and statistical power for detecting slowing/cessation of signal loss in clinical trials.

Methods: In an ongoing study 259 de novo PD subjects completed baseline,1yr, 2yr, and 4yr clinical and ¹²³I-Ioflupane/SPECT evaluation. Changes in specific binding ratio (SBR) between baseline and 2-yr, 3-yr, and 4-yr scans were calculated for ipsilateral and contralateral striatum as well as striatal subregions. In addition, monoexponential curve fitting was performed in all subregions and composite striatal SBRs, absolute and percent loss and area under curve calculated and compared to standard analyses for change in SBR relative to variance (S:N) and robustness of correlation with clinical ratings of PD symptoms Results: Longitudinal follow-up of the PPMI cohort demonstrates annualized rates of % change in DAT composite SBR to be10-12% in the first year with lower reductions in SBR by year 4, for an average annualized change of 7%. The SBR change signal is roughly identical between the two methods within each time cohort. However, the exponential analysis in both 2 and 4 year groups has superior S:N over standard methods.Similarly, correlational analyses vs UPDRS motor scores are more robust in the exponential fit analysis compared to the standard

Methods: Finally sample size estimates for detecting a 50% reduction in the rate of signal loss are 55-60% lower in the exponential fit method than the standard method. Conclusions: The use of 123-I ioflupane SPECT as a biomarker of PD progression is feasible from the perspective of the size of signal change in an early PD cohort studied over 2-4 years with best S:N properties in the exponential fit to the serial regional SBRs resulting in better correlation with motor scores and lower sample size for achieving similar power compared to the standard analysis method.