QIBA SPECT I-123 Biomarker Committee. Nov 16, 2018

Transcription

1 QIBA SPECT I-123 Biomarker Committee Nov 16, 2018

2 Agenda Roll call Profile 2.0 update, timelines, and strategy Ancillary activities PPMI data relevant to the profile AOB 2 min 10 min 5 min 25 min remaining time

3 Where are we? Profile: Version 1.0 was released for public comment. Each suggested revision was addressed by the BC and resolved. The committee s new goal is to provide an updated Version 2 by the end of 1Q Checklist: Each of the performance requirements in the Profile has been compiled as a set of checklists. These lists have been developed as tools to help actors and imaging sites evaluate their work for conformance with the Profile. Feasibility Testing: The checklists are being used as quality control tools to assess the ability (or practicality/willingness) of actors to perform each of the Profile s performance specifications. The results of these feasibility tests will then be used to streamline and tighten the Profile performance requirements.

4 Measuring PD Progression Is Problematic SLOW, VARIABLE COURSE Clinical measures of Parkinson s disease progression suggest a process which is slow with significant variability both between patients and also within the individual patient over the course of disease MEDICATION CONFOUND Assessment of disease progression is confounded by effects of antiparkinson treatments and difficulties achieving complete wash-out of drug effects PHENOMENOLOGICAL COMPLEXITY Multi-dimensional, complex, and changing clinical picture- what to track; motor scores, ADL s, milestones?

5 PPMI Study Details: Synopsis Study population Assessments/ Clinical data collection 423 de novo PD subjects (newly diagnosed and unmedicated) 196 age- and gender-matched healthy controls 64 SWEDD 67 Prodromal - Olfactory/RBD 250 LRRK2 - PD manifest and non-manifesting family members 250 GBA- PD manifest and non-manifesting family members 100 SNCA - PD manifest and non-manifesting family members Subjects will be followed through 2018 Motor assessments Neurobehavioral/cognitive testing Autonomic, Olfaction, Sleep DaTSCAN, AV133, Amyloid, DTI/RS MRI Biologic collection/ Data and Biosamples shared on website - DNA, RNA, IPSC Serum and plasma collected at each visit; urine collected annually CSF collected at baseline, 6mo 12 mo and then annually Samples aliquotted and stored in central biorepository > 800,000 Data downloads > 100 Sample requests via BRC Ancillary study development

6 PPMI is sponsored and partially funded by The Michael J. Fox Foundation for Parkinson s Research. Other funding partners include a consortium of industry players, non-profit organizations and private individuals.

7 PPMI IMAGING STUDIES Nuclear Medicine Techniques Ioflupane SPECT (DaTScan)- Dopamine transporter AV133 PET- VMAT2, vesicular transporter Florbetaben PET- Amyloid deposition Magnetic Resonance Techniques MRI- Diffusion Tensor Imaging, DTI maps integrity of brain connections Resting state MRI- describes functional connections of regions of the brain MRI T1- provides brain gross anatomy

8 Two and Four Year Longitudinal Assessment of DAT Imaging Biomarkers in a Progressing Parkinson Disease Cohort: Implications for Clinical Trial Design John P Seibyl, MD on behalf of the PPMI Investigators Institute for Neurodegenerative Disorders, and Invicro, New Haven, United States

9 Measuring DAT changes in de novo PD with 123-I Ioflupane SPECT over four years in PPMI RATIONALE: Prior studies show loss of striatal signal Parkinson's patients studied longitudinally with 123-I Ioflupane SPECT. These studies demonstrate annual loss approximately 7 to 10% of SBR per year, but with significant between subject variance. The purpose of the present investigation was to evaluate different analytic approaches in a large PD cohort studied over four years with serial DAT SPECT.

10 METHODS 343 PD patients in PPMI had serial ioflupane SPECT scans at baseiine, 1, and 2 years post enrollment, 282 PD patients had an additional 4 year scan Employed small and large region of interest template previously described for developing regional specific binding ratios (SBR). Strategies to measure SBR change follows two approaches; delta SBR and % change SBR from baseline.

11 Baseline DAT SBR, Age-corrected Mean Striatal SBR Contralateral Putamen SBR PD n= 423 HC n= 196 SWEDD = 64

12 RESEARCH QUESTIONS: 1. What is the better outcome measure of serial DAT change in PD; a delta SBR or a percent change SBR? 2. What is the better striatal sampling strategy; small or large ROIs? 3. Do analytic strategies incorporating curve fitting applied to serial within subject longitudinal SBR data increase the signal size and reduce the variance compared with standard baseline-follow-up SBR change measures? 4. What are the implications of scan analysis method on clinical therapeutic trial design and sample size estimates?

13 SBR Mean (SD) SBR by REGION 3 2 Ipsilateral = blue Contralateral =red 1 Y0 N= Ip Caud Y0 C Caud Y0 Ip Ant Put Y0 C ant Put Y0 Ip Put Y0 C Put Y0 Region Ip Stria Y0 C Stria Y0 Mean Stria Y0

14 Ip Caud Y0 Ip Ant Put Y0 Ip Put Y0 Ip Stria Y0 Mean Stria Y0 Ip Caud Y1 Ip Ant Put Y1 Ip Put Y1 Ip Stria Y1 Mean Stria Y1 Ip Caud Y2 Ip Ant Put Y2 Ip Put Y2 Ip Stria Y2 Mean Stria Y2 Ip Caud Y4 Ip Ant Put Y4 Ip Put Y4 Ip Stria Y4 Mean Stria Y4 SBR Longitudinal Mean (SD) SBR by REGION 3 SBR small ROI 2 1 Y 4 N= Region

15 Ip Caud Y0 Ip Ant Put Y0 Ip Put Y0 Ip Stria Y0 Mean Stria Y0 Ip Caud Y1 Ip Ant Put Y1 Ip Put Y1 Ip Stria Y1 Mean Stria Y1 Ip Caud Y2 Ip Ant Put Y2 Ip Put Y2 Ip Stria Y2 Mean Stria Y2 Ip Caud Y4 Ip Ant Put Y4 Ip Put Y4 Ip Stria Y4 Mean Stria Y4 SBR Baseline SBR:Large v Small ROIs Y 4 N= 285 SBR small ROI SBR large ROI Region

16 Ip Caud Y1 C Caud Y1 Ip Ant Put Y1 C ant Put Y1 Ip Put Y1 C Put Y1 Ip Stria Y1 C Stria Y1 Mean Stria Y1 Ip Caud Y2 C Caud Y2 Ip Ant Put Y2 C ant Put Y2 Ip Put Y2 C Put Y2 Ip Stria Y2 C Stria Y2 Mean Stria Y2 Ip Caud Y4 C Caud Y4 Ip Ant Put Y4 C ant Put Y4 Ip Put Y4 C Put Y4 Ip Stria Y4 C Stria Y4 Mean Stria Y4 delta SBR Ip Caud Y1 C Caud Y1 Ip Ant Put Y1 C ant Put Y1 Ip Put Y1 C Put Y1 Ip Stria Y1 C Stria Y1 Mean Stria Y1 Ip Caud Y2 C Caud Y2 Ip Ant Put Y2 C ant Put Y2 Ip Put Y2 C Put Y2 Ip Stria Y2 C Stria Y2 Mean Stria Y2 Ip Caud Y4 C Caud Y4 Ip Ant Put Y4 C ant Put Y4 Ip Put Y4 C Put Y4 Ip Stria Y4 C Stria Y4 Mean Stria Y4 delta SBR Delta SBR 1, 2 & 4 Yr Delta SBR small Delta SBR large Y 4 N= Delta SBR 1, 2 & 4 Yr Delta SBR small Delta SBR large Region Y 4 N= Region

17 Ip Caud Y1 C Caud Y1 Ip Ant Put Y1 C ant Put Y1 Ip Put Y1 C Put Y1 Ip Stria Y1 C Stria Y1 Mean Stria Y1 Ip Caud Y2 C Caud Y2 Ip Ant Put Y2 C ant Put Y2 Ip Put Y2 C Put Y2 Ip Stria Y2 C Stria Y2 Mean Stria Y2 Ip Caud Y4 C Caud Y4 Ip Ant Put Y4 C ant Put Y4 Ip Put Y4 C Put Y4 Ip Stria Y4 C Stria Y4 Mean Stria Y4 Percent change SBR Ip Caud Y1 C Caud Y1 Ip Ant Put Y1 C ant Put Y1 Ip Put Y1 C Put Y1 Ip Stria Y1 C Stria Y1 Mean Stria Y1 Ip Caud Y2 C Caud Y2 Ip Ant Put Y2 C ant Put Y2 Ip Put Y2 C Put Y2 Ip Stria Y2 C Stria Y2 Mean Stria Y2 Ip Caud Y4 C Caud Y4 Ip Ant Put Y4 C ant Put Y4 Ip Put Y4 C Put Y4 Ip Stria Y4 C Stria Y4 Mean Stria Y4 Percent change SBR % Change SBR 1, 2 & 4 Yr 40 % Change small % Change large 20 Y 4 N= % Change SBR 1, 2 & 4 Yr Region % Change small % Change large Y 4 N= 285 Region

18 Ip Caud Y1 C Caud Y1 Ip Ant Put Y1 C ant Put Y1 Ip Put Y1 C Put Y1 Ip Stria Y1 C Stria Y1 Mean Stria Y1 Ip Caud Y2 C Caud Y2 Ip Ant Put Y2 C ant Put Y2 Ip Put Y2 C Put Y2 Ip Stria Y2 C Stria Y2 Mean Stria Y2 Ip Caud Y4 C Caud Y4 Ip Ant Put Y4 C ant Put Y4 Ip Put Y4 C Put Y4 Ip Stria Y4 C Stria Y4 Mean Stria Y4 S:N% change SBR Ip Caud Y1 C Caud Y1 Ip Ant Put Y1 C ant Put Y1 Ip Put Y1 C Put Y1 Ip Stria Y1 C Stria Y1 Mean Stria Y1 Ip Caud Y2 C Caud Y2 Ip Ant Put Y2 C ant Put Y2 Ip Put Y2 C Put Y2 Ip Stria Y2 C Stria Y2 Mean Stria Y2 Ip Caud Y4 C Caud Y4 Ip Ant Put Y4 C ant Put Y4 Ip Put Y4 C Put Y4 Ip Stria Y4 C Stria Y4 Mean Stria Y4 S:N Delta SBR S:N Delta SBR 1, 2 & 4 Yr delta S:N lg delta S:N sm S:N = Mean/SD Y 4 N= 285 S:N % Change SBR 1, 2 & 4 Yr Region %change S:N small %change S:N large S:N = Mean/SD Y 4 N= 285 Region

19 Ip Caud Y1 C Caud Y1 Ip Ant Put Y1 C ant Put Y1 Ip Put Y1 C Put Y1 Ip Stria Y1 C Stria Y1 Mean Stria Y1 Ip Caud Y2 C Caud Y2 Ip Ant Put Y2 C ant Put Y2 Ip Put Y2 C Put Y2 Ip Stria Y2 C Stria Y2 Mean Stria Y2 Ip Caud Y4 C Caud Y4 Ip Ant Put Y4 C ant Put Y4 Ip Put Y4 C Put Y4 Ip Stria Y4 C Stria Y4 Mean Stria Y4 S:N% change SBR S:N % Change & Delta SBR 1, 2 & 4 Yr %change S:N small %change S:N large delta S:N sm delta S:N lg Y 4 N= 285 S:N = Mean/SD Region

20 Sample size Sample size by region Std 1 y d Std 1 y % Std 2 y d Std 2 y % Std 4 y d Std 4 y % IPSI CAUD CONTRA CAUD IPSI A P CONTRA AP IPSI PUT CONTRA PUT IPSI STRIA CONTRA STRIA MEAN STRIA ASSUMES: 50% effect size 80% power, p<0.05, 2-tailed

21 SBR ANALYSIS METHODS Method 2 Exponential fit Delta and % change from equation Method 1 Delta SBR= - SBR(y0)-SBR(y4) % change = delta SBR/SBR(y0) Years post Baseline Scan Compare variance and signal:noise for each measure as well as strength of correlation with motor symptoms and power analyses for detecting change in clinical trials.

22 Sample size Sample size by region IPSI CAUD CONTRA CAUD IPSI A P CONTRA AP IPSI PUT CONTRA PUT IPSI STRIA CONTRA STRIA MEAN STRIA Std 1 y d Std 1 y % Std 2 y d Std 2 y % Std 4 y d Std 4 y % Exp 4 y % Exp 4 y d Exp2 y % Exp2 y d ASSUMES: 50% effect size 80% power, p<0.05, 2-tailed

23 Culm. % Change SBR A Model for Non-linear Reductions in Regional Striatal SBR Change Predicts Convergence of % SBRs Over Time caudate ant. putamen putamen Time A non-linear pattern of DAT loss, such as in the simple exponential curves depicted above, identical across all striatal subregions, but with a phase shift to the right from posterior putamen to anterior putamen to caudate can explain the regional differences in the rates of specific binding loss seen in this study and their subsequent equalization. When assayed at time 1, there are regional differences in the absolute SBR, and those regions with highest SBRs are also on the fastest portion of the elimination curve. At time 2, the absolute differences in regional SBRs and the percent change per year are diminishing and converging.

24 S B R S B R Regional striatal SBR based on based on monexponential curve fits of 4 year, 4 scan data extrapolating 6 y prior and 11 y post baseline shows similar curves apparently phase shifted M e a n c u r v e s ( n = 2 7 1, 4 Y d a t a ) M e a n c u r v e s ( n = 2 7 1, 4 Y d a t a ) 4 I P C A 4 I P C A S h i f t ( y ) 0 3 C C A I P A P 3 C C A I P A P C A P I P P 2 C A P I P P C P C P Y E A R S Y E A R S Shifting the curves back to right such that they overlap results in an estimate of the years that region is ahead in the neurodegenerative process relative to ipsilateral caudate.

25 Conclusions These multicenter, multicamera data are consistent with other smaller studies indicating a 6-10% SBR loss/yr and exponential patterns of DAT signal reduction. Striatal sub-regions have different signal:noise characteristics for measures of SBR change over four years with contralateral putamen the lowest and mean striatum and ipsilateral putamen the highest and better for tracking treatment-induced slowing of DaT loss. Large and small ROI strategies are generally equivalent with regard to signal to noise, but smaller ROIs provide additional, relevant striatal subregion information. Percent change SBR is superior to delta SBR in tracking DaT signal loss. Correlation with motor UPDRS (data not shown) is moderate but significant and highest for methods using nonlinear fits. Power analyses are most robust for data analyzed with exponential fitting.

26 Agenda Roll call Profile 2.0 update, timelines, and strategy Ancillary activities PPMI data relevant to the profile AOB 2 min 10 min 5 min 25 min remaining time 26