Please Wait...

Regionally-specific SUVR values in early-stage AD patients screened with amyloid PET

David Scott1, Joyce Suhy1, Mehul Sampat1, Ping Chiao2, Jeff Sevigny2, Gregory Klein1 1BioClinica, Newark, CA, USA; 2Biogen Idec, Cambridge, MA, USA

INTRODUCTION

  • Amyloid deposition in the brain is thought to underlie a neurodegenerative cascade, leading to the development of cognitive decline and conversion to AD. In an ongoing phase 1b multi-center clinical trial (221AD103), amyloid PET is used as a screening tool to identify amyloid-positive participants as an enrichment strategy for clinical trials of Alzheimer's disease (AD).
  • PET imaging with florbetapir measures global cortical amyloid retention, allowing a trained neuroradiologist to classify a participant as positive or negative for amyloid. Each participant's amyloid burden can also be quantified through the computation of a Standard Uptake Value ratio (SUVr). We investigate which areas of the brain are most informative in yielding a positive visual reading for amyloid, and how inter-individual differences in regional SUVr are driven by diagnosis (prodromal vs. mild AD) and genetics (APOE ε4 status).

METHODS

Study Population

  • The study was conducted in the first 279 patients in a phase 1b clinical trial who fulfilled clinical criteria for either prodromal or mild Alzheimer's Disease.

Florbetapir PET Imaging

  • Florbetapir PET data were acquired  from 25 US imaging centers using 18 different PET or PET/CT scanner models manufactured by GE, Philips and Siemens.  Prior to imaging patients, each center completed a qualification process including on-site training to the imaging protocol and Hoffman phantom data acquisition to assess scanner quality and to provide calibration data for spatial resolution normalization1.
  • A 20-minute emission acquisition was acquired starting at 50 minutes after administration of 370 MBq of florbetapir. Florbetapir images were reconstructed using iterative techniques, motion-corrected, smoothed to a uniform resolution of 6.5mm in-plane and 7.5mm axially, and co-registered to T1 MRIs in a standard 91 x 109 x 91 voxel, 2 x 2 x 2 mm common space using a rigid transformation.
  • Visual Reading
  • The binary classification methodology followed guidelines described in the Amyvid™ FDA label. Visual reads were based primarily upon PET image data, while the registered MRI and fused PET/MRI data were used to provide supplemental anatomical information.
  • Scans were independently interpreted by one of two board-certified neuroradiologists trained by  the Avid Amyvid™ process.
  • Readers had the ability to interrogate and review all the data sets including the co-registered MRI/PET data in all three orthogonal orientations.
  • SUVR Method
  • Following PET imaging, semi-quantitative values of cortical to whole cerebellar standard uptake value ratios (SUVr) were computed.
  • Data were defined in native patient space using Freesurfer, and regional SUVr values for the entire parcellation were computed by normalizing to activity in a reference region.
  • Averaged SUVrs were calculated by averaging normalized activity in the frontal, anterior/posterior cingulate, lateral parietal, and lateral temporal composite regions used by the ADNI Berkeley PET Core, and referenced to whole brain, white-matter, eroded white-matter, whole cerebellum, cerebellar grey matter, cerebellar white matter, brainstem, thresholded brainstem, and a composite including whole cerebellum, brainstem and eroded white matter (Figure 1).

RESULTS

Study Population

  • Positive amyloid binding was observed in 61% of study participants, based on visual read.

SUVr Findings: Amyloid binding

  • Within each of 4 whole-brain normalized composite ROIs, highly significant group differences were observed based on visual read result. Based on Cohen's d effect size estimates, the composite frontal ROI best discriminated between visual read amyloid positive and negative subjects. An average of the 4 regional ROIs provided even better separation (figure 1, upper plot).
  • Within the average composite ROI, the selected reference region had a marked impact on group separation. Normalizing SUVrs to a whole brain template produced the best result, closely followed by a composite region including whole cerebellum, brainstem and eroded white matter masks. Normalization to white matter, eroded white matter, cerebellar white matter, brain stem, and thresholded brain stem reference regions all achieved slightly poorer separation. Whole cerebellum and cerebellar grey reference regions produced the worst separation (figure 1, lower plot).
  •  The complete FreeSurfer parcellation allowed effect sizes to be generated for each region of the brain. Again, frontal regions demonstrated the greatest separation between visual read groups, with temporal, cingulate, parietal and other subcortical regions all contributing meaningful information regarding amyloid status (figure 2; color-coded points reflect regions included in the Berkeley ROI set).

SUVr Findings: Amyloid binding

  • A participant's diagnostic category at the time of PET imaging also impacted SUVr assessments of amyloid binding. Whole-brain normalized composite ROIs demonstrated much greater variability at discriminating between mild vs. prodromal AD patients at screening; only the parietal composite and average ROIs yielded a significant difference between groups. As previously, the choice of reference region impacts the result, and follows a similar pattern to the visual read result (figure 3, top plots). In addition to the parietal composite and average ROI, 3 FreeSurfer regions (left precuneus, left paracentral gyrus, and left isthmus of cingulate gyrus) were found to significantly discriminate between diagnostic categories, following Bonferroni-correction for multiple comparisons.
  • Differences in amyloid binding between prodromal and mild AD participants were further amplified when APOE ε4 status was considered (figure 3, bottom plot). Mild AD participants showed similar levels of amyloid binding in the average composite ROI, and were indistinguishable from prodromal AD carriers. However, prodromal AD non-carriers showed significantly reduced amyloid binding compared to the other groups (t = 3.1, two-tailed p < 0.002). The group separation between prodromal APOE ε4 carriers vs. non-carriers was robust and widespread, with frontal, temporal and cingulate regions consistently demonstrating the effect; following Bonferroni-correction, and across the entire FreeSurfer parcellation, 39 regions were found to significantly discriminate between APOE ε4 status in prodromal AD. The bottom right plot in Figure 3 shows the effect in the right parahippocampal gyrus, which showed the greatest group separation of all regions tested.

 

CONCLUSIONS

  • These results confirm that subregions used in previous ADNI composite SUVR calculations show highly significant differences in amyloid tracer uptake between visually positive and negative prodromal and mild AD subjects.
  • Given the weak regional effect sizes seen in some brain areas (notably temporal lobe), these results suggest an optimized selection of FreeSurfer-based regions could improve amyloid positive vs. negative group separation.
  • Visual reading amyloid binding classification appears most robustly captured through an averaged composite SUVr using a whole-brain reference region. The composite ROI effect appears driven primarily by binding differences in frontal regions, though the choice of reference region is also important.
  • Genetic as well as diagnostic factors should be considered when enriching an AD population for clinical trials. Specifically, prodromal AD subjects who are APOE ε4 non-carriers are by far the most likely group to demonstrate negative amyloid binding patterns.

ACKNOWLEDGEMENTS

This study was funded by Biogen Idec.

LEARN MORE OR SPEAK WITH OUR EXPERTS

CONTACT US
Leader in Clinical Trial
Management Solutions

Successful clinical trials require the ability to see key details and uncover hidden insights. Bioclinica utilizes science and technology to bring clarity to clinical trials, helping companies to develop new life-improving therapies more efficiently and safely.

ASCPT 2018 Annual Meeting https://t.co/i0mjYbBq5N Join @Bioclinica Research's Open House 3/22 https://t.co/citlLotbD7
bioclinica (2 days ago)
Heading to DC for DIA's #Pharmacovigilance & Risk Mgmt Strategies Conf.? Visit @Bioclinica in booth 32. Preview our… https://t.co/alU1sCg04l
bioclinica (2 days ago)
Kevin Bishop, named a 'Top Innovator' in #clinicaltrials by @CenterWatch See how we're reducing risk to the clinic… https://t.co/rXBGvIh9Wr
bioclinica (3 days ago)
RT @CenterWatch: Congratulations to all our #2018TopInovators! Read about the innovations in the #CWMonthly special edition, https://t.co/Y…
bioclinica (2 weeks ago)
RT @CenterWatch: .@Pfizer demonstrates leadership in #clinicaltrial innovation with two spots on the @CenterWatch #2018TopInnovators, https…
bioclinica (2 weeks ago)
RT @jheanne: I just contributed towards 'Joby' on https://t.co/EHt69JFLZ9 Best wishes to Joby and all our @Bioclinica team members! @Oxfam…
bioclinica (2 weeks ago)

Latest Blogs:

Removing Risk from Clinical Trial Management System (CTMS) Implementations
Collaboration Between Clinical Operations and the Logistics and Supply Chain Teams is Key to Trial Success
The Value of Protocol Review
CTMS and RBM: Hot Topics at OCT Nordics in Copenhagen