Advertisement

DETERMINING THE UTILITY OF PMOD IMAGE QUANTIFICATION SOFTWARE FOR PROCESSING [11C]PIB AND [18F]FLUTEMETAMOL IMAGES IN A PET-ONLY PIPELINE TO GENERATE SUVR FOR QUANTIFICATION ON THE CENTILOID SCALE

      Background

      Utilising a standardised quantitative imaging analysis scale, Centiloid, for comparing β-amyloid imaging tracers in Alzheimer’s disease (AD) has been promoted by the Global Alzheimer’s Association Information Network (GAAIN), with established process pipelines reported (Rowe et al., 2016). We previously reported data confirming Centiloid pipeline implementation in PMOD Image Quantification Software produced results highly concordant with the GAAIN data. However, these pipelines require both PET and MRI images for processing. MRIs were not obtained for end-of-life subjects in GE Healthcare’s pivotal registration studies for [18F]Flutemetamol Injection. Therefore, for these subjects we evaluated the utility of PMOD for application in PET-only image analysis.

      Methods

      Using Centiloid Volume of Interest (VOI) regions, GAAIN [11C]PiB images (YC-0, n=34; AD-100, n=45) were processed using a PET-only method. Standard Uptake Value ratios (SUVr’s) and Centiloid values for the VOI regions were calculated, and compared with published values by way of validation (Klunk et.al., 2015). The utility of PMOD’s PET-only method was assessed by cross-compared to the PET/MRI method using the Centiloid standard CTX VOI. Subsequently, subjects from GE Healthcare’s pivotal autopsy studies were evaluated using PMOD’s PET-only method. Tissue-based autopsy measures were used in a Receiver Operating Characteristic (ROC) curve analysis to estimate the Centiloid value corresponding to boundaries used in CERAD pathology assessments in regards to AD classification.

      Results

      For all subjects, we found the [11C]PiB SUVr values were within recommended limits of the published set for the key reference region (whole cerebellum). When converted to Centiloid values, there was a high degree of correlation between the GAAIN data and values derived from PMOD. PET-only processing was applied to the autopsy-cohort [18F]flutemetamol images. Centiloid values were compared to the tissue based neuritic plaque assessments. ROC analysis determined the CERAD designation boundary’s in Centiloid units.

      Conclusions

      PET-only image processing, such as the one established using PMOD, provide an alternate method of Centiloid processing, and is suitable for assessing subjects without MRI. The SUVr and Centiloid values obtained from our autopsy cohort proved suitable for determining the CERAD pathological boundaries in Centiloids equivalent units, indicating that [18F]flutemetamol Centiloid values can be related to CERAD style stratification.