Authors: Willemijn J. Jansen, Valentina Ghisays, Kathryn L. DeMarco, Connie A. Boker, Kewei Chen, Yinghua Chen, Ji Luo, Hillary D. Protas, Tim West, Matthew Meyer, Kris Kirmess, Philip B. Verghese, Helen Hu, Kevin E. Yarasheski, Yi Su, Eric M. Reiman
Published: 2020-12-07
DOI: 10.1002/alz.040332
Source: Full article
AbstractBackgroundPlasma amyloid‐β (Aβ) blood tests have great promise in research and care. C2N Diagnostics has developed a mass spectrometry‐based plasma Aβ42/40 assay (APTUSTM‐ pAβ42/40), which has previously demonstrated approximately 88% accuracy (81% sensitivity and 83% specificity; CTAD 2019) to distinguish between individuals with Aβ‐positive and negative (A+ and A‐) PET scans. We sought to characterize the assay’s ability to discriminate cognitively unimpaired APOE4 homozygotes, heterozygotes, and non‐carriers from the Arizona APOE Cohort.MethodsPlasma Aβ42/40 concentration ratios and APOE4 gene dose (0, 1, 2 alleles) were characterized blindly using 0.5 cc plasma aliquots from 121 cognitively unimpaired 66±8 (47‐85) year‐old participants, including 21 APOE4 homozygotes, 37 heterozygotes and 63 non‐carriers whose blood samples and PiB or florbetapir PET scans were acquired at the same visit. Neuropathologically validated PiB 1.47 and florbetapir 1.17 SUVR cut‐offs were used to determine A+ and A‐ PET scans. Receiver operating characteristic (ROC) analyses were performed calculating the areas‐under‐the curve (AUCs) before and after adjustment for APOE4 gene status or dose, age and sex. The optimal Aβ42/40 cut‐off, sensitivity and specificity were retrospectively characterized.ResultsPlasma Aβ42/40 ratios discriminated between individuals with A+ and A‐ PET scans with 0.87 and 0.89 AUCs, respectively, before and after adjustment for APOE4 status. Further adjustment for APOE4 gene dose, age and sex did not significantly influence these results. In the APOE4 status‐adjusted ROC, a plasma Aβ42/40 0.096 cutoff, with 91% sensitivity and 82% specificity had the maximum Youden Index. We observed trends of a greater percentage of A‐ PET APOE4 homozygotes than heterozygotes or non‐carriers being plasma A+ (29%, vs. 11%; p=0.08, and vs. 11%; p=0.05, respectively). Further, individuals with an A+ plasma test but an A‐ PET scan at baseline were at 9‐fold increased risk of conversion to A+ PET compared to individuals with an A‐ plasma test.ConclusionAPOE4‐adjusted plasma β42/40 assays have great promise in AD research, prevention trials and clinical care, with demonstrated high accuracy for cognitively unimpaired persons with two, one or no copies of the APOE4 allele. This study suggests that an A+ plasma test result predicts progression from an A‐ to an A+ PET scan in the near future.