Amyloid Measures Predictably at Odds on Ends of Scales

Two measures of Aβ are used routinely to identify people who might be at risk for Alzheimer's disease—a drop in soluble Aβ42 in the cerebrospinal fluid (CSF), and the uptake of ligands that bind amyloid plaques in the brain. How do the two correlate? Very closely at the middle ranges of the scales, according to a paper in the March 30 JAMA Neurology. Scientists led by Jon Toledo and senior author John Trojanowski at the University of Pennsylvania, Philadelphia, report that the relationship between the two is weaker at the high and low ends of the scales. Experts interviewed for this article agreed that the result is to be expected. Victor Villemagne, University of Melbourne, Australia, explained that most of the dynamic range for CSF is at the low end of the scale, while most of the dynamic range for amyloid imaging is at the high end. “Instead of trying to force a linear correlation, a non-linear fit is more appropriate,” he told Alzforum.

Some previous smaller studies had suggested that CSF Aβ42 and PET imaging of parenchymal Aβ deposits correlated linearly (see Weigand et al., 2011; Jagust et al., 2009; Grimmer et al., 2009), though larger ones had hinted that the two instead had a non-linear relationship (see Landau et al., 2013; Fagan et al., 2014). To take a closer look, Toledo and colleagues examined data from 820 participants from the Alzheimer’s Disease Neuroimaging Initiative (ADNI)—259 healthy controls, 415 people with mild cognitive impairment (MCI), and 146 with Alzheimer’s disease (AD). All had CSF Aβ42 measurements and florbetapir-PET Aβ imaging done within 30 days of each other, and 150 of them had a second PET scan and CSF samples taken after two years. Included in the two-year follow up were 53 healthy people, 90 with MCI, and seven with AD.

To check how the two measures correlated, the researchers plotted baseline CSF Aβ42 values against those of florbetapir PET standardized uptake value ratios (SUVR). The relationship poorly fit a linear curve, but followed either a hyperbolic or multivariate adaptive regression splines (MARSs) model (see image below), indicating the two biomarkers exhibited a complex relationship. On the two extreme ends of the scales, neither measure predicted the other. Early in disease, when PET SUVRs were low, CSF Aβ values covered a wide range. Late in disease, amyloid PET SUVR continued to rise even after CSF Aβ levels stopped falling. The data suggest that CSF changes are more sensitive to early amyloid accumulation, while only PET reflects continued buildup late in disease, Toledo said. 

A Curvy Fit: The relationship of CSF Aβ and amyloid PET measures is hyperbolic, not a linear. Dotted lines represent the fit for people with no ApoE4 allele, and the solid ones, one copy. [Image courtesy of Toledo et al. Copyright © (2015) American Medical Association. All rights reserved.]

Henrik Zetterberg, University of Gothenburg, Sweden, said it makes sense for CSF Aβ and amyloid PET to disagree at opposite ends of the scales. People with high CSF Aβ1-42 and negative amyloid PET are likely truly Aβ-negative, so SUVR signals in that group would represent noise, he explained. The floor effect of CSF Aβ1-42 in individuals with high SUVR hints that CSF Aβ1-42 concentrations are influenced by other factors, such as Aβ production and clearance rates, he said. 

At intermediate values, CSF Aβ and amyloid PET measures correlated more tightly. ApoE genotype modulated this relationship, in that one or two copies of the ApoE4 allele corresponded to greater PET SUVRs for the same amount of CSF Aβ. ApoE4 is associated with formation of neuritic plaques, which may explain why the allele leads to higher SUVR levels for a given level of CSF Aβ, the authors wrote.

Does this lack of correlation between two widely used measures of Aβ have practical ramifications? Many studies agree that either CSF Aβ or PET imaging are good markers of brain amyloid (for a review, see Blennow et al., 2015). These data hint that each biomarker may be most useful at tracking changes at a specific point in disease progression, said Toledo. CSF might work best early, while amyloid PET may better reflect later disease progress. This may relate to PET ligands' poor affinity for diffuse plaques, which develop early in AD, the authors wrote. Though diffuse, these plaques retain Aβ, meaning less of the peptide shows up in the CSF. Zetterberg agreed, however, he advocated using the two measures in parallel until researchers know more, and some trials currently use both. Toledo said this may not always be possible. “With limited money to spend on diagnostic procedures, clinicians may eventually need to choose one biomarker over the other, so it’s important to show whether they are complementary or not,” he said.—Gwyneth Dickey Zakaib

Comments

No Available Comments

Make a Comment

To make a comment you must login or register.

References

Paper Citations

  1. Weigand SD, Vemuri P, Wiste HJ, Senjem ML, Pankratz VS, Aisen PS, Weiner MW, Petersen RC, Shaw LM, Trojanowski JQ, Knopman DS, Jack CR, . Transforming cerebrospinal fluid Aβ42 measures into calculated Pittsburgh Compound B units of brain Aβ amyloid. Alzheimers Dement. 2011 Mar;7(2):133-41. PubMed.
  2. Jagust WJ, Landau SM, Shaw LM, Trojanowski JQ, Koeppe RA, Reiman EM, Foster NL, Petersen RC, Weiner MW, Price JC, Mathis CA, . Relationships between biomarkers in aging and dementia. Neurology. 2009 Oct 13;73(15):1193-9. PubMed.
  3. Grimmer T, Riemenschneider M, Förstl H, Henriksen G, Klunk WE, Mathis CA, Shiga T, Wester HJ, Kurz A, Drzezga A. Beta amyloid in Alzheimer's disease: increased deposition in brain is reflected in reduced concentration in cerebrospinal fluid. Biol Psychiatry. 2009 Jun 1;65(11):927-34. PubMed.
  4. Landau SM, Lu M, Joshi AD, Pontecorvo M, Mintun MA, Trojanowski JQ, Shaw LM, Jagust WJ, . Comparing PET imaging and CSF measurements of Aß. Ann Neurol. 2013 Mar 28; PubMed.
  5. Fagan AM, Xiong C, Jasielec MS, Bateman RJ, Goate AM, Benzinger TL, Ghetti B, Martins RN, Masters CL, Mayeux R, Ringman JM, Rossor MN, Salloway S, Schofield PR, Sperling RA, Marcus D, Cairns NJ, Buckles VD, Ladenson JH, Morris JC, Holtzman DM, Dominantly Inherited Alzheimer Network. Longitudinal change in CSF biomarkers in autosomal-dominant Alzheimer's disease. Sci Transl Med. 2014 Mar 5;6(226):226ra30. PubMed.
  6. Blennow K, Mattsson N, Schöll M, Hansson O, Zetterberg H. Amyloid biomarkers in Alzheimer's disease. Trends Pharmacol Sci. 2015 May;36(5):297-309. Epub 2015 Apr 1 PubMed.

Further Reading

Papers

  1. Blennow K, Mattsson N, Schöll M, Hansson O, Zetterberg H. Amyloid biomarkers in Alzheimer's disease. Trends Pharmacol Sci. 2015 May;36(5):297-309. Epub 2015 Apr 1 PubMed.
  2. Fagan AM, Mintun MA, Mach RH, Lee SY, Dence CS, Shah AR, Larossa GN, Spinner ML, Klunk WE, Mathis CA, Dekosky ST, Morris JC, Holtzman DM. Inverse relation between in vivo amyloid imaging load and cerebrospinal fluid Abeta42 in humans. Ann Neurol. 2006 Mar;59(3):512-9. PubMed.
  3. Landau SM, Fero A, Baker SL, Koeppe R, Mintun M, Chen K, Reiman EM, Jagust WJ. Measurement of longitudinal β-amyloid change with 18F-florbetapir PET and standardized uptake value ratios. J Nucl Med. 2015 Apr;56(4):567-74. Epub 2015 Mar 5 PubMed.
  4. Leuzy A, Carter SF, Chiotis K, Almkvist O, Wall A, Nordberg A. Concordance and Diagnostic Accuracy of [11C]PIB PET and Cerebrospinal Fluid Biomarkers in a Sample of Patients with Mild Cognitive Impairment and Alzheimer's Disease. J Alzheimers Dis. 2015;45(4):1077-88. PubMed.

Primary Papers

  1. Toledo JB, Bjerke M, Da X, Landau SM, Foster NL, Jagust W, Jack C Jr, Weiner M, Davatzikos C, Shaw LM, Trojanowski JQ, Alzheimer’s Disease Neuroimaging Initiative Investigators. Nonlinear Association Between Cerebrospinal Fluid and Florbetapir F-18 β-Amyloid Measures Across the Spectrum of Alzheimer Disease. JAMA Neurol. 2015 May;72(5):571-81. PubMed.

Annotate

To make an annotation you must Login or Register.