Rxivist logo

Test-retest reliability and convergent validity of (R)-[11C]PK11195 outcome measures without arterial input function

By Pontus Plavén-Sigray, Granville J. Matheson, Zsolt Cselényi, Aurelija Jučaite, Lars Farde, Simon Cervenka

Posted 11 Apr 2018
bioRxiv DOI: 10.1101/298992 (published DOI: 10.1186/s13550-018-0455-8)

Background: The positron emission tomography radioligand (R)-[11C]PK11195 can be used to quantify the expression of translocator protein (TSPO), which is considered a marker for activation of glial cells. TSPO is expressed throughout the brain, and for this reason no true reference region exists. When a radioligand does not have a reference region, an arterial input function (AIF) is usually required in order to quantify binding. However, obtaining an AIF can be difficult as well as uncomfortable for participants. Alternative methods have therefore been proposed with the aim of estimating (R)-[11C]PK11195 binding without arterial measurements, such as standardized uptake values (SUVs), supervised-cluster analysis (SVCA), or the use of a pseudo-reference region. The objective of this study was to evaluate the test-retest reliability and convergent validity of these techniques. Methods: Data from a previously published (R)-[11C]PK11195 test-retest study in six healthy male subjects were reanalysed. Non-displaceable binding potential (BPND) was calculated for a set of cortical and subcortical brain regions using the simplified reference tissue model, with either cerebellum as reference region or a reference input derived using SVCA. SUVs were estimated for the time interval of 40-60 minutes. For comparison, total distribution volume (VT), specific distribution volume (VS) and BPND were estimated from the two-tissue-compartment model (2TCM) using AIF. Test-retest reliability was then assessed for all outcome measures. Convergent validity was examined by correlating all measures derived without an AIF to those derived using 2TCM. Results: Test-retest reliability for BPND estimates were poor (80% of all regional ICCs<0.5). SUVs showed, on average, moderate reliability. BPND estimates derived without an AIF were not correlated with VT, VS or BPND from the 2TCM (all R2<12%). SUVs were not correlated with any other outcome (all R2<9%). Discussion: BPND estimated using cerebellum or SVCA as reference input showed poor reliability and little to no convergent validity with outcomes derived using an AIF. SUVs showed moderate reliability but no convergent validity with any other outcome. Caution is warranted for interpreting patient-control comparisons employing (R)-[11C]PK11195 outcome measures obtained without an AIF.

Download data

  • Downloaded 598 times
  • Download rankings, all-time:
    • Site-wide: 42,414
    • In neuroscience: 6,055
  • Year to date:
    • Site-wide: 52,810
  • Since beginning of last month:
    • Site-wide: 56,622

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide


Sign up for the Rxivist weekly newsletter! (Click here for more details.)