11C]Harmine Binding to Brain Monoamine Oxidase A: Test-Retest Properties and Noninvasive Quantification.

Abstract

Inhibition of the isoform A of monoamine oxidase (MAO-A), a mitochondrial enzyme catalyzing deamination of monoamine neurotransmitters, is useful in treatment of depression and anxiety disorders. [11C]harmine, a MAO-A PET radioligand, has been used to study mood disorders and antidepressant treatment. However, [11C]harmine binding test-retest characteristics have to date only been partially investigated. Furthermore, since MAO-A is ubiquitously expressed, no reference region is available, thus requiring arterial blood sampling during PET scanning. Here, we investigate [11C]harmine binding measurements test-retest properties; assess effects of using a minimally invasive input function estimation on binding quantification and repeatability; and explore binding potentials estimation using a reference region-free approach. Quantification of [11C]harmine distribution volume (VT) via kinetic models and graphical analyses was compared based on absolute test-retest percent difference (TRPD), intraclass correlation coefficient (ICC), and identifiability. The optimal procedure was also used with a simultaneously estimated input function in place of the measured curve. Lastly, an approach for binding potentials quantification in absence of a reference region was evaluated. [11C]harmine VT estimates quantified using arterial blood and kinetic modeling showed average absolute TRPD values of 7.7 to 15.6%, and ICC values between 0.56 and 0.86, across brain regions. Using simultaneous estimation (SIME) of input function resulted in VT estimates close to those obtained using arterial input function (r = 0.951, slope = 1.073, intercept = - 1.037), with numerically but not statistically higher test-retest difference (range 16.6 to 22.0%), but with overall poor ICC values, between 0.30 and 0.57. Prospective studies using [11C]harmine are possible given its test-retest repeatability when binding is quantified using arterial blood. Results with SIME of input function show potential for simplifying data acquisition by replacing arterial catheterization with one arterial blood sample at 20min post-injection. Estimation of [11C]harmine binding potentials remains a challenge that warrants further investigation.