Department of Psychiatry and Psychotherapy, Medical University of Vienna, Austria.
Nucl Med Biol. 2013 Apr;40(3):371-7. doi: 10.1016/j.nucmedbio.2012.12.011. Epub 2013 Jan 30.
Image-derived input functions (IDIFs) represent a promising non-invasive alternative to arterial blood sampling for quantification in positron emission tomography (PET) studies. However, routine applications in patients and longitudinal designs are largely missing despite widespread attempts in healthy subjects. The aim of this study was to apply a previously validated approach to a clinical sample of patients with major depressive disorder (MDD) before and after electroconvulsive therapy (ECT).
Eleven scans from 5 patients with venous blood sampling were obtained with the radioligand [carbonyl-(11)C]WAY-100635 at baseline, before and after 11.0±1.2 ECT sessions. IDIFs were defined by two different image reconstruction algorithms 1) OSEM with subsequent partial volume correction (OSEM+PVC) and 2) reconstruction based modelling of the point spread function (TrueX). Serotonin-1A receptor (5-HT1A) binding potentials (BPP, BPND) were quantified with a two-tissue compartment (2TCM) and reference region model (MRTM2).
Compared to MRTM2, good agreement in 5-HT1A BPND was found when using input functions from OSEM+PVC (R(2)=0.82) but not TrueX (R(2)=0.57, p<0.001), which is further reflected by lower IDIF peaks for TrueX (p<0.001). Following ECT, decreased 5-HT1A BPND and BPP were found with the 2TCM using OSEM+PVC (23%-35%), except for one patient showing only subtle changes. In contrast, MRTM2 and IDIFs from TrueX gave unstable results for this patient, most probably due to a 2.4-fold underestimation of non-specific binding.
Using image-derived and venous input functions defined by OSEM with subsequent PVC we confirm previously reported decreases in 5-HT1A binding in MDD patients after ECT. In contrast to reference region modeling, quantification with image-derived input functions showed consistent results in a clinical setting due to accurate modeling of non-specific binding with OSEM+PVC.
在正电子发射断层扫描（PET）研究中，图像衍生输入函数（IDIFs）代表了一种有前途的非侵入性替代动脉采血的方法，用于定量分析。然而，尽管在健康受试者中进行了广泛的尝试，但在患者中常规应用和纵向设计仍然很大程度上缺失。本研究的目的是将以前经过验证的方法应用于接受电惊厥治疗（ECT）前后的重度抑郁症（MDD）患者的临床样本中。
从 5 名接受静脉采血的患者中获得了 11 次扫描，使用放射性配体[羰基-(11)C]WAY-100635 在基线、ECT 前后 11.0±1.2 次时进行。通过两种不同的图像重建算法 1）OSEM 后进行部分容积校正（OSEM+PVC）和 2）基于点扩散函数的重建（TrueX）来定义 IDIFs。通过双组织室（2TCM）和参考区模型（MRTM2）来量化 5-羟色胺 1A 受体（5-HT1A）结合势（BPP、BPND）。
与 MRTM2 相比，当使用 OSEM+PVC 的输入函数时，5-HT1A BPND 具有良好的一致性（R（2）=0.82），但使用 TrueX 时则不一致（R（2）=0.57，p<0.001），这进一步反映了 TrueX 的 IDIF 峰值较低（p<0.001）。在 ECT 后，使用 OSEM+PVC 的 2TCM 发现 5-HT1A BPND 和 BPP 降低（23%-35%），除了一名患者仅显示出轻微的变化。相比之下，MRTM2 和 TrueX 的 IDIF 对于该患者的结果不稳定，最有可能是由于非特异性结合的 2.4 倍低估。
使用 OSEM 后进行的图像衍生和静脉输入函数定义，我们证实了先前报道的 ECT 后 MDD 患者 5-HT1A 结合减少。与参考区建模相比，由于 OSEM+PVC 对非特异性结合的准确建模，使用图像衍生输入函数进行定量分析在临床环境中显示出一致的结果。
