Analysis of [11C]PE2I
This a review of publications on the analysis of [11C]PE2I PET studies and SPECT studies with PE2I labelled with other tracers.
Unlike many other DAT ligands, [11C]PE2I is a highly selective ligand for DAT, and does not accumulate in regions rich in the serotonin or noradrenaline transporter. The extrastriatal PE2I binding is very low (Hall et al. 1999). In vitro binding of [125I]PE2I in substantia nigra was about 50% of binding in caudate and putamen, and 15% in thalamus (Hall et al. 1999). In displacement and pre-treatment measurements the cerebellum-to-blood ratio did not change (Poyot et al. 2001; Halldin et al. 2003), supporting the use of cerebellum as a reference region.
[11C]PE2I shows relatively high non-specific binding to white matter of both cerebrum and cerebellum: it is evident in most published in vitro binding assays and PET images (Hall et al. 1999; Poyot et al. 2001); in vitro accumulation of [125I]PE2I corresponds to 30-35% of the specific binding in caudate and putamen (Hall et al. 1999). This is probably due to its relatively high lipophilicity (Hall et al. 1999).
Compartment model analysis
Pinborg et al. (2002, 2005) have fitted traditional one- and two-tissue compartment models to human [123I]PE2I-SPECT data. They found that for all ROIs, AIC indicated a significantly better fit using the two-tissue compartment model compared with the one-tissue compartment model, and that this was most evident in the receptor poor cortical regions. They assumed that the second tissue compartment may be explained by parallel compartment (heterogeneity) instead of serial compartment (non-specific binding), which might in turn explain cause the non-physiological rate constants and compartment distribution volumes (Pinborg et al. 2002). At early publications, because of the poor resolution of SPECT and PET scanners, it was expected that the relatively high accumulation in white matter could be forming most of the “parallel compartment”. Another explanation - presence of a radioactive metabolite passing the blood-brain barrier - was at that time suggested to be unlikely on basis that the two main metabolites of [11C]PE2I are polar compounds (Halldin et al. 2003).
But now with more recent studies it has been confirmed, both with rat and human studies, that metabolites do exist (Jucaite et al. 2006, Shetty et al. 2007 and Hirvonen et al. 2008). At least two or three metabolites have been identified for [11C]PE2I in human studies (Jucaite 2006, Hirvonen 2008).
Pinborg et al. (2002) estimated that the first-pass extraction fraction of [123I]PE2I was about 0.72 in striatum and about 0.34-0.42 in cortical regions. With [11C]PE2I the first-pass extraction fraction might be even higher. This suggests that when vascular radioactivity concentration is concerned, arterial blood concentration represents only the arterial fraction of it, and venous volume fraction can possibly be neglected.
According to Pinborg et al. (2005), the full reference tissue model and the simplified reference tissue model performed equally well with [123I]PE2I SPECT data based on Akaike information criterion values, and the BP estimates were similar than with Logan analysis when study length was 90 min or longer (Pinborg et al., 2005). But recently DeLorenzo et al. (2009) found that SRTM does actually not perform as well as Logan.
A non-iterative two tissue compartment model (2TCNI) has been suggested by DeLorenzo et al. (2009). With this non-iterative method, the data is compared to a set of precalculated functions, rather than performing non-linear least squares fit. (Ogden et al. 2007)
DeLorenzo et al. have compared results from one- and two-tissue compartment models both with iterative and non-iterative solutions, basis pursuit method, Logan, LEGA and bloodless Logan, LEGA and SRTM. 2TCNI model with 100 min scans is suggested. 2TCNI is found to perform best in terms of test-retest percentage difference, with-in subject mean sum of squares, variance, identifiability (data stability) and time stability (method reaches the steady state earlier than others).
Still, it is not evident if the metabolites have been taken into account in interpretation of method comparison. It might be that metabolites are affecting the results in an unknown way. For example, the determination of the length of the scan may be affected by the fact, that it is actually the metabolites that finally reach a stable state instead of specifically bound ligand. So we cannot be sure if the specific binding has reached stable state actually earlier.
In Turku, the iterative version of 2TC model could also be considered. DeLorenzo et al. suspected that the method was trapped into local minima, but for us it should not be a problem because of a developed global optimization solution.
Poyot et al (2001) validated the use of multi-injection protocol with arterial input in primates for estimation of B’max. In putamen and caudate, B’max correlated strictly with [125I]PE2I determined in vitro, although the values were different.
Non-compartment analysis (Logan plot)
SPECT studies in healthy volunteers, Parkinson’s disease (PD) patients, and nonhuman primates with [123I]PE2I have validated the use of this tracer and Logan graphical method with occipital cortex input for studying the binding of PE2I and the clinical features of PD (Pinborg et al. 2002 and 2005; Prunier et al. 2003a; Prunier et al. 2003b). In SPECT studies, occipital cortex is used as reference region instead of cerebellum because occipital cortex is readily visible and well defined in SPECT images.
Pinborg et al. (2002) applied Logan plot to the [123I]PE2I SPECT studies with metabolite corrected plasma input, and reference tissue input with and without reference tissue k2 correction (k’2=0.013±0.006 min-1 from six subjects, determined from the intercept of the plasma input Logan plot of the occipital cortex). Their Logan plot linearity is not reached before 120 min in all cases, neither with plasma or occipital cortex input. Surprisingly, applying k’2 correction did not decrease the time to reach linearity, but even increased it in some cases. Pinborg et al. (2002) recommend using Logan plot with occipital cortex input without k’2 correction.
Hirvonen et al. have used Logan analysis with [11C]PE2I (Hirvonen 2008) with arterial plasma input.
DeLorenzo et al. (2009) have studied Logan both with blood input, and reference region (cerebellum) input in a [11C]PE2I PET study. They suggest that 2TCNI method as the best option, but both Logan versions are also reported to perform reasonably well. The results correlate well with 2TCNI, even though it was studied that the assumption of only one tissue compartment in reference region is actually not holding true for [11C]PE2I. DeLorenzo also notes that it has been shown that Logan plot can underestimate non-displaceable binding potential.
Likelihood estimation graphical analysis (LEGA) was studied by DeLorenzo et. al. (2009) both with blood and reference region (cerebellum) input. It was found to give similar results as Logan, with no significant differences. Although, Logan performed slightly better in terms of test-retest percentage difference, with-in subject mean sum of squares and variance.
Seki et al. (2010) have suggested the original multilinear reference tissue model (MRTMo). In their publication it was found to perform better than SRTM method. MRTMo is one of the variations of graphical approach, where a multilinear regression is obtained after a certain equilibrium time (Seki 2010, Ichise 1996).
Peak equilibrium (ratio) analysis
Peak equilibrium analysis, (striatum-occipital cortex)/occipital cortex ratio at the peak of (striatum-occipital cortex) curve, provided similar BP estimates and SEM than simplified reference tissue model for [123I]PE2I (Pinborg et al., 2005). However, the ratio was highly variable from 0 to 70-80 min after injection, making identification of the correct scan period important and thus in practice requiring dynamic acquisition. Therefore, Pinborg et al. (2005) suggest using bolus/infusion protocol even in clinical studies, because then a certain scan time (after 120 min of constant infusion) can be applied for all subjects.
Simplified reference tissue model (SRTM)
At least Jucaite et al. (2005), Odano et al. (2006), Leroy et al. (2007), Ciumas et. al (2008) and Arakawa et. al. (2009) have used simplified reference tissue model with cerebellum as reference region to estimate regional BP values. Jucaite et al. (2006) compared SRTM and plasma input models, and suggest that SRTM can be used in clinical studies when arterial sampling need to be avoided.
In a more recent study by DeLorenzo et. Al. (2009), the SRTM has been noted to perform worse than bloodless Logan analysis or LEGA. On the other hand Seki et al. (2009) have suggested that original multi-linear reference tissue model (MRTMo) should be used instead of SRTM.
Jonasson et al (2013) compared different methods for voxel level analysis of [11C]PE2I studies, and concluded that basis function implementation of SRTM is the best.
About refence regions
DeLorentzo et al. have studied thoroughly the kinetics of [11C]PE2I and at first they studied a large set of brain regions to find a reference region, where best fitting model would be 1TC. This is an assumption made by most reference region methods.
The cerebellum, cuneus, gyrus, dorsal and lateral prefrontal cortex, fusiform gyrus, gyrus rectus and uncus were run through cluster analysis and all clusters were fit with both 1TC and 2TC model. It turned out that all of these regions and their subregions fit better with 2TC model (lower AIC measure than 1TC model).
DeLorenzo et al. still do not exclude the usage of reference region models. But when analysing [11C]PE2I data with these methods, it should be noted that the reference region models are not fitting the data perfectly.
- Arakawa et al. Increase in thalamic binding of [11C]PE2I in patients with schizophrenia: A positron emission tomography study of dopamine transporter. J phychiatric research 2009; 43:1219-23
- DeLorenzo et. al Modeling considerations for in vivo quantification of the dopamine transporter using [11C]PE2I and positron emission tomography. J cereb Blood Flow Metab 2009; 29:1332-1345.
- Ciumas et al. Reduced dopamine transporter binding in patients with juvenile myoclonic epilepsy. Neurology (e-Pub ahead of print on May 7, 2008, at www.neurology.org.) 2008;788-794
- Hall H et al. Visualization of the dopamine transporter in the human brain postmortem with the new selective ligand [125I]PE2I. Neuroimage 1999; 9: 108-116.
- Halldin C et al. [11C]PE2I: a highly selective radioligand for PET examination of the dopamine transporter in monkey and human brain. Eur J Nucl Med 2003; 30: 1220-1230.
- Hirvonen et al. Measurement of striatal and extrastriatal dopamine transporter binding with high-resolution PET and [11C]PE2I: quantitative modelling and test-retest reproducibility. 2008; J Cereb Blood Flow Metab 28:1059-69
- Ichise M. et al. Noninvasive quantification of dopamine D2 receptors with iodine-123-IBF SPECT. J Nucl Med. 1996;37(3):513–20.
- Jonasson M, Appel L, Frick A, Nyholm D, Askmark H, Danfors T, Sörensen J, Furmark T, Lubberink M. Validation of parametric methods for [11C]PE2I positron emission tomograhy. NeuroImage 2013 (in press).
- Jucaite A, Fernell E, Halldin C, Forssberg H, Farde L. Reduced midbrain dopamine transporter binding in male adolescents with attention-deficit/hyperactivity disorder: association between striatal dopamine markers and motor hyperactivity. Biol Psychiatry 2005; 57: 229-238.
- Jucaite A, Odano I, Olsson H, Pauli S, Halldin C, Farde L. Qunatitative analyses of regional [11C]PE2I binding to the dopamine transporter in the human brain: a PET study. Eur J Nucl Med Mol Imaging 2006; 33: 657-668.
- Leroy C, Comtat C, Trébossen R, Syrota A, Martinot J-L, Ribeiro M-J. Assessment of 11C-PE2I binding to the neuronal dopamine transporter in humans with the high-spatial-resolution PET scanner HRRT. J Nucl Med 2007; 48: 538-546.
- Odano et al. Dopamine transporter binding corrected for tissue heterogeneity and partial volume effect using [11C]PE2I and PET. Poster Presentations / NeuroImage 31 2006; T44–T186.
- Ogden RT, Ohja A, Erlandsson K, Oquendo M, Mann JJ, Parsey RV. In vivo quantification of serotonin transporters using [11C]DASB and positron emission tomography in humans: modelling considerations. J Cereb Blood Flow Metab 2007; 27:205-17.
- Pinborg LH et al. Quantification of [123I]PE2I binding to dopamine transporters with SPET. Eur J Nucl Med 2002; 29: 623-631.
- Pinborg LH et al. Quantification of 123I-PE2I binding to dopamine transporter with SPECT after bolus and bolus/infusion. J Nucl Med 2005; 46: 1119-1127.
- Poyot T et al. Anatomic and biochemical correlates of the dopamine transporter ligand 11C-PE2I in normal and Parkinsonian primates: comparison with 6-[18F]fluoro-L-dopa. J Cereb Blood Flow Metab 2001; 21: 782-792.