CHAPTER 77 Functional Imaging in Movement Disorders
Functional Imaging of Presynaptic Dopamine Terminals
The pathologic hallmark of PD is degeneration of pigmented dopaminergic neurons in the substantia nigra pars compacta in association with the formation of intraneuronal Lewy bodies. Loss of nigral cells results in profound dopamine depletion in the nigrostriatal projections, with terminals in the posterior part of the putamen being most affected. A loss of dopaminergic nigrostriatal neurons, however, is not exclusive to PD; it can be observed in other neurodegenerative disorders associated with parkinsonism, including multiple system atrophy (MSA), progressive supranuclear palsy (PSP), and corticobasal degeneration. Three different strategies have been developed for imaging the presynaptic function of dopamine terminals: (1) measurement of the binding of presynaptic dopamine transporter (DAT), the plasma membrane protein involved in the reuptake of dopamine from the synaptic cleft, by several PET and SPECT tracers; (2) measurement of dopamine synthesis from exogenous dopa and its vesicular storage, by 18F-dopa PET; and (3) measurement of the binding of type 2 vesicular monoamine transporter, the protein responsible for the transport of dopamine from the cytoplasm into secretory vesicles, with 11C-dihydrotetrabenazine (DTBZ) PET. Figure 77-1 is a schematic representation of a striatal presynaptic dopamine terminal.
Several PET (11C-CFT, 18F-CFT, 18F-FP-CIT, 11C-RTI-32) and SPECT (123I-β-CIT, 123I-FP-CIT, 123I-altropane, 11C-methylphenidate, 99mTc-TRODAT-1) ligands are now available for the imaging of DAT. Most of these ligands bind to dopamine, noradrenaline, and serotonin transporters. However, because the majority of transporters within the striatum are dopaminergic, reductions in the striatal uptake of these tracers provide a useful marker of DAT functional loss in that region. It must be taken into account that DAT binding may underestimate true terminal density owing to the relative downregulation of DAT in the remaining neurons as a response to reductions in synaptic dopamine. Additionally, there is evidence that DAT binding decreases with age (3.3% to 10% per decade) in healthy subjects.1–3
SPECT is more widely accessible than PET, so more DAT binding data are available with this modality in parkinsonian syndromes. In patients with PD, striatal 123I-β-CIT uptake correlates well with stage of disease and symptom severity, particularly limb bradykinesia and rigidity, but not with rest tremor severity.4–6 Patients with hemiparkinsonism show a pronounced reduction in contralateral striatal 123I-β-CIT binding. Binding in the ipsilateral striatum is also lower than that in healthy controls.7,8 Parallel findings are provided by 123I-FP-CIT SPECT (DaTSCAN), and this is frequently used in routine clinical practice to differentiate idiopathic PD from non–dopamine-deficient syndromes such as secondary nondegenerative parkinsonism and essential tremor.
Measurement of striatal aromatic amino acid decarboxylase activity with 18F-dopa PET is another way to assess the integrity of dopaminergic neurons in vivo. After intravenous administration, 18F-dopa is taken up by the terminals of dopaminergic projections, where it is converted to 18F-dopamine by aromatic amino acid decarboxylase and stored in vesicles. Measurement of 18F-dopa uptake in PD striatum therefore provides a functional measure of the number of remaining dopaminergic terminals. This is supported by pathologic studies, which have demonstrated that levels of striatal 18F-dopa uptake correlate well with nigral cell counts in both humans with various disorders and in nonhuman primates with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-induced parkinsonism.9,10 Uptake of 18F-dopa may, however, underestimate the degenerative process owing to the compensatory upregulation of aromatic amino acid decarboxylase in the preserved terminals.11
To date, 18F-dopa PET has been the most informative neuroimaging technique. In agreement with the results of neuropathologic studies, PD patients show a gradient of reduced striatal 18F-dopa uptake along a rostrocaudal axis. Patients with hemiparkinsonism have the greatest 18F-dopa uptake reduction in the dorsal posterior putamen contralateral to the symptomatic side,12 whereas patients with more advanced bilateral disease show additional decreases within the ventral and anterior putamen and dorsal caudate (Fig. 77-2). Reductions in 18F-dopa uptake within the ventral head of the caudate are usually observed only in advanced stages of disease. Putaminal uptake of 18F-dopa in PD inversely correlates with the degree of motor disability as measured by the Unified Parkinson’s Disease Rating Scale (UPDRS).13–15 In patients with MSA and PSP, 18F-dopa PET studies reveal reductions in striatal uptake that resemble those observed in PD patients; however, the caudate nucleus is generally more severely affected than in PD, and the reduction of uptake is more homogeneous within the striatal structures.16,17 Figure 77-3 shows striatal 18F-dopa uptake in a patient with MSA.
11C-DTBZ PET is a marker of type 2 vesicular monoamine transporters. It has been suggested that 11C-DTBZ binding provides the most reliable measurement of the density of dopaminergic terminals.18 At present, however, 11C-DTBZ PET is available in only a few centers, and not many studies have been performed with this technique.
Functional Imaging of Dopamine Receptor Availability
In PD patients, striatal 123I-IBZM binding is normal in early disease, and it remains normal or becomes mildly reduced in more advanced disease.19–21 With 11C-raclopride PET, a mild increase in putaminal D2 binding has been observed in untreated PD patients, suggesting either increased D2 receptor availability due to reduced occupancy by endogenous dopamine or receptor upregulation secondary to deafferentation.22–24 As the disease progresses and dopaminergic therapy is started, 11C-raclopride binding normalizes in the putamen and mildly decreases in the caudate,13,24–27 reflecting either disease progression or an effect of treatment.
11C-FLB 457 PET also reveals decreased dopamine receptor availability in the thalamus, anterior cingulate, and dorsolateral prefrontal and temporal cortices in patients with advanced PD.28,29 Scherfler and colleagues30 recently reported that untreated patients with parkinsonism linked to parkin gene mutations show an increase in putaminal 11C-raclopride binding compared with normal subjects. However, at variance with observations in PD patients, levodopa-treated parkin-linked patients show significant reductions in 11C-raclopride binding in both the caudate and the putamen. This could reflect a greater susceptibility to the exposure to dopaminergic medication in parkin-linked patients than in idiopathic PD patients.
In patients with MSA and PSP, both 11C-raclopride binding and 123I-IBZM binding are reduced, suggesting that a degeneration of striatal D2 receptors occurs in these conditions.16,20–22,26,31 Unfortunately, a degree of overlap is seen across the ranges of individual D2 binding data in MSA, PSP, and PD patients and normal subjects. Given this overlap, striatal dopamine binding does not appear to be a sensitive means of discriminating PD from other neurodegenerative causes of parkinsonism.
Finally, reduction of dopamine receptor binding has been proposed as a biomarker of striatal functional integrity in HD. The neurodegenerative process in HD leads to progressive loss of striatal medium-spiny GABA-ergic neurons bearing both dopamine D1 and D2 receptors, along with opioid and benzodiazepine receptors. Decreases of striatal D2 binding in HD patients and asymptomatic HD carriers have been detected with both 11C-raclopride PET and 123I-IBZM SPECT, whereas 123I-epidepride binding was reduced only in patients with advanced HD. Reductions in striatal D1 binding have also been reported with 11C-SCH23390 PET.32–39 Voxel-based statistical parametric mapping of 11C-raclopride PET images has localized a significant reduction of D2 dopamine receptor availability in frontal and temporal areas,38 suggesting that this may play a role in the pathophysiology of cognitive disturbances in early HD.
Functional Imaging of Dopamine Release
Over the past decade, numerous studies have demonstrated the sensitivity of 11C-raclopride PET to fluctuations in synaptic concentrations of dopamine following pharmacologic or behavioral challenges. Rises in synaptic dopamine levels translate into decreases in dopamine D2 receptor availability, which can be detected as reductions in 11C-raclopride binding.40 It has been estimated that a 10% reduction in the availability of D2 receptors for 11C-raclopride binding reflects a fivefold increase in synaptic dopamine levels.41 Using this paradigm, endogenous release of dopamine has been measured after the administration of methamphetamine or levodopa,42–44 during repetitive transcranial magnetic stimulation,45 and during the performance of behaviorally rewarded tasks46 or sequential finger movements47,48 in both normal subjects and PD patients.
In a group of PD patients, the mean reductions in caudate and putamen 11C-raclopride binding following intravenous methamphetamine administration were significantly attenuated compared with normal subjects (8% versus 17% in the caudate; 7% versus 25% in the putamen). The percentage reduction in putamen 11C-raclopride binding after amphetamine administration correlated with both putaminal 18F-dopa uptake and UPDRS scores of individual patients not taking their medication. In the same cohort of patients, statistical parametric mapping localized similar levels of dopamine release in dorsal and ventrolateral prefrontal and orbitofrontal areas in PD patients and normal subjects. These findings indicate that dopamine release in frontal areas after a pharmacologic challenge is preserved even in late stages of PD.42
11C-raclopride PET has also been employed to assess the changes in synaptic levels of dopamine that result after exogenous levodopa administration in PD patients (Fig. 77-4).43,44,49,50 All these studies have shown that striatal reductions in 11C-raclopride binding after levodopa become larger as motor disability increases and the disease progresses. The increased synaptic dopamine levels that result from levodopa in more advanced PD probably reflect the reduced dopamine storage capacity of the putamen and fewer dopamine transporters available to clear the transmitter. The relationship between clinical improvement and synaptic dopamine release after a single oral dose of levodopa has been evaluated in a group of patients with advanced PD.43 Although individual improvements in rigidity and bradykinesia correlated well with putaminal dopamine release, this was not the case for tremor or axial symptoms. Interestingly, large putaminal 11C-raclopride binding changes were associated with higher dyskinesia scores. These findings indicate that in advanced PD, the improvement of rigidity and bradykinesia and the presence of dyskinesia after a single dose of oral levodopa are governed by the level of dopamine generated at striatal D2 receptors. They also confirm that relief of parkinsonian tremor and axial symptoms is not related to striatal synaptic dopamine levels and presumably occurs via extrastriatal mechanisms.
Functional Imaging of Cerebral Blood Flow and Glucose Metabolism
Resting Glucose Metabolism
Absolute levels of rCMRGlc are usually normal in PD, but patients show a characteristic profile of relative lentiform nucleus and thalamic hypermetabolism, along with hypometabolism of the lateral and mesial premotor cortex, supplementary motor area (SMA), dorsolateral prefrontal cortex (DLPFC), temporoparietal cortex, and parieto-occipital association regions.51–55 Expression of this PD-related profile can be quantitated and correlates with disease severity as assessed by the UPDRS. The pallidal and putaminal hypermetabolism observed in patients with early PD may represent a compensatory response to deafferentation. Increased 18F-dopa uptake in the globus pallidus interna has also been reported at the onset of PD symptoms.56 Although striatal metabolism is preserved in PD, reduced rCMRGlc is observed in the majority of patients with atypical neurodegenerative parkinsonism such as MSA and PSP.16,57 The patterns of altered regional glucose metabolism in each of these parkinsonian conditions have been used to differentiate them from PD. It has been reported that 18FDG PET has 95% sensitivity and 94% specificity for discriminating patients with PD from healthy volunteers when computer-assisted methodologies are applied.58 However, levels of rCMRGlc are sensitive to coexisting cognitive impairment and depression, and they normalize following treatment with dopaminergic drugs. Using 18FDG PET, it has been possible to detect subclinical cortical dysfunction in about one third of nondemented patients with established PD.53,59,60
The functional effects of HD pathology have also been extensively investigated in vivo with both 18FDG PET and cerebral perfusion SPECT studies. Most studies have reported reduced glucose metabolism and cerebral perfusion in the basal ganglia, particularly the heads of the caudate, and also in association with cortical areas targeting the frontal cortex. Reductions in both cortical and subcortical blood flow and glucose metabolism reportedly correlate with cognitive function in HD. Reduced basal ganglia rCMRGlc is detectable in asymptomatic HD gene mutation carriers. It has been suggested that in many cases this precedes the morphologic changes, such as caudate nucleus atrophy, that are observed on computed tomography and magnetic resonance imaging. In practice, SPECT imaging of caudate blood flow is a less sensitive indicator of caudate dysfunction than is 18FDG PET.61,62
Activation Studies
Several H215O PET studies have compared patterns of regional brain activation during the execution of finger or hand movements by normal subjects, PD patients, and HD patients. Compared with normal subjects, PD patients have reduced regional cerebral blood flow (rCBF) in the rostral SMA and right DLPFC when performing paced movements of a joystick in freely selected directions with the right hand.63 Similar findings were observed in PD patients learning a sequence of finger movements.64 An inability to activate the SMA and DLPFC during learning or selection of movements could explain the difficulty PD patients have acquiring new motor skills and initiating volitional movements. Interestingly, the administration of apomorphine normalized this aberrant activation pattern as soon as the patients switched to an “on” state.65
Other H215O PET and, more recently, functional magnetic resonance imaging studies have shown raised activation in the cerebellum and lateral premotor and parietal areas in PD patients compared with healthy subjects during sequential manual movements.66–69 The hyperactivity in these areas may represent the use of alternative cortical-subcortical circuits to avoid basal ganglia connections, but this is still being debated. It may also explain the beneficial effects of visual and auditory cues when PD patients perform volitional movements. In a recent longitudinal study, 13 patients with recent-onset PD were studied twice with H215O PET. Imaging was performed with the patients off medication at baseline and again after 2 years.70 On both occasions, subjects were asked to perform paced reaching movements toward targets presented in a predictable order. Increasing activation in the pallidum bilaterally and in the left putamen was detected as the disease progressed. In the cortex, motor-related activation increased in the right pre-SMA, anterior cingulate cortex, and left postcentral gyrus. Increases in the right dorsal premotor cortex correlated well with progressive delays in movement initiation, whereas slowing of movement velocity was associated with declining activation in the left DLPFC and dorsal premotor cortex. These findings suggest that with advancing PD, motor performance is associated with the recruitment of brain regions normally involved in the execution of more complex tasks. This may reflect a progressive loss of functional selectivity or inefficient compensatory activation.
There have been only a few activation studies with H215O PET in patients with HD. In one study, changes in rCBF were assessed during the execution of paced joystick movements in freely chosen directions.71 HD patients showed impaired activation of premotor and prefrontal areas in a pattern similar to that observed for PD patients. This finding helps explain the bradykinesia observed in HD patients. Another H215O PET study showed that HD patients have reduced activation of frontostriatal circuitry, along with enhanced activation of parietal areas during a finger opposition task, suggesting compensatory recruitment of normally redundant circuitry.72
Clinical Applications of Functional Neuroimaging in Parkinson’s Disease
Differential Diagnosis
18F-dopa PET and markers of DAT binding such as 123I-β-CIT and 123I-FP-CIT SPECT can differentiate patients with dopamine-deficient parkinsonian syndromes, such as PD, from normal subjects and those with essential tremor with a sensitivity of around 90%.73 It has been reported that 123I-β-CIT SPECT is 100% sensitive and specific for the diagnosis of PD in patients younger than 55 years.74 These approaches can also reveal the presence of dopamine terminal dysfunction in drug-associated parkinsonism74–76 and apparently psychogenic PD.77
The ability of 18F-dopa PET and markers of DAT binding to differentiate among the different dopamine-deficient parkinsonian syndromes is poor owing to the overlap in individual ranges of putamen tracer uptake in PD, MSA, and PSP patients.78,79 In patients with vascular parkinsonism, 123I-β-CIT and 123I-FP-CIT uptake is heterogeneous; some patients show normal uptake, and others show reduced uptake.74,76 In one study, 99mTc-TRODAT-1 was reported to reliably discriminate between PD and vascular parkinsonism.80 Table 77-1 gives an overview of neuroimaging findings in parkinsonian syndromes.
Detection of Preclinical Disease
18F-dopa PET has demonstrated subclinical dopaminergic dysfunction in one quarter of asymptomatic adult relatives in kindreds with known familial PD.81 One third of those relatives with abnormal imaging went on to develop clinical parkinsonism during a 5-year longitudinal follow-up. 123I-β-CIT SPECT revealed loss of striatal DAT binding in 10% of asymptomatic relatives of PD patients shown to have hyposmia on the University of Pennsylvania Smell Identification Test (UPSIT), another known risk factor for PD.82 Another 18F-dopa PET study reported reduced putaminal 18F-dopa uptake in asymptomatic co-twins of PD patients with apparently sporadic disease. Dopaminergic dysfunction was more common in monozygotic co-twins than in dizygotic co-twins (56% and 18%, respectively). Two of 18 monozygotic co-twins developed clinical parkinsonism over a 4-year follow-up.83 These findings strongly support a role of inheritance in apparently sporadic PD.
Efficacy of Putative Disease-Modifying Agents
Longitudinal changes in striatal 18F-dopa PET or 123I-β-CIT SPECT have been used as in vivo biomarkers of PD progression in several clinical trials performed to evaluate the efficacy of dopamine agonists, levodopa, and putative neuroprotective agents in modifying the natural course of this condition. Neither the glutamate release inhibitor riluzole nor the mixed lineage kinase inhibitor CEP1347 had a significant effect on disease progression as rated both clinically and with dopaminergic imaging. Use of dopamine agonists slowed the loss of dopamine terminal function in PD relative to levodopa in early PD, but clinically the patients were more improved with the latter medication. In general, functional imaging has either failed to show a neuroprotective effect of the agents on trial or produced findings discordant with clinical outcomes. Table 77-2 gives a brief overview of these studies.
