Introduction

Parkinson’s disease (PD) is characterized by the motor symptoms bradykinesia, rigidity, resting tremor and postural instability in later stages and the diagnosis is based on these symptoms. Typically the symptoms start asymmetrically on one body side, gradually affect both sides and respond well to levodopa. In many cases the clinical picture is very characteristic and the diagnosis clear. Sometimes, it is even possible to diagnose this disorder by just observing a patient as it is described in James Parkinson’s original monography (1817). However, in early stages the symptoms can be uncharacteristic or incomplete and in some patients there may be overlap with other syndromes.

Marttila and Rinne (1976) investigated the accuracy of clinical diagnosis in 800 patients referred to their department with the diagnosis of parkinsonism and 25% were re-diagnosed as essential tremor (ET), 4% as vascular pseudo-parkinsonism and 8% as drug induced parkinsonism. In a community-based study (Meara et al. 1999) 402 patients with presumed PD diagnosed by general practitioners only 53% fulfilled the clinical criteria of PD. The commonest causes of misdiagnoses were ET, Alzheimer’s disease and vascular pseudo-parkinsonism. On the other hand a population-based study in London found that in patients who were examined because of clinical symptoms of Parkinsonism or tremor the diagnosis of PD was overlooked in 20% of cases (Schrag et al. 2002). Even in the diagnosis by movement disorder specialists, error rates can be as high as 25%, as was shown in a clinico-pathological study by Hughes et al. (1992). Using stringent diagnostic criteria the diagnostic accuracy was increased to 91%. The major source of error was in the differential diagnosis of PD and atypical degenerative parkinsonian syndromes such as multi system atrophy (MSA) and progressive supranuclear palsy (PSP). Vascular pseudo-parkinsonism and Alzheimer’s disease were also sometimes misdiagnosed (Hughes et al. 2001, 2002).

Therefore, there is a need for an objective diagnostic test or biological marker for PD.

Because the evaluation of disease progression based on clinical neurological examination is difficult due to the effects of antiparkinsonian treatment there is also a need for an objective measure of the dopaminergic nerve cell loss in vivo to determine the rate of progression and to test possible neuroprotective strategies. Imaging methods with either positron emission tomography (PET) or single photon emission tomography (SPECT) are one way to show a dysfunction or degeneration of nigrostriatal dopaminergic terminals. Several of these methods have been developed during the last decades and some of them have become clinical routine.

18F-dopa PET

18F-dopa PET is a marker of dopaminergic terminal function and reflects dopa transport into the terminal, dopa decarboxylase activity and dopamine storage capacity (Poewe and Scherfler 2003). 18F-dopa PET studies were the first, and for a long time, the only approach for the in vivo assessment of dopaminergic function in PD.

The first report of the successful visualization of dopaminergic terminals in the striatum with 18F-dopa PET in vivo in humans was published by Garnett et al. (1983). A severe reduction of 18F-dopa uptake in the striatum of PD patients was shown by Leenders et al. (1986).

Since then a large number of studies using 18F-dopa PET have been published. Although it has been proven to be a reliable method to study the dopaminergic system, the high hardware costs and the necessity of a cyclotron facility for the production of 18F in close vicinity have restricted its use in clinical practice.

Development of dopamine transporter (DAT) imaging

DATs are localized on dopaminergic axons in the striatum and regulate extracellular dopamine levels. DAT internalization and recruitment back to the plasma membrane is regulated by phosphorylation and de-phosphorylation which in turn regulate DAT activity and dopamine reuptake (Piccini 2003).

In the 1980s in vitro studies in post-mortem human brain samples of PD patients showed a loss of DAT binding which was due to the degeneration of dopaminergic nigro-striatal neurons (Pimoule et al. 1983; Janowsky et al. 1987). These findings could be replicated in vivo with the PET ligand 11C-nomifensine (Aquilonius et al. 1987). 11C-nomifensine has low selectivity and high nonspecific binding and was, therefore, not suitable for clinical application (Piccini 2003). 11C-cocaine was not suitable either because of its high nonspecific binding and rapid uptake and clearance from the brain (Innis 1994).

1989 a group of cocaine analogues (phenyltropanes) with very high affinity for DATs and low nonspecific binding was synthesized (Madras et al. 1989). Postmortem studies with one of these analogues (3H-CFT) in PD brains showed severe depletion of DAT binding (Kaufman and Madras 1991). PET studies with the 11C labelled phenyltropane CFT also demonstrated reduced binding in PD patients (Frost et al. 1993). ß-CIT (RTI 55) is an iodinated analogue of CFT with further increased affinity for monoamine transporters (Boja et al. 1991; Neumeyer et al. 1991). ß-CIT has a 150-fold higher affinity for DATs than cocaine and slower brain kinetics (Innis 1994). Iodinated compounds can be labelled with 123I and thus can be used for SPECT. Image resolution in SPECT is in the range of 8–10 mm, whereas image resolution in PET can be as good as 4–5 mm. However, due to the longer half-life of 123I in comparison to 11C and 18F which are used for labelling tracers for PET as well as due to the slow kinetics of phenyltropanes SPECT offers advantages over PET. 123I-ß-CIT was extensively studied in animal experiments and its binding studied in post-mortem human brain samples (see Brücke et al. 1997).

The first studies with 123I-ß-CIT SPECT in humans demonstrated marked reduction of striatal binding in PD patients as compared with normal controls (Brücke et al. 1993a; Figs. 1, 2; Innis et al. 1993).

Fig. 1
figure 1

(= Fig. 1 in Brücke et al. 1993a). 123I-ß-CIT binding in a 27-year-old healthy control. 3,125 mm thick adjacent slices oriented parallel to the canto-meatal plane are shown. Subjects right side on the left, left upper corner bottom, right lower corner top. The ratio of specific striatal binding to nonspecific cerebellar binding is 8.8

Full size image
Fig. 2
figure 2

(= Fig. 4 in Brücke et al. 1993a). Activity ratios between target areas (corresponding to the striatum and hypothalamus–midbrain area) and the cerebellum in four controls (C) and two patients with Parkinson’s disease (P). To our knowledge this publication (Brücke et al. 1993a) demonstrated for the first time that DAT-SPECT with 123I-ß-CIT can clearly visualize and measure the dopaminergic deficit in PD. The paper by Innis et al. (1993) appeared a few months later and independently showed the same results

Full size image

Maximal binding in the striatum was measured 20 h after administration of the tracer. Binding equilibrium of total uptake in the striatum and nonspecific binding in the cerebellum during the acquisition was demonstrated in normal controls and PD patients. A ratio of specific binding in the striatum and nonspecific uptake in the cerebellum (total striatal uptake—nonspecific uptake in the cerebellum/nonspecific uptake in the cerebellum = ratio of total striatal binding/ nonspecific cerebellar uptake-1) during a period of binding equilibrium is directly related to the so called binding potential (transporter density (Bmax) × tracer affinity 1/Kd = Bmax/Kd) of DATs. Thus a semi-quantitative analysis of ß-CIT SPECT by a simple measurement of this ratio is possible and was used in further studies with this ligand.

It was considered a disadvantage and inconvenience that outpatients had to be injected on the first day and image acquisition had to be performed on the next day with ß-CIT. Another problem was seen in the fact that ß-CIT not only had very high affinity for DATs but also bound to serotonin transporters (5HTTs). However, it had been shown that the binding in the striatum was almost exclusive to DATs and that 5HTT binding in the brainstem could easily be separated. This even offered the possibility to measure DAT and 5HTT-binding simultaneously with ß-CIT (Pirker et al. 1995).

Because of the putative disadvantages of 123I-ß-CIT, especially the fact that tracer application and image acquisition could not be performed on one day, other compounds with faster kinetics and higher selectivity were synthesized, such as 123I-FP-CIT, 99mTc-TRODAT and 123I-IP-CIT. 123I-FP-CIT has a lower DAT binding affinity and a lower ratio of specific vs. non-specific binding (“target to background ratio”) than 123I-ß-CIT (Kd 3.5 vs. 1.4 nM; ratio about 3 vs. 7 to 8 in normal controls) and reaches its maximal binding after 3–4 h with a shorter binding equilibrium (Booij et al. 1997). Because semi-quantification is probably less accurate with 123I-FP-CIT, images are often assessed visually. 123I-FP-CIT has been marketed for several years as DAT-Scan® for the diagnosis of PD and the differential diagnosis of Lewy body dementia (LBD) and Alzheimer’s disease, and is now most widely used in clinical practice. Most of the early studies were performed with 123I-ß-CIT and the results subsequently confirmed with 123I-FP-CIT.

All SPECT studies with 123I-ß-CIT and 123I-FP-CIT have shown a high sensitivity for detecting a binding reduction in PD compared with controls. In patients with hemi-Parkinson a reduction of 40–53% of 123I-ß-CIT binding in the striatum contra-laterally to the affected side and a 30–39% reduction ipsi-laterally was measured (Marek et al. 1996; Brücke et al. 1997). This demonstrates that the preclinical lesion can also be detected with this method. An age related loss of 123I-ß-CIT binding of 3–7% per decade was found in normal controls (Volkow et al. 1994; van Dyck et al. 1995; Brücke et al. 1997; Asenbaum et al. 1997). This is consistent with an age-associated reduction of dopaminergic neurons in the substantia nigra (Fearnley and Lees 1991) and underlines the fact that the reduction in PD patients has to be compared with age matched controls or an age matched control value. The early studies of DAT imaging also showed a good correlation of the degree of DAT binding reduction in PD patients with measures of clinical severity like the Hoehn and Yahr (H/Y) stage or UPDRS motor and ADL scores and subscores of bradykinesia, rigidity and axial symptoms (Brücke et al. 1994, 1997; Seibyl et al. 1995; Rinne et al. 1995; Asenbaum et al. 1997; Pirker 2003). No correlation was found with tremor scores which suggests a different pathology (Seibyl et al. 1995; Brücke et al. 1997; Benamer et al. 2000a; Pirker 2003).

All studies show that the reduction in DAT binding in PD is more marked in the putamen than in the caudate nucleus. The reason for this is the selective degeneration of dopaminergic neurons in the ventrolateral part of the substantia nigra, which project to the putamen (Fearnley and Lees 1991). In the visual assessment of DAT-SPECT images the striatum in normal controls has a comma-shaped structure, which is lost in PD patients. This is due to the greater loss of DAT binding in the dorsal part of the comma, which reflects the putamen.

DAT imaging in clinical practice

DAT imaging with SPECT is now part of the routine examinations in patients with Parkinsonism or tremor, where the differential diagnosis is unclear, and dopaminergic degeneration has to be proven or ruled out. A normal DAT SPECT result has been included as an absolute exclusion criterion for a diagnosis of PD (Postuma et al. 2015). It is important that imaging with CT or MRI is performed to detect possible structural pathology. As mentioned in the introduction, patients with essential tremor (ET) or vascular pseudo-parkinsonism are often misdiagnosed by general practitioners or doctors not familiar with movement disorders. Atypical degenerative parkinsonian disorders like multiple system atrophy (MSA) and progressive supranuclear palsy (PSP) and also cases with early PD and isolated rest tremor pose diagnostic problems, even for movement disorder specialists.

Tremor

ET is the most common type of tremor with a prevalence of 3 to 4 per 1000. Typically there is a relatively symmetrical postural and kinetic tremor of the upper limbs (6–8 Hz) which sometimes also affects head, voice and trunk, and rarely the legs. About 10% have additional rest tremor (Marshall and Grosset 2003). A positive family history in about 60% and a variable response to alcohol and beta-blockers is reported (Rajput et al. 1975). Sometimes ET can also be asymmetrical. In elderly patients mild bradykinesia and reduced arm swing may occur and that can cause diagnostic problems. In ET there is no dopaminergic degeneration, and therefore, DAT scan should be normal. This was shown in several studies with 123I-ß-CIT and 123I-FP-CIT with high sensitivity and specificity (Asenbaum et al. 1998; Benamer et al. 2000b).

Monosymptomatic rest tremor is defined as pure rest tremor without other parkinsonian symptoms and a duration of at least 2 years. In some of these patients DAT imaging showed a dopaminergic degeneration (Benamer et al. 2003). Therefore, it seems that these patients had early tremor dominant PD but that a PD diagnosis could not be made according to diagnostic criteria.

Holmes tremor (formally called rubral tremor) is a relatively rare condition. In most cases there is a lesion of dopaminergic neurons and cerebello-thalamic fibres in the midbrain which is often caused by vascular lesions or lesions in MS. It is characterized by a very slow (< 4.5 Hz) rest- and intention tremor with large amplitude (Marshall and Grosset 2003). DAT-SPECT can show a unilateral loss of striatal binding on the side of the lesion and normal uptake on the other side (Pirker 2015).

Dystonic tremor is mainly postural and kinetic with variable frequency and occurs in a body part affected by dystonia. DAT-SPECT is normal.

Postural and kinetic tremor may also occur in demyelinating peripheral neuropathies (Marshall and Grosset 2003).

Vascular Parkinsonism (lower body Parkinson)

The term “vascular parkinsonism” is misleading, because it suggests a pathology similar to PD and a dopaminergic lesion. Typically patients do not have parkinsonian symptoms in the upper limbs and often have a broad based shuffling gait with normal or even exaggerated arm swing. CT and MRI often show a subcortical vascular encephalopathy, and there is usually no response to levodopa. Despite this many of these patients are unnecessarily treated with anti-parkinsonian therapy. DAT-SPECT in these cases is normal (Gerschlager et al. 2002). PD is common and so are vascular lesions in elderly patients. Therefore, many patients with the diagnosis of “vascular parkinsonism” are cases with PD and severe vascular comorbidity (Winikates and Jankovic 1999). In these patients DAT-SPECT is abnormal and a levodopa therapy indicated.

A vascular lesion in the basal ganglia can lead to a reduction of DAT binding in the area of the defect with a normal binding on the other side. Very rare cases with a vascular lesion of the substantia nigra typically also show a unilateral loss of striatal DAT binding. Bilateral structural lesions in the striatum can sometimes cause DAT results, which cannot be distinguished from PD. In these cases MRI or CT is necessary for the interpretation of the DAT findings and for the right diagnosis (Pirker 2015).

Hydrocephalus

Normal pressure hydrocephalus (NPH) NPH is characterized clinically by a frontal type gait disorder, bladder dysfunction and cognitive decline. In some atypical cases parkinsonian symptoms can be present. Typically CT and MRI show marked enlargement of the cerebral ventricles without increase of the subarachnoid space. Sometimes a slight diffuse reduction of striatal DAT binding can be seen (Pirker 2015).

Obstructive hydrocephalus Recently a case report about a reversible DAT reduction in a case of obstructive hydrocephalus with gait disorder was published (Nishida and Yokota 2021). 123I-FP-CIT SPECT apparently showed a diffusely reduced uptake in both striata without asymmetry. A diagnosis of aqueduct stenosis was made and ventriculostomy was performed. Postoperatively neurological symptoms and DAT binding improved. The DAT-SPECT images even made it to the front page of the Annals of Neurology.

Drug-induced parkinsonian syndromes

Many different drugs can cause a blockade of dopamine D2 receptors in the striatum and thereby induce symptoms resembling PD. In most of the cases widely used neuroleptics are responsible. A dopamine D2 receptor blockade can also be induced by antiemetics like metoclopramide or the calcium antagonists flunarizine and cinnarizine used as anti-migraine drugs (Brücke et al. 1995). Neuroleptics which are marketed as “atypical”, such as olanzapine and risperidone, are supposed to cause fewer extrapyramidal side-effects. However, it was shown that they induce high rates of striatal dopamine D2 receptor blockade: Olanzapine 75% (Tauscher et al. 1999), risperidone up to 68% (Küfferle et al. 1996; Dresel et al. 1998). Therefore, they often also cause parkinsonian symptoms.

Truly atypical are clozapine and quetiapine which block D2 receptors only by 20–30% (Küfferle et al. 1997) and have practically no parkinsonian side-effects. Clozapine and quetiapine are also the only neuroleptics which can be used in PD patients.

Tetrabenazine is used to treat dyskinesias and choreatic movements. It acts by depleting dopamine from dopaminergic nerve-endings and, therefore, can also cause parkinsonian symptoms.

Old age is a risk factor for drug induced extrapyramidal symptoms, which can be explained by an age-related loss of dopaminergic neurons and dopamine D2 receptors (Brücke et al. 1995). D2 receptor blocking drugs can also demask preclinical PD. This can be documented with pathological DAT imaging, whereas a normal DAT scan proves a drug induced symptomatology.

Psychogenic (functional) Parkinsonism

Psychogenic Parkinson is a rare condition and was seen only in 2–6% of psychogenic movement disorders (Miyasaki et al. 2003; Fahn and Williams 1988). Often an abrupt onset is common and no history of progression reported. Tremor in psychogenic Parkinson is complex, occurring at rest, posture and voluntary movement. It varies in frequency and rhythmicity and lessens with distraction. In contrast to rest tremor in PD it can be entrained. Slowness can be profound and rigidity resembles voluntary resistance. Spontaneous remissions can occur and a good effect of levodopa is often reported. DAT-SPECT is usually normal and rules out true PD. However, in some cases, coexistence of psychogenic Parkinson and PD may occur (Tolosa et al. 2003).

Atypical neurodegenerative parkinsonian syndromes (APS)

MSA and PSP

The accuracy of the clinical diagnosis of MSA, and PSP is imperfect and the differential diagnosis with PD is difficult, especially in early clinical stages. These syndromes have a more rapid and malignant clinical course and respond less to dopaminergic therapy. The neuropathology is more widespread than in PD with a loss of nigrostriatal dopaminergic neurons and intrinsic neurons in the striatum. In vivo studies with 18F-DOPA PET showed equally severe impairment of putamen and caudate uptake in PSP and a more heterogenous picture in MSA. PD patients, as expected, showed a more severe involvement of the putamen (Brooks et al. 1990). DAT SPECT studies with 123I ß-CIT in MSA and PSP patients demonstrated a marked striatal binding reduction in MSA and PSP in the same magnitude as in PD. However, the finding of an equally impaired putamen and caudate uptake in PSP could not be confirmed. In comparison with PD no significant differences in caudate/putamen ratios or asymmetry between left and right striatum in MSA and PSP were found (Brücke et al. 1997; Pirker et al. 2000).

The first study comparing DAT SPECT findings in patients with autopsy-confirmed MSA and PD was in line with these results. Overall striatal and subregional analyses of 123I ß-CIT binding could not differentiate MSA from PD. Contrary to earlier findings of lower asymmetry of striatal DAT binding in MSA compared with PD (Varrone et al. 2001) a tendency of greater asymmetry in autopsy confirmed MSA patients was found (Perju-Dumbrava et al. 2012).

Whereas DAT SPECT clearly shows the dopaminergic degeneration in MSA and PSP, a differentiation from PD, however, is not possible. Dopamine D2 receptor imaging with 123I-IBZM (Tatsch et al. 1991; Brücke et al. 1993b), MIBG Szintigraphy (Braune et al. 1999) or MRI based techniques are more helpful in differentiating these disorders.

Corticobasal syndrome, corticobasal degeneration (CBS/CBD)

The findings in this rare neurodegenerative disorder are heterogenous. DAT binding can be normal or pathologic. Contrary to MSA and PSP, dopamine D2 receptor imaging with 123I-IBZM is normal in most cases (Pirker 2015).

Wilson’s disease

The most common neurological symptoms in Wilson’s disease are Parkinsonism and dystonia. A 123I-ß-CIT SPECT study demonstrated severe striatal loss of DAT binding in the same range as in PD suggesting severe damage in presynaptic nigrostriatal dopaminergic nerve terminals. This probably contributes to the neurological manifestations of this disorder (Jeon et al. 1998b).

Dementia with Lewy bodies (DLB)

DLB is the second most common form of dementia after Alzheimer’s disease (AD). According to a community based study DLB accounts for 11% and AD for 31% of patients with dementia (Stevens et al. 2002). It is often preceded by REM sleep behaviour disorder (RBD). Fluctuations of cognitive deficits occur, visuo-spatial symptoms are prominent and visual hallucinations are typical. Parkinsonian symptoms are probably less responsive to levodopa and there is hypersensitivity to neuroleptic medication (Swanberg and Cummings 2002). Neuropathological examination shows dopaminergic degeneration and Lewy bodies in the substantia nigra and in the cortex, and Alzheimer type pathology often coexists (Walker et al. 2007). Clinical differentiation of DLB and AD is difficult. Pathologic DAT-SPECT confirms the dopaminergic lesion in DLB, whereas DAT binding in AD is normal (Donnemiller et al. 1997; Ransmayr et al. 2001; Walker et al. 2002). In an autopsy controlled study one patient out of 8 with DLB had normal DAT imaging, while DAT-SPECT was normal in all 9 cases with AD. Thus the sensitivity of DAT based diagnosis of DLB was 88% and the selectivity 100%. The initial clinical diagnosis had a sensitivity of 75% but was specific in only 42% (Walker et al. 2007) (Table 1).

Table 1 DAT-SPECT results in PD and related disorders
Full size table

Dopa responsive dystonia

Dopa responsive dystonia (DRD) is a rare genetic disorder with a defect of dopamine synthesis but intact dopaminergic nerve endings and a normal DAT binding (Naumann et al. 1997; Jeon et al. 1998a). DAT SPECT can help to differentiate this disease from juvenile PD, where DAT binding is reduced.

Other reasons for reduced motility

Subtle parkinsonian signs can also be seen in healthy older persons, in cases of Alzheimer’s disease and in patients with depression. Reduced mobility is often a consequence of orthopaedic problems or rheumatic disease. In all these conditions DAT-SPECT is normal and rules out PD (Pirker 2015).

Technical and practical issues

There are a number of technical and practical issues with DAT-SPECT imaging which have to be observed. Dual or triple head or other dedicated SPECT cameras should be used for brain imaging. Low energy high (or ultrahigh) resolution (LEHR or LEUHR) parallel-hole collimators are most suitable. SPECT should be started when there is an equilibrium of specific striatal tracer uptake (total striatal—nonspecific uptake in the cerebellum or occipital cortex). The time window is 18–24 h post-injection for 123I-ß-CIT and 3–6 h post-injection for 123I-FP-CIT. Each center should maintain a standardized post-injection imaging time for comparability between subjects and intra-individual follow-up studies. Detailed parameters for acquisition and image processing are outlined in the EANM practice guidelines (Morbelli et al. 2020). Amongst others, the amount of tracer dose injected and head movement during the acquisition period can influence and markedly reduce the quality and the resolution of the images obtained. For visual assessment it is important to judge the uptake in the posterior part of the striatum, corresponding to the putamen, in comparison with the caudate nucleus in the anterior striatum and to assess if striatal uptake is symmetric or not. In PD patients reduction of tracer uptake is pronounced in the putamen and thus the normal “commaform” of DAT binding is lost, and the striatal uptake usually is asymmetric (Fig. 3).

Fig. 3
figure 3

Striatal 123I-FP-CIT binding in an age matched healthy control (A); 56-year-old PD patient 1 year after diagnosis (B); 2 years after diagnosis (C). Axial slices parallel to the canto-meatal plane. Marked loss of DAT binding in the putamen on the left more than on the right side. The posterior part of the right putamen is still visible, on the left only the anterior part can still be delineated (B). Further decrease of DAT binding in the left caudate and posterior and anterior right putamen. The left putamen cannot be delineated anymore (C). Example of significant loss of DAT binding in early PD. (Images from Isotopix institute, Vienna)

Full size image

Semi-quantitative measurement of the results is influenced by the right placement of regions of interest (ROIs) and the size of these regions which are usually drawn on axial slices. The most common outcome measure is the specific binding ratio (striatal target to background ratio) which is used both in clinical routine and research. Misplacement and too large ROIs lead to wrong results. If MRI or CT is available, ROIs drawn with overlay techniques or volumetric regions of the basal ganglia can be used. Another approach is voxelwise analysis combined with statistical parametric mapping. Fully automated methods are available with registration of the SPECT scan to a template or to an averaged, spatially normalized brain volume. The data obtained are dependent on the acquisition and reconstruction parameters and the corrections applied. As a result no universal cutoffs for normal vs. abnormal exist, and therefore, each site should ideally establish its own reference values using a healthy age-matched control group (EANM practice guidelines; Morbelli et al. 2020). Test retest variability is usually in the same range or even larger than annual DAT reductions in longitudinal SPECT studies. It, therefore, does not make sense to determine progression rates in individual patients. Such studies can only be performed in patient groups. It also has to be kept in mind that the image resolution in SPECT studies is in the range of 8–10 mm (whereas resolution in PET reaches 4–5 mm).

In the end, results in a clinical setting also depend on the experience of nuclear medicine physicians or radiologists. Not seldom there is a tendency to overdiagnose the results, leading to misdiagnosis of a dopaminergic lesion and unnecessary, sometimes harmful treatment.

In multicenter trials the acquisition protocol and image reconstruction and filtering should be harmonized as much as possible. To overcome variability, image reconstruction can also be performed in one central facility. The use of an anthropomorphic striatal phantom can help to develop correction factors for standardization between different cameras.

DAT imaging in clinical research

The most important topic and goal in PD research is finding a neuroprotective or disease modifying treatment. Several trials with this aim have been unsuccessful. One major problem is that measurements of disease progression based on clinical examinations alone are influenced by anti-parkinsonian treatment. Dopaminergic imaging either with PET or SPECT based methods might provide an objective measure and biological marker of nigrostriatal dopaminergic dysfunction. These methods had shown high sensitivity in detecting dopaminergic degeneration in PD and APS, and a correlation of the degree of striatal dopaminergic deficit and measures of motor impairment were found (see above). To study disease progression many longitudinal studies with F-DOPA PET and DAT-SPECT were performed.

Measurement of disease progression

Several studies with dopaminergic imaging in PD have demonstrated a loss of either striatal DAT binding or F-DOPA uptake of about 40–60% at the time of emergence of PD symptoms (Marek et al. 1996; Asenbaum et al. 1997; Brücke et al. 1997; Booij et al. 1997; Morrish et al. 1998). Early longitudinal studies with F-DOPA PET and DAT-SPECT imaging reported annual reduction rates of 5–11% (Morrish et al. 1998; Marek et al. 2001; Nurmi et al. 2000; Pirker et al. 2002). In the study by Pirker et al. the progression rate in PD patients was compared with a group of patients with atypical parkinsonian syndromes (APS) and ET. The annual progression rate in early PD patients was 7%, whereas no significant progression was found in PD patients with longer disease duration and in ET patients. APS patients showed the fastest annual progression with 15%, which is in line with a faster decline of clinical symptoms in these disorders. These findings also show that the disease progression in PD seems to slow down in the course of the disease and suggests an exponential decline. Another progression study in PD patients with relatively long disease duration of 7 years, found a low but significant reduction of striatal DAT over 2 years. Sub-regional analysis demonstrated a more marked progression rate in those regions which were less affected at the time of the first scan (caudate nuclei ipsi- and contralateral to the more affected side 2% progression of the normal mean per year versus 0.6% in the contralateral putamen, which was most affected at the time of the first scan). This suggests that the process of degeneration in these patients had already levelled off in the contralateral putamen which showed pronounced degeneration at the time of the first scan, but was still ongoing in the caudate nuclei which had been less affected at baseline (Brücke et al. 2000). A DAT-SPECT study by Staffen et al. (2000) found a 6.8% decrease of DAT binding in PD patients with H/Y stage I and only 1.25% in H/Y stage III during a 15 month period. These results also point to an exponential decline of dopaminergic function in PD. An example of a relatively marked loss of DAT binding between the 1. and 2. year after PD diagnosis is given in Fig. 3.

Two larger longitudinal DAT-SPECT and F-DOPA PET studies aimed to prove a possible neuroprotective effect of the dopamine agonist’s pramipexol and ropinirol, respectively, versus levodopa (Parkinson Study Group 2000; Whone et al. 2003). Early PD patients initially treated with pramipexol or ropinirole were compared with patients treated with levodopa during a 22–46 month evaluation period (Parkinson Study Group 2002). Both the CALM-PD CIT and the REAL-PET trials demonstrated a relative reduction in the rate of loss of striatal 123I-ß-CIT and 18F-DOPA uptake of 35–47% in patients treated with pramipexol or ropinirol, respectively. These imaging data suggested that pramipexole or ropinirole may retard and/or that levodopa may accelerate the degeneration of dopaminergic nigrostriatal neurons. However, the clinical outcome did not show that these drugs accelerate or retard PD progression. The question remains whether the imaging markers are directly up- or down-regulated by longterm treatment with these drugs or if they really reflect dopaminergic nerve cell loss in the substantia nigra (Marek et al. 2003). Similar considerations resulted from the ELLDOPA trial comparing the progression of PD patients on placebo versus levodopa. This trial also showed a more rapid decline of striatal 123I-ß-CIT over 40 weeks in the levodopa group, but the clinical data suggested that levodopa slowed the progression of PD. The clinical assessment at the end of the study was performed 2 weeks after stopping treatment, and a potential long term effect of levodopa could also have explained the clinical results (Fahn et al. 2004).

The PRECEPT trial (Parkinson Study Group 2007) is a very large longitudinal study on neuroprotection in early PD patients not yet requiring dopaminergic therapy. The antiapoptotic compound CEP-1347, which had shown neuroprotective effects in a variety of in vitro and in vivo animal models was tested in a placebo controlled trial in 806 PD patients in different doses. The primary endpoint was the time to development of disability requiring dopaminergic treatment. The trial was stopped after 21,4 months of follow-up because of futility. 123I-ß-CIT SPECT was performed in 799 patients at baseline and after 22 months. In fact ß-CIT uptake in patients in the active treatment arms deteriorated faster than in the placebo group. This could be due to an accelerated DAT loss or to a down-regulation of DAT expression by CEP-1347. Changes in UPDRS scores and ß-CIT imaging results showed similar patterns. The annualized percent changes of striatal, caudate and putamen ß-CIT binding were highly significantly larger in those patients who developed disability requiring dopaminergic therapy (− 5.5%; − 4.2% and − 7.9%, respectively) compared with subjects who did not reach this endpoint (− 2.6%; − 1.5%; − 4.5%, respectively). Although this study demonstrated that CEP-1347 was ineffective in early PD it shows that DAT imaging seems to be a reliable biomarker for disease progression in very early stages of the disease.

A recent meta-analysis of presynaptic dopaminergic imaging studies in PD included 141 PET or SPECT studies and 3605 patients. This analysis included 3 longitudinal aromatic amino acid decarboxylase (AADC) studies with PET (using 18F-DOPA or 11C-DOPA) and 15 longitudinal DAT studies with PET or SPECT (using 123I-ß-CIT, 123I-FP-CIT, 18F-FP-CIT, 11C-ß-CIT) (Kaasinen and Vahlberg 2017). The large number of dopaminergic imaging studies included in this meta-analysis demonstrates the enormous interest and activity in the field. The results showed a consistently smaller AADC than DAT reduction, suggesting up-regulation of AADC function in PD. Crossectionally the correlation with disease severity and loss of dopaminergic function appears linear, but the majority of longitudinal studies point to a negative exponential progression pattern of the dopaminergic lesion.

The Parkinson’s progression marker initiative (PPMI) is a large ongoing international multicenter study, designed to identify clinical, DAT imaging, genetic and biospecimen PD markers to accelerate disease modifying therapeutic trials (Marek et al. 2018). This study includes more than 400 untreated patients with early PD and 196 controls. A reduction of 45% in mean striatal and of 68% in contralateral putamen DAT binding was found. The study aims to establish longitudinal progression biomarkers for future treatment trials in PD.

SWEDD

This term refers to the absence of an imaging abnormality in patients clinically presumed to have PD. In the above listed larger longitudinal studies with DAT-SPECT or F-DOPA PET 4–20% of the patients included initially had normal imaging results. There has been a long discussion and controversy in the medical literature about the question what these patients actually had. Clinical and imaging follow-ups in the vast majority, neither showed progression in clinical markers of motor impairment, nor in the imaging results. It now seems quite clear that most “SWEDD” cases were due to a clinical misdiagnosis. Most of these patients probably have dystonic tremor and in the others all above mentioned differential diagnoses might apply. However, singular cases did show clinical progression, a positive levodopa response, imaging and/or genetic evidence in the follow-ups. It also is quite probable that some scan results were incorrect at baseline, because DAT scan is not 100% sensitive and specific (Erro et al. 2016).

DAT imaging in preclinical PD

DAT SPECT enables to detect a nigrostriatal dopaminergic deficit in very early or presymptomatic stages. A preclinical DAT defect has been demonstrated in patients with hemiparkinson in the ipsilateral striatum, as has been mentioned above (p4). Subjects with pre-motor symptoms of PD like hyposmia and RBD can have pathological DAT-SPECT results suggesting a dopaminergic degeneration (Ponsen et al. 2004; Eisensehr et al. 2000; Iranzo et al. 2017).

Studies in non-manifest and symptomatic carriers of PD genes

DAT SPECT has also been used in subjects carrying mutations of genes associated with PD. Varrone and Pellecchia (2018) have reviewed DAT SPECT imaging results in familial PD from a large number of studies. Cases with GBA, alpha-synuclein, LRRK2, Parkin and PINK1-induced mutations were included. The largest cohorts were patients with Parkin mutations and patients and non-symptomatic carriers with LRRK2 mutations. Up to 44% of LRRK2 non-symptomatic carriers and almost all symptomatic carriers in all groups had pathologic DAT scan results. Asymmetry of DAT binding seems more pronounced in the LRRK2 group.

The PPMI study included 208 LRRK2 and 184 GBA mutation carriers without manifest PD. 286 had DAT-SPECT imaging. Of these 18 (11%) LRRK2 and four (3%) GBA carriers had pathological DAT scans. Compared with healthy controls there was a significant increase of subtle motor signs of PD in the whole group of non-manifesting carriers that can precede DAT deficit (Simuni et al. 2020).

Correlation of dopaminergic imaging with neuropathology in PD

Dopaminergic imaging results and clinical endpoints in PD trials are often discordant, questioning the validity of these imaging markers to reflect the degree of dopaminergic degeneration and disease severity. Literature on the correlation of dopaminergic imaging with post-mortem neuropathological findings in PD is scarce. An early study in a small group of patients with various neurodegenerative disorders [1 PD, 3 PSP, 1 amyotrophic lateral sclerosis, 1 Alzheimers disease (AD)] found that the striatal 18F-dopa uptake was correlated with cell density in the substantia nigra and striatal dopamine levels. The correlation with cell density was strictly proportional, while dopamine concentrations were relatively more reduced (Snow et al. 1993).

That DAT imaging also can give a useful insight into the status of the nigrostriatal system was shown by two studies in humans comparing imaging results with neuropathological findings (Colloby et al. 2012; Kraemmer et al. 2014). Colloby et al. studied an autopsy confirmed group of patients with AD (n = 4), Parkinson’s disease with dementia (n = 12) and DLB patients (n = 7) and compared nigral dopaminergic cell loss and alpha-synuclein, tau or ß-amyloid pathology with striatal 123I-FP-CIT binding. A correlation was found for DAT binding and nigral cell loss but not for the other parameters.

Kraemmer and colleagues studied one patient with PD, two with DLB, one with MSA, two with CBD, one with atypical parkinsonism, one AD patient and one with Creutzfeld–Jakob disease who had been studied with 123I-ß-CIT SPECT and compared the imaging results with neuropathological findings. Striatal DAT binding correlated highly (r = 0.97) with post-mortem substantia nigra cell counts (tyrosin hydroxylase- and neuromelanin-positive cells).

In contrast to these studies Saari et al. (2017) found no association of striatal DAT binding and substantia nigra cell counts (tyrosin hydroxylase- and neuromelanin-positive cells) in 11 PD, 5 MSA, 1 CBD and 1 PSP patients. Thirteen of these patients had been scanned with123I-FP-CIT and 5 with123I-ß-CIT. The same group (Honkanen et al. 2019) performed tyrosine–hydroxylase positive nerve fibre counts in the putamen of 10 PD patients and 4 patients with atypical parkinsonian syndromes [subsample of the previous study of Saari et al. (2017)] and correlated the results with striatal DAT binding. Again no correlation was found. These data were questioned because of long intervals between scans and death (up to 13 years) and death and autopsy (up to 8 days). A long interval between scans and death means that imaging was done early in the disease process with still high striatal DAT binding and a long interval between death and autopsy can affect tyrosine–hydroxylase positive axon counts due to autolysis. In fact a positive correlation exists for striatal DAT binding and the time interval of imaging and death and a negative correlation for putaminal fiber counts and the interval between death and autopsy in the study by Honkanen et al. (Pifl 2019).

Several animal studies correlating striatal DAT binding and post-mortem findings of a nigrostriatal dopaminergic lesion also showed somewhat inconsistent results (Kraemmer et al. 2014). A study in monkeys with a one-sided MPTP lesion and PET imaging with 18F-Dopa, 11C-dihydrotetrabenazine (DTBZ), a marker for the vesicular monoamine transporter (VMAT2) and the DAT tracer 11C-CFT found a correlation with cell density of dopaminergic neurons in the substantia nigra only when this cell loss was < 50%. However, all PET measures correlated with striatal dopamine levels over the full range of dopamine depletion. A linear correlation was found between the severity of hemiparkinsonism in the monkeys and the full range of nigral cell loss (Karimi et al. 2013). On the other hand a strong correlation of DAT uptake as measured with11C-CFT and 11C-DTBZ binding to VMAT2 in the substantia nigra and nigral dopaminergic cell counts was found in the same monkey model. In addition, uptake of both markers had a linear relationship with motor impairment (Brown et al. 2013). The discrepancy between striatal measures and nigral cell loss above a certain threshold points to a levelling off effect of striatal dopaminergic markers, and questions if these markers truly reflect disease severity in later more progressed stages of the disease (Perlmutter and Stoessl 2019).

In fact, a large recent longitudinal study in early PD patients over 5 years found that the rate of change of 123I-FP-CIT DAT imaging was largest during the first year. The rate of change of clinical rating scores and DAT imaging over the whole 5 year period did not correlate (Simuni et al. 2018). As mentioned above, this is also in line with earlier DAT imaging studies suggesting a levelling off of annual reduction rates with disease progression (Staffen et al. 2000; Brücke et al. 2000; Pirker et al. 2002). Thus striatal DAT imaging may not be an ideal measure of PD severity beyond the early stage of the disease and, therefore, not optimal as a biomarker to assess the effect of possible neuroprotective or disease-modifying therapies (Perlmutter and Stoessl 2019).

Conclusion and outlook

DAT–SPECT has established its role in clinical practice as a diagnostic test which proves or disproves dopaminergic degeneration in patients with an unclear diagnosis. It can distinguish patients with PD or atypical parkinsonian syndromes (APS) with dopaminergic degeneration from patients with disorders which are often misdiagnosed as PD, such as ET or vascular Parkinsonism. However, differentiation of PD and APS with DAT imaging is not possible. In patients with suspected early PD a normal DAT SPECT can refute the diagnosis, which is important for clinical trials. DLB patients can be distinguished from patients with AD with high specificity and selectivity. A preclinical dopaminergic dysfunction can be detected in patients with prodromal symptoms of PD such as RBD and hyposmia and in carriers of different PD genes.

The role of DAT SPECT as a possible biomarker in larger clinical trials on neuroprotection is still unclear and complex. The decrease of striatal DAT expression over the course of PD is not linear and levels off in later stages. It probably correlates with dopaminergic cell loss in the substantia nigra only in early disease stages when this cell loss is still below 50%. DAT expression is also down-regulated as a compensatory mechanism to keep up synaptic dopamine levels early in the disease course and might be influenced directly by long-term drug treatment.

Interest in functional imaging of the dopaminergic system is constantly rising. There are recent technical improvements in the performance of cameras for SPECT and PET, new image analysis techniques are in the offing and new radiopharmaceuticals are being developed. Large-field cadmium–zinc–telluride (CZT) SPECT cameras provide high performances in terms of count sensitivity and spatial resolution and are now also used for brain imaging. Image quality is much higher with these cameras and detector ring-configuration may further enhance resolution (Verger et al. 2021; Bordonne et al. 2020; Bani Sadr et al. 2019). A whole range of new PET radiotracers are being developed not only for dopaminergic imaging but also for targeting alpha-synuclein and tau-aggregates and for studying neuroinflammation. It is hoped that these advances will help to increase our understanding of different neurodegenerative diseases, and to develop disease-modifying treatments