Early perfusion and dopamine transporter imaging using ¹⁸F-FP-CIT PET/CT in patients with parkinsonism

Abstract

Combined use of ¹⁸F-N-(3-fluoropropyl)-2β-carboxymethoxy-3β-(4-iodophenyl)nortropane (FP-CIT) for dopamine transporter imaging and ¹⁸F-fludeoxyglucose (FDG) for glucose metabolism shows good diagnostic performance for differential diagnosis of Parkinson disease (PD) and Parkinson plus syndrome (multiple system atrophy, progressive supranuclear palsy, corticobasal degeneration, and dementia with Lewy bodies). A recent study showed that ¹⁸F-FP-CIT positron emission tomography (PET) with early perfusion imaging is useful for the differential diagnosis of PD and Parkinson plus syndrome with lower radiation exposure, time, and cost. In this review, we summarize the advantages of using ¹⁸F-FP-CIT PET for perfusion and dopamine transporter imaging, as well as clinical features useful for the differential diagnosis of PD and Parkinson plus syndrome.

Introduction

Parkinsonism is a syndrome of movement disorders characterized by a combination of six cardinal features including, bradykinesia, rigidity, resting tremor, postural instability, flexed posture, and freezing [1]. Parkinsonism shows the highest frequency among referral movement disorders in the clinic, and the causes of parkinsonism are variable. Parkinson disease (PD) is the most common form of parkinsonism. Parkinson plus syndrome, such as multiple system atrophy (MSA), progressive supranuclear palsy (PSP), corticobasal degeneration (CBD), and dementia with Lewy bodies (DLB), is another cause of parkinsonism that should be distinguished from PD [2].

Parkinson plus syndrome exhibits the following additional neurological features: autonomic dysfunction and cerebellar ataxia in MSA; supranuclear ophthalmoparesis in PSP; a combination of dystonia, myoclonus, apraxia, cortical sensory deficit, and alien hand in CBD; and early, prominent dementia in DLB [3].

To date, the therapeutic strategies for these chronic neurodegenerative disorders are based on symptomatic treatments. Dopaminergic medications are well-known representative drugs and are remarkably effective for symptom relief in patients with PD [4]. In contrast, no effective treatment has been developed for patients with Parkinson plus syndrome, and the prognosis of patients with Parkinson plus syndrome is poorer than that of patients with PD [5]. Accurate differential diagnosis will help not only the clinician to determine treatment regimens and prognosis but also the patient to plan rest of their life. Recently, several clinical trials of patients with Parkinson plus syndrome have been conducted to investigate new therapeutic drugs, and the importance of achieving an accurate differential diagnosis in the early stage of Parkinson plus syndrome has been emphasized. As there are no reliable biomarkers for any of these disorders, the proposed clinical diagnostic criteria based on the presence of clinical features are used for differential diagnosis of parkinsonism. However, this condition is challenging to accurately diagnose because patients with parkinsonian disorders show similar symptoms, and distinct symptoms do not occur in the early stage [6].

Dopamine transporter imaging

To evaluate the presynaptic nigrostriatal dopamine neurons in striatum, positron emission tomography (PET) and single-photon emission computed tomography (SPECT) technology have been used since the 1980s. ¹⁸F-dihydroxyphenylalanine (DOPA), which reflects aromatic L-amino acid decarboxylase (AADC) activity, was initially introduced, but there were several factors limiting its widespread use in clinics [7]. ¹⁸F-DOPA PET is an indirect imaging method of the presynaptic nigrostriatal neurons, and some reports observed upregulation of AADC activity in patients with early PD [8]. Therefore, other targets for dopamine neurons were developed.

The terminals of nigrostriatal dopaminergic neurons express dopamine transporters (DAT), which are responsible for dopamine uptake from the synaptic cleft, making DAT a distinct molecular target (Figure 1) [9]. Recently, many radiotracers were developed to evaluate presynaptic DAT function: ¹²³I-2β-carbomethoxy-3β-(4-iodophenyl)tropane (β-CIT) (Figure 2A), ¹²³I-N-(3-fluoropropyl)-2β-carboxymethoxy-3β-(4-iodophenyl)nortropane (FP-CIT) (Figure 2B), ¹²³I-altropane, ^99mTc-TRODAT for SPECT imaging, and ¹¹C-RTI 32, ¹¹C-CFT, ¹¹C-methylphenidate, ¹¹C-nomifensine, and ¹⁸F-FP-CIT (Figure 2C) for PET imaging [9].

Figure 1.

Schematic diagram of dopamine neuronal synapse in striatum. Nigrostriatal dopamine neuron releases the dopamine into the synaptic cleft in striatum, and the released dopamine binds to dopamine receptor in postsynaptic neuron. Dopamine transporters (DAT) on the membrane of presynaptic dopamine neuron reuptake the dopamine in synaptic cleft. Vesicular monoamine transporters (VMAT) transport the dopamine into the vesicles, and vesicles release dopamine to the synaptic cleft. DA, dopamine; VMAT, vesicular monoaminergic transporter; DAT, dopamine transporter; DR, dopamine receptor.

Figure 2.

Chemical structures of (A) ¹²³I-2β-carbomethoxy-3β-(4-iodophenyl)tropane (β-CIT), (B) ¹²³I-N-(3-fluoropropyl)-2β-carboxymethoxy-3β-(4-iodophenyl)nortropane (FP-CIT), and (C) ¹⁸F-FP-CIT.

¹²³I-β-CIT shows a good striatal/cerebellar uptake ratio, but 24 h are required to equilibrate the tracer after intravenous injection [9]. ¹²³I-β-CIT also non-selectively binds to noradrenaline and serotonin transporters with a similar or lower nanomolar affinity [10]. Therefore, new tracers with better kinetics and DAT selectivity were developed. DAT imaging using ¹²³I-FP-CIT can be performed between 3 and 6 h after intravenous injection and shows better DAT selectivity than ¹²³I-β-CIT [11,12]. DAT imaging using ¹²³I-FP-CIT is widely used and commercially available in Europe (since 2000) and the United States (since 2011). Several multicenter studies have evaluated the use of ¹²³I-FP-CIT [13,14].

FP-CIT also can be labeled with ¹⁸F for PET imaging without modifying the chemical structure. ¹⁸F-FP-CIT has several advantages for DAT imaging compared to ¹²³I-FP-CIT. PET provides higher resolution than SPECT. The half-life of ¹⁸F (109.8 min) is relatively longer than those of other PET radionuclides (e.g., ¹¹C), and ¹⁸F is better for the production and commercialization of PET probes. One of the minor metabolites (<4%) of FP-CIT is nor-β-CIT, which is hydrophilic [15]. ¹²³I still labels nor-β-CIT, but ¹⁸F is detached from nor-β-CIT and converted to a hydrophilic metabolite. As nor-β-CIT can be taken up by the brain, ¹⁸F-FP-CIT shows better kinetics than ¹²³I-FP-CIT [11]. A recent study comparing ¹²³I-FP-CIT and ¹⁸F-FP-CIT showed that visual analyses using both methods did not affect diagnostic accuracy, while semi-quantitative analysis resulted in better contrast of ¹⁸F-FP-CIT PET/CT relative to ¹²³I-FP-CIT SPECT/CT [16]. The low radiochemical yield of ¹⁸F-FP-CIT was a major factor limiting its widespread clinical use; however, recent advancements in ¹⁸F labeling methods of FP-CIT have resulted in high radiochemical yields (>50%) with high reproducibility, enabling ¹⁸F-FP-CIT to be used clinically and commercially [17].

¹⁸F-FP-CIT binding in the normal putamen decreases with aging at the rate of 5.3-6.4% per decade [11,18], which is comparable to the results of multicenter studies examining healthy controls by ¹²³I-FP-CIT [13,14]. ¹²³I-FP-CIT is not effective for differentiating PD from Parkinson plus syndrome [19], but ¹⁸F-FP-CIT shows significant diagnostic value for the differentiation of not only PD but also Parkinson plus syndrome [20]. As PET provides higher anatomic resolution than SPECT, ¹⁸F-FP-CIT can provide better regional characteristic features of PD, PSP, and MSA. Especially, decreased uptake of caudate nucleus in PSP and ventral putamen of MSA were reported using ¹⁸F-FP-CIT PET/CT [20], but there is no report related to regional characteristics of Parkinsonism plus syndrome in ¹²³I-FP-CIT SPECT in our knowledge.

Perfusion imaging of ¹⁸F-FP-CIT

¹⁸F-FDG is the most commonly used radiotracer for assessing regional cerebral glucose metabolism as a marker of neuronal function. Previous studies and preliminary meta-analysis showed that ¹⁸F-FDG PET is highly accurate (>90%) for distinguishing between PD and Parkinson plus syndrome [21]. As it is classically considered that cerebral blood flow and glucose metabolism are tightly coupled [22,23], perfusion imaging can be used as an alternative to glucose imaging. A recent comparison study using ¹⁵O-water for perfusion imaging and ¹⁸F-FDG for glucose metabolism revealed some differences [22], that validating role of perfusion imaging for the differential diagnosis of Parkinson plus syndrome might be needed.

¹⁸F-FP-CIT shows rapid increase of tracer uptake in brain with high extraction fraction rate after intravenous injection, which suggests that it has a high membrane permeability [11]. Early perfusion uptake of ¹⁸F-FP-CIT in dopamine-poor regions (such as the cerebral cortex or cerebellum) shows peak around first 10 minutes after injection, DAT binding might be affected after 10 min [11,24]. Therefore, early imaging within 10 min after intravenous injection of ¹⁸F-FP-CIT well represents perfusion flow and mimics glucose metabolism in the brain, and several studies have been performed to evaluate the usefulness for perfusion imaging using ¹⁸F-FP-CIT for the differential diagnosis of PD and Parkinson plus syndrome [24,25]. However, perfusion images are typically obtained for 5-10 min after tracer injection, producing images with relatively lower quality than those obtained using ¹⁸F-FDG, which hamper imaging interpretation in small regions such as the midbrain [24].

Differential diagnosis using ¹⁸F-FP-CIT

Parkinson disease

Typical DAT images of PD show reduced tracer uptake in the putamen and caudate nucleus with a rostrocaudal gradient (anteroposterior gradient), and the caudate nucleus is less affected than the putamen [26]. These findings are well-known for DAT images using SPECT. As PET provides much higher resolution than SPECT, maximum intensity projection (MIP) images and coronal images are helpful for the differential diagnosis of parkinsonism. Not only the rostrocaudal gradient, but also the ventrodorsal gradient of putaminal DAT loss is a hallmark of PD in ¹⁸F-FP-CIT images [20]. Therefore, PD patients show preserved DAT in the caudate nucleus and ventral putamen compared to in the posterior putamen. MIP images well visualize these characteristic findings of DAT images (Figure 3). As essential tremor or drug-induced parkinsonism shows normal DAT in the striatum, DAT imaging is very helpful for differentiating the causes of tremor from PD. Additionally, a recent study showed that lower DAT uptake may be related to the partial recovery of drug-induced parkinsonism [27]. However, it is difficult to differentiate PD from Parkinson plus syndrome using DAT images.

Figure 3.

Parkinson disease (PD) (A) Maximum intensity projection image of DAT imaging using ¹⁸F-FP-CIT showed pronounced decreased tracer uptake in the posterior putamen and relative preservation of the caudate nucleus and ventral putamen. (B) Transaxial DAT image shows the rostrocaudal gradient of the putamen (yellow arrow, posterior putamen), and (C) coronal DAT image shows the ventrodorsal gradient of the putamen (yellow arrow, posterior putamen; white arrow, ventral putamen). (D) Early perfusion image of ¹⁸F-FP-CIT and (E) glucose metabolism of ¹⁸F-FDG did not show abnormalities in the striatum.

¹⁸F-FDG PET scans of PD patients appear normal. Upon voxel-based statistical analyses and close inspection, PD is characterized by hypometabolism of the posterior temporoparietal, occipital cortex, and sometimes frontal cortex [21]. Perfusion images also show hypoperfusion in similar cortical regions [28]. From a clinical perspective, the role of early perfusion and metabolism imaging is more focused on the differential diagnosis of PD to Parkinson plus syndrome than on PD to normal cases.

Multiple system atrophy

MSA presents parkinsonian features of varying severity, cerebellar ataxia, autonomic failure, urogenital dysfunction, and corticospinal disorders, and is divided into two categories: MSA with predominant parkinsonism (MSA-P) and MSA with predominant cerebellar ataxia (MSA-C) [29]. The 2008 consensus statement on the diagnosis of MSA includes ¹⁸F-FDG and DAT imaging as additional features of possible MSA [30].

It is known that MSA-P patients are more predominantly affected by striatal DAT loss than MSA-C patients [26]. Kim et al. reported that all MSA-P patients showed striatal DAT loss, while 33% of MSA-C patients showed striatal DAT loss [31]. The higher resolution of PET using ¹⁸F-FP-CIT enables discrimination of the ventral putamen and posterior putamen. MSA shows striatal DAT loss without a ventrodorsal gradient, with MSA suspected based on DAT images (Figure 4) [20]. However, the posterior putamen to ventral putamen ratio overlaps between patients with PD and MSA, making it difficult to establish a cut-off value of the ratio for differential diagnosis of PD and MSA. At the cutoff value for the ratio of the posterior putamen to ventral putamen (>0.65), sensitivity and specificity for differentiating MSA from PD were 90% and 45% respectively [20].

Figure 4.

Multiple systemic atrophy with predominant parkinsonism (MSA-P) (A) Maximum intensity projection image of DAT imaging using ¹⁸F-FP-CIT showing decreased tracer uptake in the posterior putamen and ventral putamen, and relative preservation of the caudate nucleus. (B) Transaxial DAT image showing decreased uptake of the putamen with the predominant left side, and (C) coronal DAT image showing equally decreased uptake in the dorsal and ventral putamen with loss of the ventrodorsal gradient. (D) Early perfusion image of ¹⁸F-FP-CIT and (E) glucose metabolism of ¹⁸F-FDG showing decreased perfusion and glucose hypometabolism in the posterior putamen. (F) FLAIR MR image showing atrophic changes in the left putamen and high signal intensity at the rim of the left putamen.

MSA shows decreased glucose metabolism at the striatum (particularly the posterior putamen) and cerebellum [21]. MSA-P patients showed decreased glucose metabolism at the striatum, with 50% of patients with MSA-P showing cerebellar hypometabolism. Patients with MSA-C showed decreased glucose metabolism at the cerebellum and 13% of patients with MSA-C showed striatal hypometabolism [31].

In early perfusion imaging using ¹⁸F-FP-CIT, the cerebellum in MSA-C and putamen in MSA-P showed significant hypoperfusion compared to PD [25]. Particularly, MSA-C showed distinct hypoperfusion at the cerebellum compared to PD, resulting in clearer discrimination than ¹⁸F-FDG [24]. In a comparison of putamen uptake between MSA-P and PD, perfusion images using ¹⁸F-FP-CIT were less distinguishable than glucose metabolism using ¹⁸F-FDG [24].

The designation of MSA-P or MSA-C refers to the predominant feature at the time the patient is evaluated, which can change over time [30]. Thus, predominant imaging features (such as DAT loss or cerebellar hypometabolism) are not used for the designation of MSA-P or MSA-C using current diagnostic criteria. However, combined interpretation of predominant clinical features and imaging finding is very helpful for diagnosing MSA subtypes. In patients with parkinsonian features without evident ataxia, demonstration of cerebellar hypometabolism or hypoperfusion can indicate a diagnosis of MSA-P rather than PD. In contrast, in the absence of parkinsonian features in a patient with cerebellar ataxia, evidence of DAT loss may point to the diagnosis of MSA-C (Figure 5) [32].

Figure 5.

Multiple systemic atrophy with predominant cerebellar ataxia (MSA-C) (A) Maximum intensity projection image of DAT image using ¹⁸F-FP-CIT showing decreased tracer uptake in the posterior putamen and ventral putamen, and relative preservation of the caudate nucleus. (B) Transaxial and (C) coronal DAT images showing decreased uptake in the dorsal and ventral putamen and loss of the ventrodorsal gradient. (D) Transaxial T2 MRI showing atrophic changes in the cerebellum and pons (“hot-cross bun sign”, yellow arrow). (E) Transaxial early perfusion image of ¹⁸F-FP-CIT and (F) glucose metabolism of ¹⁸F-FDG showing no abnormality in the striatum. (G) Sagittal early perfusion imaging of ¹⁸F-FP-CIT and (H) glucose metabolism of ¹⁸F-FDG showing decreased perfusion and glucose hypometabolism in the cerebellum.

Progressive supranuclear palsy

The National Institute of Neurological Disorders and Stroke (NINDS) and Society for PSP (SPSP) international workshop proposed criteria for diagnosing classic PSP [33,34]. The International Parkinson and Movement Disorder Society (MDS)-endorsed PSP Study Group provided revisions of the NINDS-SPSP criteria in 2017 [35]. Classic PSP, Richardson syndrome (PSP-RS), accounts for only 24% of 100 autopsy-confirmed PSP cases [36], and various subtypes of PSP have been proposed; initial predominance of ocular motor dysfunction (PSP-OM), postural instability (PSP-PI), Parkinsonism resembling idiopathic Parkinson disease (PSP-P), frontal lobe cognitive or behavioral presentations (PSP-F), including behavioral variant frontotemporal dementia (bvFTD), progressive gait freezing (PSP-PGF), corticobasal syndrome (PSP-CBS), primary lateral sclerosis (PSP-PLS), cerebellar ataxia (PSP-C), and speech/language disorders (PSP-SL), including nonfluent/agrammatic primary progressive aphasia (nfaPPA) and progressive apraxia of speech (AOS) [35].

DAT images show markedly reduced tracer uptake in the putamen and caudate nucleus (small rostrocaudal gradient and symmetric) in PSP-RS [26,37], and the caudate nucleus is more affected than in PD [20,24,25,38,39]. Although advanced PD can show a decreased anterior caudate nucleus, the symptom duration of PSP would be much shorter than that of advanced PD because of the rapid progression of PSP [20]. The ratio of the anterior caudate to the ventral striatum showed excellent diagnostic value for differentiating PSP-RS from PD (cutoff, <0.7; sensitivity, 94%; specificity, 92%) [20]. Some studies did not show consistent diagnostic performance using caudate uptake [40,41], but they performed analysis using ¹²³I-β-CIT SPECT, and the lower resolution may have affected the results. ¹⁸F-FP-CIT PET shows higher resolution than SPECT and can clearly visualize decreased tracer uptake in the caudate nucleus (Figure 6). Additionally, midbrain tracer uptake is decreased in PSP-RS, with lower uptake than in PD but similar uptake as in MSA [42,43]. Few studies have evaluated DAT imaging features in the other subtypes of non-PSP-RS [37]. PSP-PGF (also known as pure akinesia with gate freezing), PSP-P, and PSP-RS showed similar DAT imaging patterns of the striatum [37,38,44-46].

Figure 6.

Progressive supranuclear palsy-Richardson syndrome (PSP-RS) (A) Maximum intensity projection image of DAT imaging using ¹⁸F-FP-CIT showing prominently decreased tracer uptake in the caudate nucleus and decreased tracer uptake in the putamen. (B) Transaxial and (C) coronal DAT images showing decreased tracer uptake in both the anterior caudate nucleus (left dominant). (D) Early perfusion image of ¹⁸F-FP-CIT and (E) glucose metabolism of ¹⁸F-FDG showing decreased perfusion and glucose hypometabolism in the cerebellum. (F) Sagittal T1 MRI showing atrophic changes in the midbrain with preserved pons, known as the “hummingbird sign” or “penguin sign”.

PSP-RS shows hypometabolism at the medial and dorsolateral frontal cortex, caudate nucleus, thalamus, and midbrain [21,37] and greater frontal hypometabolism than in PD and MSA, providing good diagnostic performance for diagnosing PSP-RS (sensitivity, 76%; specificity, 98%) [47,48]. Early perfusion of ¹⁸F-FP-CIT showed significant hypoperfusion in the medial frontal cortex compared to in MSA and PD, but not in the midbrain [24].

MDS-PSP criteria adopt only two imaging findings as supportive features: predominant midbrain atrophy or hypometabolism and postsynaptic striatal dopaminergic degeneration [35]. As current imaging techniques have not been compared to the neuropathological gold standard, the MDS-PSP criteria do not consider imaging features as strong evidence of PSP, and additional studies are needed to assess the value of these imaging studies [35,37].

Corticobasal degeneration

CBD is a clinicopathologic entity, but various clinical features can occur in other pathologies, making differential diagnosis quite difficult [49]. The classic presentation with asymmetric rigidity, dystonia, and ideomotor apraxia is now referred to as corticobasal syndrome (CBS), but CBD has also been increasingly recognized to present with features that may overlap with frontotemporal lobe dementia, primary progressive aphasia, Alzheimer disease (AD), and PSP [3].

DAT images revealed reduced tracer uptake in the putamen and caudate nucleus (small rostrocaudal gradient), which mimics PSP; however, marked asymmetricity is a key feature of CBD compared to in PSP and MSA (Figure 7) [26]. Reduced FP-CIT uptake in patients with CBS was quite variable, and there was no significant correlation between tracer uptake values and clinical features such as disease duration and severity [50].

Figure 7.

Corticobasal degeneration (CBD) (A) Maximum intensity projection image of DAT imaging using ¹⁸F-FP-CIT showing prominent asymmetricity of the striatum. (B) Transaxial and (C) coronal DAT images showing decreased tracer uptake in the left caudate nucleus and putamen. (D) Early perfusion image of ¹⁸F-FP-CIT and (E) glucose metabolism of ¹⁸F-FDG showing decreased perfusion and glucose hypometabolism in the left cerebral hemisphere, striatum, and thalamus.

CBD is characterized by asymmetric hypometabolism of the frontoparietal areas, striatum, and thalamus [21]. Cortical hypometabolism may be pronounced in the parietal cortex and typically extends across the sensorimotor cortex into the cingulate gyrus and premotor-to-posterior prefrontal areas [21]. Brain perfusion images can also be used to visualize asymmetric brain perfusion with similar patterns as ¹⁸F-FDG [51,52]. Additionally, as AD, PSP, CBD, and frontotemporal lobe dementia share tau pathology [53], characteristic metabolism and perfusion patterns of the diseases must be considered to ensure an accurate diagnosis.

Combined with pathologically confirmed CBD cases and consensus, 4 CBD phenotypes have been proposed: CBS, frontal behavioral-spatial syndrome, nonfluent/agrammatic variant of primary progressive aphasia, and PSP syndrome [49]. As the pathology of CBD is predicted antemortem in only 25-56% of cases [49], combined analysis of functional images and the subtypes of pathologic confirmed CBD are needed to increase the value of imaging studies.

Dementia with Lewy bodies

DLB is an early-onset rapidly progressive dementia that is part of the spectrum of PD [3]. The fourth consensus report of the DLB consortium considered the loss DAT imaging as an indicative biomarker and metabolism/perfusion change as a supportive biomarker [54].

DAT images of DLB showed decreased tracer uptake in the caudate nucleus and putamen. The rostrocaudal gradient in DLB may be flatter than in PD because of the relatively early involvement of the caudate nucleus [55,56]. DAT imaging shows good diagnostic accuracy (86%) for distinguishing autopsy-confirmed DLB from AD [57]. Normal DAT uptake was reported in 10% of autopsy-confirmed DLB, and these patients showed synucleinopathy in the limbic and neocortical areas but minimal pathology was found in the substantia nigra [57].

Glucose hypometabolism of the occipital cortex is correlated with the visual neuropathy of DLB [54]. Additionally, widespread lateral frontal and temporoparietal glucose hypometabolism was observed [21]. As DLB does not involve the posterior cingulate, which is commonly involved in AD, the posterior cingulate island sign can be observed in DLB patients [58]. Many perfusion imaging analyses were performed to differentiate DLB, revealing a similar hypoperfusion pattern as ¹⁸F-FDG [54].

Future perspectives

Recent advancements have enabled simultaneous acquisition of PET images and magnetic resonance imaging (MRI) data. Conducting ¹⁸F-FP-CIT and MRI together reduces the radiation burden and acquisition time. Recently, comparison studies of ¹⁸F-FP-CIT PET/CT and PET/MR were performed, which showed similar image quality for visual analysis [59,60]. MRI has better image resolution, better anatomical correlation, and more optimal subregional analysis; FP-CIT imaging may be available using PET/MR. A recent study revealed a correlation between the binding ratio of ¹⁸F-FP-CIT and gray matter density using PET/MR [61]. However, ¹⁸F-FP-CIT PET/MR caused spatial bias in quantification, although attenuation maps accounted for cortical bones, and improvements in attenuation correction are needed for data comparison [59].

Recently, deep learning-based artificial intelligence (AI) has been widely applied in the medical imaging field [62]. Most of the characteristic imaging features can be detected by direct observation but applying deep learning-based AI might provide new imaging biomarkers for PD and Parkinson plus syndrome automatically. A recent study using deep-learning based AI effectively classified PD with ¹²³I-FP-CIT SPECT and classified clinical PD in patients whose scans showed no evidence of dopaminergic deficit [63]. However, deep-learning technology requires a large data set that relatively smaller number patients with Parkinson plus syndrome is a limiting factor, and difficulty in determining a definite diagnosis limit the application of this technology for diagnosing the disease. Multicenter pathologic-confirmed imaging data are required to use deep learning technology for disease analysis.

The other application of deep learning-based AI is for image reconstruction is to improve image quality [64]. Because of the limited time for acquisition (less than 10 min) of the perfusion phase using ¹⁸F-FP-CIT, improvements in image quality are difficult to achieve using conventional reconstruction procedures. Recent image generation approaches using deep learning may be helpful for reconstructing perfusion images, which is more practical in clinical practice.

Conclusion

¹⁸F-FP-CIT PET with early perfusion and DAT imaging is useful for the differential diagnosis of PD and Parkinson plus syndrome (MSA, PSP, CBD, and DLB). Early perfusion images can be acquired with less radiation exposure, time, and cost. Further studies are needed to verify the results using autopsy-confirmed Parkinson plus syndromes and their subtypes.

Disclosure of conflict of interest

None.