Efficiency of 123 I-ioflupane SPECT as a marker of basal ganglia damage in acute methanol poisoning: a six-year prospective study CURRENT STATUS: POSTED

Purpose Investigate whether 123 I-ioflupane SPECT (DaT SPECT) has the potential as a marker of basal ganglia damage in acute methanol poisoning. Methods Prospective, single-centre, cohort study of patients with confirmed methanol poisoning was conducted. DaT SPECT was performed twice with semi-quantification using DaTQUANT TM and MRI-based volumetry was calculated. Specific binding ratios (SBR) of striatum, caudate nucleus, and putamen were correlated with laboratory parameters of outcome, volumetric data, and retinal nerve fibres layer (RNFL) thickness measurements. Results Forty-two patients (mean age 46.3±4.2 y; 8 females), including 15 with putamen lesions (Group I) and 27 patients with intact putamen (Group II), underwent DaT SPECT. Volumetry was calculated in 35 of the patients assessed. SBR values for the left putamen correlated with putamen volume (r=0.665; p<0.001). Decreased bilateral SBR values were determined for the striatum and the putamen, but not for the nucleus caudate, in Group I (p<0.05). A strong correlation was observed between the SBR of the posterior putamen and arterial blood pH (r=0.574; p<0.001) and other toxicological parameters of severity of poisoning/outcome including serum lactate, glucose, and creatinine concentrations (p<0.05). The SBR of the posterior putamen positively correlated with the global RNFL thickness (p<0.05). ROC analysis demonstrated a significant discriminatory ability of SBR of the posterior putamen with AUC=0.753 (95%CI 0.604–0.902; p=0.007). The multivariate regression model demonstrated that arterial blood pH, age, and gender were the most significant factors associated with SBR of the posterior putamen. DaT SPECT demonstrates significant potential for the diagnosis of methanol-induced basal ganglia damage. tomography retinal nerve fibres layer (RNFL) thickness measurement. RNFL thickness was measured by SD-OCT Spectralis Tracking Laser software and was compared to a normative database. The following biochemical tests were performed: serum glucose, glycohaemoglobin, albumin, prealbumin, liver and renal function tests, lipids, thyroid-stimulating hormone (TSH), vitamins B 1 and B 12 , carbohydrate-deficient transferrin (CDT), and ethyl glucuronide screening in urine.

Introduction different brain structures. Patients with MRI signs of methanol-induced brain damage possess significantly decreased basal ganglia volumes, and the most significant changes occur in the putamen [32]. Basal ganglia neurons and their axons are susceptible to damage induced by hypoxia and oxidative stress [33,34]. Damaged brain tissue undergoes oedema of affected areas, and later, necrosis, glial scar and postmalatic pseudocyst formation result in decreased brain volume in affected regions [35]. The putamen receives extensive dopaminergic projections from the substantia nigra pars compacta [36]. Altered putaminal volume and shrinkage of the striatum due to decreases in neuronal numbers may affect diverse regions of the brain, since the substantia nigra, thalamus and frontal cortical regions form a topographically organized, cortical-striatal-thalamic-cortical loop [37].
Dopamine transporter (DaT) is a protein that is expressed in the membrane of dopaminergic cells. It facilitates the re-uptake of dopamine into presynaptic terminals. Modern imaging techniques applying 123 I-N-v-fluoropropyl-2b-carbomethoxy 3b-(4-iodophenyl)nortropane ([ 123 I]FP-CIT, or 123 I-ioflupane) for labelling DaT provide a useful tool for evaluating dopaminergic terminal function by single photon emission computed tomography (DaT SPECT), an established method that has widely been applied in both clinical practice and research [38]. Dopamine transporter imaging is key for diagnosing idiopathic parkinsonism (e.g., Parkinson's disease, multisystem atrophy, progressive supranuclear palsy, Lewy body dementia) to distinguish between the occurrence of nigrostriatal degeneration and the preservation of dopaminergic function in clinically uncertain cases of suspected parkinsonian syndrome.
Striatal DaT binding may contribute to axonal dysfunction or DaT expression in the nigrostriatal pathway in patients acutely exposed to methanol. However, no systematic studies have investigated striatal dopaminergic system integrity via functional imaging in survivors of acute methanol poisoning, and only episodic case reports have been published [39,40]. Therefore, it remains unclear whether putaminal necrosis in methanol-poisoned patients decreases striatal DaT binding as a result of axonal denervation. The association between volumes of interest (VOI) measurements determined via MRI and specific binding ratios (SBR) determined via 123I-ioflupane on DaT SPECT is unclear.
Furthermore, researchers have not identified whether quantitative indices of DaT SPECT are associated with poisoning severity, which is determined using acute toxicological laboratory parameters (e.g. base deficit, serum methanol, formate, lactate, glucose, and creatinine concentrations), prognostic factors (e.g. arterial blood pH, serum ethanol concentration, poisoning severity score, PSS), and hospital treatment modalities [30,41,42]. To answer these questions, the relationship between 123 I-ioflupane for the striatum, putamen, nucleus caudate SBRs and MRI-based volumes of these brain structures was investigated. The association between DaT SPECT values and the severity of acute methanol poisoning and clinically applied prognostic factors of long-term outcomes was explored.

Study design and setting
We carried out a prospective, longitudinal, single-centre, observational cohort study of patients with confirmed acute methanol poisoning that were treated in hospitals during a mass methanol poisoning outbreak that occurred in the Czech Republic from September to December of 2012 [11]. The Institutional Ethics Committee approved the prospective study and written informed consent was obtained from all patients prior to study enrolment. Clinical and laboratory tests indicating acute exposure were collected using standardised data collection protocols. Information regarding treatment modalities used and outcomes was collected from discharge reports. Patients who survived methanol poisoning and were discharged from the hospital were examined 3-8 months and then 2, 4 and 6 years after discharge using the same clinical examination and equipment, hardware and image acquisition, reconstruction, and analysis protocol.

Selection of participants and treatment methods
The national monitoring and reporting system of all cases of acute methanol poisoning was established during the methanol poisoning outbreak. All hospitalised patients with analytically confirmed methanol poisoning were eligible for the study. The patients underwent standard neurological examinations and complete ocular examinations while they were hospitalised and also when they were discharged from the hospital. The examination protocol included either a brain CT or MRI. The patients were considered to have CNS sequelae of acute methanol intoxication if MRI or CT brain scans revealed lesions of the basal ganglia.

Treatment was provided according to the European Association of Poisons Centres and Clinical
Toxicologists and the American Academy of Clinical Toxicology guidelines for treatment of methanol poisoning [41]. Alcohol dehydrogenase (ADH) blockers, ethanol and fomepizole, were used as antidotes [43]. Other treatment modalities administered included folate substitution, bicarbonate, enhanced elimination methods, intermittent haemodialysis or continuous veno-venous haemodialysis / haemodiafiltration [44].

Brain imaging protocol and clinical follow-up
During the follow-up, all subjects were scanned using 3T imaging (MAGNE-TOM Skyra; Siemens Healthcare, Erlangen, Germany). The MRI protocol included a T1-weighted 3D MPRAGE sequence for brain volumetry, and the following parameters were applied: repetition time (TR) 2,300 ms, echo time (TE) 2.26 ms, inversion time (TI) 732 ms, slice thickness 1.0/0 mm and flip angle (FA) 8°. Further, sequences were used to describe brain pathology via T2WI, 3D FLAIR, SWI and coronal T2WI with fat saturation was used to visualise optic nerves. All MR data were processed using the MorphoBox prototype software. These data allowed researchers to estimate single brain structure volumes in cm 3 . The estimation was calculated in five steps: (i) a brain template was created each input MR using nine-parameter affine spatial transformation (three translational and three rotational parameters followed by anisotropic scaling); (ii) the input image was corrected for bias field using an expectation-maximisation algorithm; (iii) a total intracranial volume (TIV) template mask was resampled and applied to the transformation obtained in step i for skull stripping; (iv) a template-free tissue classification algorithm was used on the TIV-restricted image, which produced three tissue posterior probability maps (grey matter, white matter and cerebrospinal fluid); and (v) the tissue probability maps were combined with the masks that were resampled in the first step to produce regional volume estimates [45,46].
The follow-up clinical examination protocol, aside from brain MRI and DaT SPECT performed on the same day, included the neurological examination of motor and sensory function, reflexes, cerebellar function, cranial nerves function, and Neuroprotection and Natural History in Parkinson's Plus Syndromes scale, an ophthalmological examination, visual evoked potentials, and optical coherence tomography with retinal nerve fibres layer (RNFL) thickness measurement. RNFL thickness was measured by SD-OCT Spectralis Tracking Laser Tomography (Heidelberg Engineering GmbH, Heidelberg, Germany; software version 5.8.3) and was compared to a normative database. The following biochemical tests were performed: serum glucose, glycohaemoglobin, albumin, prealbumin, liver and renal function tests, lipids, thyroid-stimulating hormone (TSH), vitamins B 1 and B 12 , carbohydrate-deficient transferrin (CDT), and ethyl glucuronide screening in urine. Commercially available DaTQUANT TM (GE Healthcare, Little Chalfont, UK) software was used for automatic semi-quantitative image analyses. After SPECT reconstruction, transaxial slices were used as software input, which applies an automatic VOI-based semi-quantitative evaluation of the image data and compares it to an in-built set of reference data for statistical analysis. Pre-defined template VOIs were automatically positioned in the striatum, putamen, and caudate regions in each hemisphere, and in the occipital reference region. The program calculated semi-quantitative SBR as the difference between the mean counts in each VOI and the mean counts in the background, divided by the mean counts in the background.

Statistical analysis
For all variables, basic descriptive statistics (mean, CI, SD, skewness and kurtosis) were calculated, and the data were tested for normality using Shapiro-Wilk's test. The Pearson's c 2 test was used to compare the frequencies of categorical demographic and clinical variables between the groups of methanol poisoning survivors with and without MRI signs of necrotic lesions of the putamen.
Continuous variables (volumetric data, SBR, and others) were compared using t-test for independent groups. Bivariate associations were assessed using Pearson's correlation coefficient with the twotailed significance test option. Reliability analysis was conducted using the intraclass correlation coefficient (ICC) to index the reliability of DaT SPECT measurements in the first and the second examination. The ICC between the two examinations was calculated using the two-way randomeffects model for absolute agreement between measurements. A multivariate regression model was used to predict the SBR for the striatum, nucleus caudate, putamen, anterior putamen, and posterior putamen posterior based on several predictors including age, gender, chronic alcohol abuse, smoking, and arterial blood pH at admission. Receiver operating curves (ROC) and the area under the curve (AUC) were used to compare discriminatory capacities of the SBR in different VOI for indicating methanol-induced basal ganglia damage. The best cut-off scores within the ROC analysis were identified in terms of sensitivity and specificity. The level of significance was set to α = 0.05. The statistical analyses were performed using Statistical Product and Service Solutions software (IBM SPSS Statistics for Windows, Version 25.0).

Patient characteristics
Of 108 patients hospitalized with acute methanol poisoning, 84 survived. These survivors were asked whether they would be willing to participate in a study that involved a long-term assessment of the effects of methanol poisoning. Of the 54 patients who agreed, in 42 patients (8 females), brain MRI was performed four times during a six-year period and DaT SPECT was performed twice (four and six years after discharge). The patients had not previously been diagnosed with neurological, neurodegenerative, neurovascular, or psychiatric disorders other than chronic alcohol abuse. In 15/42 patients (36%), brain MRI revealed signs of necrotic lesions of the putamen (patients classified as Group I) and in 27 patients, no signs of putaminal damage were detected (classified as Group II).  Table 1. The two groups did not differ with respect to age, gender, proportion of chronic alcohol abusers or proportion of smokers when admitted to the hospital with acute methanol poisoning. The patients with MRI-detected necrotic lesions of the putamen tended to be more severely poisoned and have lower arterial blood pH and bicarbonate concentrations. They also tended to have increased base deficit, anion gap, and serum lactate levels, which are characteristics of metabolic acidosis. All patients admitted in a coma belonged to Group I and 80% of the patients with PSS 3 poisoning (severe poisoning) were the group (p < 0.001). No difference in treatment modalities (administration of ethanol versus fomepizole for ADH blocking, folate substitution) and in the follow-up laboratory parameters was observed between the two groups.
Patients in Group I had lower left putamen, nucleus caudate, and globus pallidus volumes than those of Group II, but did not display different right VOI, which revealed a degree of asymmetry associated with MRI-volumetric findings (Table 1).

Correlation between mean specific binding ratios on DaT SPECT, clinical and laboratory parameters of poisoning severity and outcomes
There was no difference observed between the two groups with regard to the relative variation of background signals, ruling out significant differences in image quality. An assessment of reproducibility between the first and the second examination revealed strong agreement (intraclass correlation coefficients for the SBR for all VOI ranged from 0.853 to 0.933, all p < 0.001). Therefore, mean quantitative indices as the averages of the first and the second measurements were further applied. DaT SPECT results produced significantly lower mean SBR for the whole striatum and for the bilateral putamen, but not for the nucleus caudate, in the patients with necrotic lesions of the putamen detected on brain MRI ( Table 2). The greatest difference in the SBR was observed for the bilateral posterior putamen. Certain asymmetry in the SBR for the right and left hemispheres was observed, with lower indices determined for the left putamen, but the difference was not significant (p > 0.05). The SBR for the left putamen was strongly positively correlated with its volume (Fig. 2). Correlations between SBR and the volume of the right putamen were significant only for the posterior putamen (r = 0.386, p = 0.022). In contrast, the volume of the nucleus caudate was not correlated with its SBR, and the SBR for the striatum and the putamen (p > 0.05). Finally, the volume of left, but not right, globus pallidus positively correlated with the SBR for the whole putamen (r = 0.344, p = 0.043), and for the posterior putamen (r = 0.378, p = 0.025). No correlation was found between the SBR for the whole striatum and the volume of nucleus caudate or globus pallidus (p > 0.05).
The SBR for the striatum, putamen and nucleus caudate in both hemispheres did not correlate with age, gender, and duration between methanol exposure and hospital admission (p > 0.05). There was a significant positive correlation between chronic alcohol abuse and the SBR for the right, but not left, A multivariate regression model demonstrated that arterial blood pH, age at admission to the hospital, and gender were significant association factors for the SBR for the bilateral posterior putamen in survivors of acute methanol poisoning ( Table 3). The association of smoking with SBR demonstrated borderline significance for the right, but not left, posterior putamen. The same model was relevant for the whole putamen, but the degree to which arterial blood pH and the age associated with SBR decreased since values of the SBR for the anterior putamen associated to a lesser degree than those for the posterior putamen to laboratory parameters of the severity of methanol poisoning. Notes: CI -confidence interval; Arterial pH -arterial blood pH at admission to the hospital with acute methanol poisoning; Age -age at admission to the hospital; Chronic alcohol -chronic alcohol abuse.
The level of significance is p < 0.05 (bold numbers).

Discussion
Our study demonstrated that DaT imaging with 123I-ioflupane SPECT is extremely accurate and is capable of distinguishing between the patients with and without methanol-induced basal ganglia damage. These results revealed that the SBR for the putamen was significantly lower in patients with MRI signs of necrotic putaminal lesions and was correlated with volumetric data. DaT SPECT reflects dopaminergic axonal dysfunction in the striatum, whereas MRI-volumetry directly registers the size of relevant brain structures. The SBR for the putamen as the region of interest showed better classification performance compared to the SBR for the whole striatum. Furthermore, the SBR for the putamen was strongly positively correlated with arterial blood pH at admission to the hospital, acute laboratory parameter reflecting the severity of methanol poisoning and main prognostic indicator of hospital outcome [9,11,41,42]. Of all studied VOI, the SBR for the posterior putamen demonstrated highest sensitivity and specificity for detection of methanol-induced basal ganglia damage.
Importantly, DaT expression in the posterior putamen was positively associated with RNFL thickness, which is a basic morphological feature of retinal neurodegeneration after acute methanol poisoning [30].
An association exists between striatal 123 I-ioflupane uptake and both the number of dopaminergic neurons in the substantia nigra pars compacta and the functional state of those neuron's axon terminals [47,48]. Outputs of the substantia nigra pars compacta are directed to the spines of dendrites of GABAergic medium spiny neurons in the striatum. Therefore, DaT SPECT, by evaluating the function of dopaminergic terminal axons, provides indirect information on the numbers of surviving GABAergic neurons in the putamen, especially within its posterior part, the caudal putamen, containing motor circuits where the reductions of dopamine transporters are most severe. Our data are in accordance with results of previous studies that revealed the selective vulnerability of the posterior putamen to hypoxic-ischemic brain injury, which researchers determined to be due to vascular and biochemical factors [49]. Further, we observed certain asymmetry in dopaminergic terminal function decrease with relatively lower indices for the SBR for the left striatum. It is interesting that Parkinsonian disease pathology occurs asymmetrically as well, and clinical motor symptoms manifest unilaterally in the early stages of the disease [50].
Common risk factors of poor prognosis in acute methanol poisoning, such as severity of metabolic acidosis with high anion gap, base deficit, serum lactate, and low arterial blood pH [41,42], were is typically inferior to that of MRI, and partial volume effects or smoothing may prevent SPECT from detecting small focal lesions in the brain of the patients with mild poisoning [51].
Age dependency has been reported for both males and females for DaT SPECT measurements [52]. In the present study, no correlation of the SBR with age and gender was observed. This lack of association can be explained by the relatively homogenous young age of the study population (IQR 35-58 years), small group size, and the capacity of methanol-induced basal ganglia damage to mask age-related DaT decline. Nevertheless, in the multivariate regression model, both age and gender significantly associated with DaT availability in the putamen. This model demonstrated that arterial blood pH at admission most strongly affected DaT SPECT measurements ( Table 3). The toxic effect of formic acid, an inhibitor of mitochondrial respiration, depends on the degree of the metabolic acidosis produced: in a highly acidotic environment, formic acid will be more toxic than in a less acidotic one.
The dissociation constant of formic acid (pKa) is 3.8, therefore, a pH-drop by 0.3 would double the undissociated formic acid levels and produce significant increases in toxicity. This occurs because only undissociated formic acid crosses the blood-brain barrier and reaches neurons of the basal ganglia [53].
In the current study, no significant differences were observed in the SBR for the nucleus caudate between the patients with and without MRI-signs of basal ganglia damage. Differences in the SBR for the anterior putamen between two groups were smaller than for the posterior putamen. These data It has been previously reported that higher putaminal volume was associated with longer lifetime cigarette use and younger age of smoking initiation [54].
Abnormal thickness of RNFL on optical coherence tomography as a result of the toxic effect of formic acid on retinal ganglion cells, followed by chronic retinal neurodegeneration, has been previously reported in survivors of acute methanol poisoning [30,55]. DaT SPECT measurements revealed a positive correlation between RNFL thickness and the SBR of the posterior putamen. This finding indicates that OCT with RNFL measurement can be used as a reliable, cheap and simple screening method, indicating the group of methanol-poisoning survivors with possible basal ganglia damage. In this group, DaT SPECT with 123 I-ioflupane can be used for verifying the diagnosis.
The current study has several limitations. First, number of study participants who underwent MRIvolumetry was smaller than the number of patients who underwent DaT SPECT. This may have affected correlations between the SBR and volumetry data. Nevertheless, his is the first study that prospectively assessed a homogenous cohort of survivors of acute methanol poisoning from single clinical setting. The patients of the study were subjected to the same study protocol and equipment for six years, and were recruited after a single, mass methanol poisoning event. Brain MRI was performed on all patients to detect the signs of methanol-induced damage. Secondly, patients with MRI data indicative of necrotic lesions of the putamen tended to be younger than those without signs of putaminal damage, differences were not significant. Third, DaT SPECT was not performed on healthy controls for to ethical reasons. The diagnosis of acute methanol poisoning was confirmed via toxicological laboratory analysis in all patients. A reliable diagnosis of long-term CNS sequelae of acute methanol poisoning requires a relatively long follow-up. The strength of the study consisted in its prospective longitudinal character. MRI of the brain was performed, and patients were examined consecutively four times within a six-year period post-discharge from the hospital. For each patient, DaT SPECT was performed twice in the same medical facility using the same camera, hardware, and acquisition/reconstruction protocol to minimize variability.

Conclusion
DaT SPECT with 123 I-ioflupane demonstrates high potential for the assessment of dopaminergic function in the striatum, especially within its posterior region, as a valuable diagnostic tool for evaluating long-term CNS sequelae in survivors of acute methanol poisoning. Importantly, DaT SPECT demonstrated high diagnostic correspondence with MRI-volumetry, and the SBR decrease in the striatum correlated with acute laboratory parameters of poisoning severity and prognostic indicators of hospital outcome, which indicated that DaT SPECT has the potential to more accurately identify patients with basal ganglia damage than brain MRI-volumetry. Therefore, DaT SPECT with 123 I-

Ethical approval and consent to participate
The Ethics Committee of the General University Hospital in Prague approved this prospective study, and written informed consent was obtained from all patients before examinations. The approval No. is 31/15.

Consent for publication
Written informed consent was obtained from all patients before examinations.

Availability of supporting data
All data generated or analysed during this study are included in this published article.