The Feasibility Of Co-Registered Solid-State 123i-Mibg Spect And Ct Imaging For The Assessment Of Ventricular And Atrial Sympathetic Innervation In Healthy Individuals

Background. To evaluate global and regional ventricular and atrial cardiac iodine-123 metaiodobenzylguanidine ( 123 I-mIBG) uptake and consistency over time in healthy individuals using co-registered single-photon emission computed tomography (SPECT) and computed tomography (CT) imaging. Fifteen healthy individuals (median age 31 years [26; 41]) were included in the study. All participants underwent CT and subsequent baseline 123 I -mIBG SPECT imaging (early and late acquisition) using a dedicated cardiac solid-state gamma camera. The heart-to-mediastinum (H/M) ratio, wash out rate (WR), summed 123 I-mIBG defect score (SDS) as well as presence and patterns of left atrium (LA) discrete 123 I-mIBG uptake areas (DUA) were assessed. Follow-up SPECT imaging was acquired 5– 7 days after initial procedure.

. Areas of reduced 123 I-mIBG uptake were detected in 60% (9/15) of cases and the median SDS was 1 [0 ; 2]. No signi cant changes were observed in global and regional 123 I-mIBG cardiac uptake between baseline and follow-up studies. At baseline 36 DUA were identi ed, 16 (44%) of which (median per individual 1 [1;1]) had moderate-high con dence score (CS). Conclusion. Co-registered solid-state 123 I-mIBG SPECT and CT is feasible for the evaluation of left ventricle sympathetic innervation as well as for the detection of discrete sympathetic areas in the typical anatomical sites of left atrial ganglionated plexi with modest variations in the constancy of uptake locations. There were no signi cant changes in the global and regional 123 I-mIBG left ventricular uptake over a short time interval in healthy individuals.

Background
The autonomic nervous system (ANS) plays an important role in regulating the function of cardiomyocytes, the conducting system, coronary vessels, contractile function of the myocardium, and also affects the electrophysiological properties of the myocardium [1]. The heart is innervated by both the extrinsic (central) and the intrinsic cardiac autonomic nervous system (CANS) [2]. The CANS, in particular presented as ganglionated plexi (GP) which contain autonomic ganglia and nerves anatomically located around the left atrium (LA), plays an important role in the initiation and maintenance of atrial brillation (AF). However, the effectiveness of GP ablation in AF patients remains disputed and one of the major reasons could be the lack of an e cient method to localize the GPs. The adrenergic part of CANS can be noninvasively assessed with Iodine-123 metaiodobenzylguanidine ( 123 I-mIBG), which is a sympathetic neurotransmitter radionuclide analog and aids in the detection of sympathetic innervation abnormalities [3]. 123 I-mIBG scintigraphy is widely used for the qualitative and quantitative global and regional assessment of cardiac sympathetic innervation and shows cardiac denervation in a variety of pathologies [4].
In recent years, nuclear cardiac imaging has advanced technologically, with the introduction of novel high-sensitivity, rapid-acquisition solid-state dedicated cardiac single-photon emission computed tomography (SPECT) cameras. The use of solid-state semiconductor cadmium zinc telluride-based (CZT) detectors, coupled with their ability to fan across the speci c region of interest, has demonstrated signi cantly improved sensitivity, spatial resolution and energy resolution, enabling cardiac SPECT imaging with a spatial resolution of < 5 mm [5,6]. Several studies have assessed the comparative performance of Anger-type and CZT SPECT systems for detection of left ventricular myocardial tracers [7,8,9], but there are few studies evaluating left atrial 123 I-mIBG uptake using CZT systems, particularly in healthy individuals. In patients with atrial brillation, 123 I-mIBG CZT SPECT co-registered with cardiac computed tomography (CT) has been reported to non-invasively identify discrete 123 I-mIBG uptake areas (DUAs) that correlate with left atrial ganglionated plexi (LAGP) identi ed invasively by high-frequency stimulation (HFS) during pulmonary vein isolation (PVI) catheter ablation [10]. However, imaging patterns of LAGP in healthy individuals have never been demonstrated. The aim of this study was to evaluate global and regional ventricular and atrial cardiac 123 I-mIBG uptake and their consistency over time in healthy individuals using co-registered SPECT and CT imaging.

Study protocol
The study was initiated to test the feasibility of identifying left atrium (LA) DUA in healthy individuals using an imaging technique that enables indirect measurement of cardiac sympathetic innervation, and additionally to evaluate the constancy of CANS imaging patterns.
All study participants underwent an initial assessment including medical history, cardiac CT imaging and 123 I-mIBG SPECT imaging. Follow-up 123 I-mIBG SPECT imaging was acquired 5-7 days after initial procedure.

Patient Selection
Fifteen healthy individuals who were invited to participate due to their interest in science and medicine gave their voluntary consent to participate in the study. Inclusion criteria were healthy men or women aged 18-75 years without history or symptoms of heart disease. Exclusion criteria included any history of heart disease, hypertension, diabetes, chronic kidney disease, neuropathy and any contraindication to 123 I-mIBG or iodinated contrast media.

Page 4/19
Cardiac computed tomography All cardiac CT (Aquilion ONE, Toshiba Medical Systems Corp., Japan) scans were performed 1 day prior to 123 I-mIBG SPECT imaging. For acquisition of the volume dataset, a biphasic intravenous iodinated contrast protocol was used, with 60 mL contrast at 5 mL/s followed by 30 mL 30%/70% mixture of contrast and normal saline respectively at the same rate. The scan was acquired with prospective ECG triggering in the mid-diastolic phase and in held expiration. Images were imported into dedicated software (Shina Systems Limited, Caesarea, Israel) for cardiac chamber segmentation. After manual corrections to the segmentation, a representative 3D surface mesh le was created for each chamber, which was then used for co-registration with SPECT images. The mean radiation exposure for the CT procedure was 3.8 +/-2.1 (3.1-6.5 95% CI) mSv.

I-mIBG SPECT image acquisition
Imaging procedures were conducted in accordance with established guidelines [11,15]. All medications known to affect the uptake of 123 I-mIBG were ceased and oral thyroid blockade with potassium iodate 170 mg was provided for all patients 1 day before the planned 123 I-mIBG administration and continued for 1-2 days afterwards. Medications were re-started following the SPECT procedure. Data were acquired 15 minutes (early acquisition) and four hours (late acquisition) after intravenous injection of 313 +/-37 (299-327 95% CI) MBq 123 I-mIBG, on a dedicated cardiac solid-state SPECT camera (D-SPECT, Spectrum Dynamics Medical, Caesarea, Israel) with cardiac-and respiratory-gating. Study acquisition time was 20 minutes. The mean radiation exposure for the SPECT procedure was 4.11 +/-0.5 (3.9-4.3 95% CI) mSv [12]. 123 I-mIBG SPECT and CT image processing and co-registration To process the corresponding SPECT and CT images, 123 I-mIBG late acquisition datasets were reconstructed with a high-resolution reconstruction algorithm in the diastolic phase of the cardiac cycle and in the expiratory phase of the respiratory cycle.
The co-registered images were generated using a dedicated workstation (SUMO D-SPECT, Spectrum Dynamics Medical, Caesarea, Israel). SPECT images were automatically co-registered to the CT images using left ventricular (LV) myocardial uptake as the SPECT reference and LV myocardial segmentation as the CT reference. The SPECT tomograms could be translated in the (x, y, z) direction in order to achieve accurate location matching with the CT-derived myocardial segmentation using identi able anatomic reference points (ie. left ventricular apex and mitral annulus). Co-registration accuracy was reviewed by scrolling through horizontal long, vertical long and short axis views and, once satisfactory alignment was con rmed, nal 3D images of cardiac innervation combined with anatomical information were generated. Data analysis Global and regional sympathetic cardiac innervation The planar equivalent image (planogram) was used to derive heart-to-mediastinum (H/M) ratio for the early and late acquisitions, which in turn were used to calculate cardiac wash out rate (WR) [13]. The H/M ratio was determined from the average pixel counts/number of pixels in a visually drawn heart region of interest (ROI) divided by the average pixel counts/number of pixels in a visually drawn mediastinum ROI in the midline upper chest positioned to re ect the region with lowest background activity. The WR was calculated as described elsewhere [15].
For SPECT image analysis, regional myocardial 123 I-mIBG uptake was classi ed using a 17-segment model of the left ventricle and a semiquantitative ve-point scale (0 = normal uptake, 1 = mildly reduced uptake, 2 = moderately reduced uptake, 3 = severely reduced uptake and 4 = no uptake). Readers were instructed that scoring could be adjusted if artefact was thought to be present. A summed 123 I-mIBG defect score (SDS) was calculated as the sum of all segmental defect scores in the late acquisition; patients with no defect were scored zero. Scoring was performed independently by two experienced nuclear physicians and consensus was reached in case of divergence.
Co-registered 123 I-mIBG SPECT/CT data analysis Focal areas of increased 123 I-mIBG activity around the left atrium were reviewed and accepted or rejected by the reader as discrete uptake areas (DUAs). Each DUA was assigned a con dence score (CS) based on discreteness, extent of overlap with non-LA activity (e.g. LV myocardium or lung) and proximity to known anatomical GP cluster locations. DUAs that met all three criteria -discrete, distinct from adjacent extracardiac activity and located in an area known to be typical for GPs -were scored with high CS. DUAs meeting only 2 or 1 of the criteria were scored with moderate and low CS respectively. Areas of uptake adjacent to areas of known high 123 I-mIBG activity, such as the basal lateral wall of the LV, lung or anterior to the esophagus were scored with lower con dence or excluded from consideration. Images were reviewed for the identi cation of DUA and a nal CS assigned by a consensus of blinded analyses by 3 trained physicians (1 radiologist and 2 electro-physiologists).

Statistics
Descriptive statistics are presented as median [interquartile range] for continuous variables and are frequencies and percentages of patients for categorical variables. Continuous variables in paired samples were compared using Wilcoxon signed rank test. Bland-Altman plots were generated, along with estimation of associated statistics, to describe the agreement in detected DUA numbers between baseline and follow-up studies. To assess proportional bias (i.e. increasing difference between the measures with increasing the average of the measures), linear regression with differences as dependent variable and means as explanatory variable was used and statistical signi cance of the slope was tested using Wald test. All analyses were done using R (R Core Team (2019). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/). All p values were based on a two-sided test and a p value < 0.05 was considered statistically signi cant.

Results
All 15 healthy participants in the study completed both baseline and follow-up imaging protocols. The clinical and demographic data are detailed in Table 1. Global and regional sympathetic cardiac innervation No signi cant changes were observed in global and regional 123 I-mIBG cardiac uptake between baseline and follow-up studies (   Table 3 Co-registered 123 I-mIBG SPECT/CT data analysis. n -numbers; Me -median; IQR -interquartile range; DUA -discrete uptake area; CS -con dence score;  There was no conclusive evidence of proportional bias (p = 0.08) (Fig. 3a). Similarly, detection of moderate-high CS DUA numbers showed modest agreement, with no statistically signi cant proportional bias (p = 0.07) (Fig. 3b). Moderate-high CS DUAs had rather widest anatomical distribution -2/16 DUAs were located in interatrial septum, left superior, right anterior clusters; 5/16 DUAs were located at the junction of SVC and RA (Fig. 4).
Ten of 16 (63%) DUAs identi ed at baseline with moderate-high CS were seen in the same location at follow-up. This was only true for 2/20 (10%) DUAs identi ed with low con dence at baseline. 6/16 (37%) moderate-high CS DUAs identi ed at baseline were not reproduced on the follow-up study.

Discussion
In this study, we demonstrated for the rst time the feasibility for identi cation of left atrium 123 I-mIBG discrete uptake areas (DUA) in healthy individuals using co-registered images from 123 I-mIBG solid-state SPECT and a corresponding CT. Also, we evaluated imaging patterns of atrial and ventricular cardiac sympathetic innervation and their variation over time in a small cohort of healthy individuals.
Originally developed as a radionuclide for imaging adrenal tumors, 123 I-mIBG has emerged as a promising tool for the prediction of heart failure progression [16], arrhythmic events [17], and even prognosis of AF recurrence after interventional treatment [18]. More recently, imaging with dedicated cardiac SPECT cameras using solid-state CZT detectors has been shown to identify discrete areas of sympathetic activity that correlate with GPs identi ed invasively using high frequency stimulation (HFS) [10].

Global and regional sympathetic cardiac innervation
Late H/M ratio determination using the planar images from Anger SPECT cameras (A-SPECT) is a wellstandardized and reproducible parameter in assessment of global sympathetic cardiac innervation and its prognostic value is widely recognized. However, previous studies using A-SPECT cameras have shown that the acquisition protocol used signi cantly affects quanti cation of late 123 I-mIBG H/M ratio. In particular the collimators used and the stopping power of the detector material may have an impact on the quanti cation of the 123 I-mIBG H/M ratio [19].
Further, it has been noted that there are a growing number of nuclear medicine sites that are using the new generation of cardiac-centered CZT-based gamma cameras for routine clinical practice [20]. This suggests the importance of developing further clinical evidence regarding the assessment of 123 I-mIBG cardiac uptake features using such dedicated cardiac SPECT cameras. It has been demonstrated that these cameras have a better count detection sensitivity and improved energy resolution, enabling reductions in acquisition times and injected radiopharmaceutical doses [5,6]. However, there are only a small number of studies evaluating cardiac sympathetic innervation imaging with these new generation detectors [21,22,23] and only a few studies have compared H/M ratio determined using CZT acquisition versus that determined using A-SPECT [13,14]. Bellevre et al. demonstrated that determination of the late H/M ratio of 123 I-MIBG uptake using a parallel-collimator CZT camera (D-SPECT, Spectrum Dynamics Medical, Caesarea, Israel) was feasible, reliable and equivalent to H/M ratio using A-SPECT in patients with heart failure [13]. It has also been noted that H/M ratio values from D-SPECT were signi cantly higher than those from A-SPECT, however after applying a correction factor, there was no signi cant difference between A-SPECT H/M ratio and corrected D-SPECT H/M ratio. In contrast, Blaire et al. using a multi-pinhole CZT camera (Discovery NM 530c, GE Healthcare) found almost perfect concordance between H/M ratio derived from transaxial SPECT images and H/M ratio from planar ones however concordance between H/M ratio from re-projected images and planar ones was only moderate. Blaire et al. considered that was related to the multi-pinhole collimation, which is responsible for a truncation artifact that interferes with the mean counts of the myocardial ROI [14].
In our study we used the methodology of H/M ratio calculation proposed Bellevre et al. [13]. Thus, quantitative characteristics of global sympathetic activity (H/M ratio, WR) from our study are comparable to those in a recent study of healthy adults [24] that used planar images for calculation and are lower than reported in ADMIRE-HF study controls participants that used SPECT-derived data [25]. This fact may support the use of planogram images and a correction factor as the methodology for H/M ratio determination for data from the cardiac dedicated parallel-collimator CZT camera (D-SPECT, Spectrum Dynamics Medical, Caesarea, Israel) used in this study.
A recent study of healthy adults (n = 15) with mean age 54.6 ± 5.4 reported heterogeneous patterns of regional 123 I-mIBG uptake predominantly affecting the LV apex, base and inferior wall [24]. This is in concordance with earlier observational studies that reported the effects of age and sex on myocardial 123 -mIBG uptake [26,27]. The recent ADMIRE-HF study, however, failed to show an in uence of age on 123 I-mIBG uptake in older healthy controls [25]. In our study, we also observed a slight decreasing of 123 I-mIBG regional myocardial uptake with increased age, comprised mostly of mild uptake reduction in LV apex and LV inferior wall segments. However, the low median age of our participants (31 years [26;41]) makes the abovementioned age-related denervation an unlikely contributor to our ndings. Our study subjects also did not have hypertension, diabetes mellitus or any other disease known to affect the autonomic nervous system. Decreased regional myocardial 123 I-mIBG uptake has previously been reported in athletes, in the setting of both normal heart rate and sinus bradycardia [28]. In this cohort, inferior, apical and septal defects were all demonstrated, with a signi cant reduction in percentage regional 123 I-mIBG uptake in the inferior region for athletes with sinus bradycardia. Additionally, a normal SPECT database accumulated by the Japanese Society of Nuclear Medicine (JSNM) working group evaluating normal values and standardization of parameters in nuclear cardiology found patterns of lower count distribution of 123 I-mIBG, particularly in the inferior region [32], as we have shown in the cohort of healthy individuals studied here. Hence, the slight decreases in 123 I-mIBG regional myocardial uptake in our study subjects can be interpreted as a heterogeneous pattern of regional myocardial 123 I-mIBG uptake and indicate normal physiological variation.
Co-registered 123 I-mIBG SPECT/CT data According to the results of a recent study, pre-operative non-invasive imaging of 123 I-mIBG DUAs may be helpful in targeting locations for LAGP ablation as a part of interventional AF treatment. This may potentially increase the e cacy of the ablation procedure without touching viable myocardium during LAGP detection using HFS [29]. Hence, knowledge of the distribution and characteristics of DUAs in humans without cardiovascular disease may be helpful for future studies of image-guided AF ablation procedures.
Our study of healthy individuals demonstrates the presence of discrete uptake of 123 I-mIBG, a radiolabeled analog of sympathetic neuronal transmitter norepinephrine, around the left atrium. This fact allows us to assume the presence of sympathetic activity areas in sites that correspond to known anatomical clusters of LAGP. Several studies have shown that LAGPs contain both sympathetic and parasympathetic elements, as well as a variety of neuropeptides and neuromodulators [30,31]. Hence, physiological 123 I-mIBG uptake in typical anatomical sites serves as a marker for LAGP location. For ethical reasons, we did not proceed to con rm the DUAs identi ed in our normal patient cohort invasively with HFS, but this has been done in patients with paroxysmal AF [10].
In comparison with a recent study of AF patients (n = 21) [10] we found a much smaller number of DUAs in healthy individuals (4 [3-4.5] vs 1 [1;1]). However, it should be noted that discrete 123 I-mIBG uptake could indicate only the presence of local sympathetic neurotransmission functional activity, but not the presence of anatomical (or material) structures. The relationship between left atrial sympathetic activity and atrial brillation requires further investigation but our data could suggest that healthy individuals seem to have fewer functionally active GPs than has been reported for patients with AF.
Our ndings demonstrated poor agreement for interstudy DUA number variation, that can be caused by small number of cases and probably could re ect the contribution of artifacts and adjustment areas of known high 123 I-mIBG activity to DUA identi cation with low CS. However, moderate-high CS DUA numbers demonstrated modest agreement between studies, indicating the value of additional criteria for DUA identi cation methodology. That is shown by the fact that DUAs corresponding to the described anatomical LAGP sites are more frequently reproduced in follow-up studies (10/16 (63%)) than DUAs located in other LA areas, corroborating the idea that these sites re ect true LAGPs. It seems that in general, modest variations in reproducibility of DUAs imaging, hypothetically could be related to functional changes in CANS activity. However, in order to assess the DUA imaging reproducibility and extrapolate the methodology to the routine clinical practice larger cohort of study subjects is required.

Limitations Of The Study
The primary limitation of the present study is the small number of subjects included. In addition, we did not apply any cardiac testing to con rm the healthy status of study participants. Also, the variability in the positions of the 123 I-mIBG discrete uptake areas between the baseline and follow-up studies may, at least in part, be camera related or due to subject positioning or even the appearance of true new uptake areas. A larger study is required to con rm our ndings. Also, there is need for further research to correlate imaging ndings with morphological studies of presumed LAGP locations.
New Knowledge Gained Global and regional 123 I-mIBG cardiac uptake, taken as an indirect marker of cardiac autonomic nervous system functional activity, remained constant over a short time interval in healthy individuals. Discrete uptake areas of 123 I-mIBG in the typical anatomical sites of left atrial ganglionated plexi could be visualized in healthy individuals with modest variations in the constancy of uptake locations. Thus, the further investigation on imaging patterns and their diagnostic signi cance is needed.

Conclusion
Co-registered solid-state 123 I-mIBG SPECT and CT is feasible for the evaluation of left ventricle sympathetic innervation as well as for the detection of discrete sympathetic areas in the typical anatomical sites of left atrial ganglionated plexi with modest variations in the constancy of uptake locations. There were no signi cant changes in the global and regional 123 I-mIBG left ventricular uptake over a short time interval in healthy individuals. Consent for publication of results was obtained through the informed consent form which was signed by each subject before any trial-related procedures were performed.
Availability of data and material.
The datasets generated during the current study are available from the corresponding author on reasonable request.