Diagnostic Value of Choline PET in the Preoperative Localization of Hyperfunctioning Parathyroid Gland(s): A Comprehensive Overview

Hyperparathyroidism is a metabolic disorder characterized by the excessive production of the parathyroid hormone. The diagnosis is based on clinical and laboratory data. In most cases the only treatment is surgery and a correct preoperatory localization of the hyperfunctioning parathyroid gland(s) is essential. Currently, ultrasonography combined with [99mTc]Tc-MIBI parathyroid scintigraphy, optionally associated with single photon emission computed tomography/computed tomography (SPECT/CT), represent the standard preoperative imaging. In recent years, a number of studies have evaluated the potential role of choline positron emission tomography (PET) in hyperparathyroidism with promising results. Most of the recent evidence underlined its higher sensitivity and diagnostic accuracy in the localization of hyperfunctioning parathyroid glands. Choline PET has a higher spatial resolution that is useful for the detection of smaller parathyroid glands and it also has shorter examination times and favorable radiation exposure. These are just a few of the aspects that support it to overcome traditional imaging. Moreover, from the preliminary data, the choline uptake mechanism seems to also have an impact on its better performance. For these reasons, if first used as second level imaging in patients with negative or inconclusive traditional imaging results, several authors have supported its use as a first line investigation. This comprehensive overview aims to provide an accurate description of the preliminary results available in the literature about the use of choline PET/CT in hyperparathyroidism and to compare these results with the performance of traditional imaging methods.


Introduction
Hyperparathyroidism (HPT) is the third most common endocrine disorder and is characterized by elevated or inappropriately normal parathyroid hormone (PTH) secretion, which regulates calcium and phosphorus metabolism [1].
This disorder is classified into three subgroups according to the pathogenesis. Primary HPT is the most common form, caused by a solitary parathyroid adenoma (80%), followed by four gland hyperplasia (10-15%), multiple adenomas (5%) and parathyroid cancer (<1%) [2]. HPT can also be due to an alteration of calcium homeostasis, frequently associated with a chronic disease that induces the overproduction of PTH. This condition, known as secondary HPT, can evolve into a tertiary form where one or more parathyroid glands begin to produce PTH independently.
In most cases, surgery is the only curative treatment and the correct presurgical detection of the hyperfunctioning parathyroid gland(s) represents a crucial step for patient management.
However, both of these methods have a few limitations. Low PTH serum levels and ionized calcium, a small adenoma, multiglandular disease and/or concomitant thyroid disease represent some factors associated with a lower scintigraphy performance [8]. Similarly, neck US is highly operator dependent and, in a few cases, is not able to differentiate parathyroid adenomas from reactive cervical lymph nodes as well as to detect ectopic parathyroid glands located in the upper mediastinum [5,7].
Recently, based on occasional findings, the role of positron emission tomography/CT (PET/CT) using choline-labeled radiopharmaceuticals has been highlighted [9].
The rationale of the use of choline-labeled radiopharmaceuticals in this setting of patients could be explained by the different mechanism of the uptake of parathyroid cells compared with [ 99m Tc]Tc-MIBI.
Of note, the parathyroid gland has two major types of cells, chief and oxyphilic cells. Chief cells are small with a big nucleus and poor cytoplasm and are responsible for PTH secretion in which also choline-kinase, a phospholipid/ Ca 2± dependent enzyme, seems to play a role. Oxyphilic cells are characterized by an abundant typically acidophilic cytoplasm due to large amounts of mitochondria [10].
While it is well established that [ 99m Tc]Tc-MIBI, a lipophilic cation, concentrates in the mitochondria of parathyroid cells through the transmembrane electrical potential, the exact biological process underlying the choline uptake in HPT is not yet clearly understood.
Similar to [ 99m Tc]Tc-MIBI, thanks to its positive electric charge, choline-labeled radiopharmaceuticals enter through a membrane transporter and accumulate in the mitochondria of both oxyphilic and chief cells. Furthermore, in the chief cells, choline is also phosphorylated by choline-kinase, which is overexpressed in patients with HPT, and used as a component of cell membranes [11][12][13][14].
Therefore, if [ 99m Tc]Tc-MIBI metabolism depends only on the mitochondria uptake, choline-labeled radiopharmaceuticals can exploit a double uptake mechanism that could represent the advantage of choline-PET to [ 99m Tc]Tc-MIBI scintigraphy [15].
The aim of this review is to provide an accurate description of the preliminary results available in the literature about the use of choline PET/CT in HPT and to compare these results with the performance of traditional imaging methods.
The different mechanism uptake of both radiotracers in hyperfunctioning parathyroid cells is represented in Figure 1. The aim of this review is to provide an accurate description of the preliminary results available in the literature about the use of choline PET/CT in HPT and to compare these results with the performance of traditional imaging methods.

Materials and Methods
A bibliographic literature search until 30

Materials and Methods
A bibliographic literature search until 30 November 2020 was conducted on three electronic databases (Pubmed, Scopus, Google Scholar). Based on a PICO (Problem/Population, Intervention, Comparison, Outcome) strategy search the following keywords were entered: (Hyperparathyroidism AND hyperfunctioning parathyroid glands AND/OR adenoma) and ([ 18 F]FCH PET/CT OR [ 18 F]FCH PET OR [ 11 C]Choline PET/CT AND choline positron emission tomography) and ([ 99m Tc]Tc-MIBI AND parathyroid scintigraphy) and (diagnostic performance AND detection rate AND sensitivity AND specificity). Only original articles edited in English and about humans only that analyzed the role of [ 11 C/ 18 F]cholinelabeled radiopharmaceuticals in the localization of hyperfunctioning parathyroid gland(s) were finally included in this review. Articles with a sample size < 10 patients, review, meta-analysis, case report or case series were not included in the discussion. To identify supplementary eligible articles, additional references were searched from the retrieved review articles.

Literature Results
The literature search revealed 62 articles. Thirty of them were excluded after reviewing titles, abstracts and full texts. Therefore, 32 studies were considered suitable for the analysis: 14 studies were retrospective whereas 18 studies were prospective. In 28 studies choline was labeled with Fluorine-18 [ 18 F] while only four studies evaluated HPT with Carbon-11[ 11 C] choline. Most authors (84%) examined the role of choline in primary HPT, in particular in the detection of parathyroid adenoma. Only 16% of authors included patients with secondary and/or tertiary HPT. In 25 studies choline PET/CT was compared with [ 99m Tc]Tc-MIBI scintigraphy. In the articles where sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and overall accuracy were calculated, postsurgical histological examinations and/or the biochemical resolution of HPT after surgery were considered as the gold standard.
The characteristics of the included studies are detailed in Table 1. The majority of adenomas (71.1%) demonstrated an increase in SUVmax from early to late phase. A positive correlation was observed between SUVmax in late phase and PTH level and between ABR in late phase and PTH level.
The surgical specimen volume was positively correlated with the PET/MR volume and PTH level.

Positron Emission Tomography Procedure and Radiation Exposure
As can be seen in Table 1, there was no standard dose administered. The mean dose was 2.5-3 MBq/kg in most of the studies but even lower doses did not appear to have an impact on performance.
A crucial point in parathyroid imaging is the correct timing of the post-injection acquisition that mainly differs in the studies according to radioisotope decay used (only 20 min half-life for 11 C compared with 110 min for 18 F). In 15 studies, the authors chose to perform only a single acquisition whereas a dual-phase protocol was the strategy of choice in 16 studies. In particular, a dynamic acquisition was performed in nine studies, an early static acquisition (set at 5 min in most cases) was taken by 12 authors, while a 60 min post-injection acquisition protocol was the most used timing set by the authors. Rep et al. were the only ones to perform a triple time-point protocol.
PET/CT was acquired in the supine position, extending the field of view (FOV) from the base of the skull to the diaphragm in order to include possible ectopic adenomas. None of the authors specified the preferred position of the arms. However, considering a head-neck district investigation, it is highly recommended to perform the scan with the arms down, hyperextending the neck, to avoid any attenuation by the body and reduction of the area visualization. In addition, Michaud et al. suggested extending the acquisition up to mid thighs in men in order not to miss potential prostate foci [17,21].
Regarding radiation exposure, few results are available but promising. Rep

Diagnostics Performance
In Table 2 the detailed diagnostic performances of the main studies used for the comparison between nuclear medicine imaging methods are reported.

Discussion
We firstly detected occasional findings on choline PET/CT performed for other reasons (i.e., Figure 2). In recent years interest has grown in the use of this multimodality imaging method in the preoperative localization of hyperfunctioning parathyroid glands.
Michaud et al. were one of the first to investigate the role of [ 18 F]FCH PET/CT in HPT as second line imaging. In the two studies performed by the same group, a heterogeneous population of patients suffering from both primary and secondary HPT was considered, reaching a lesion-based sensitivity of 89% but a detection rate on a per patient level of 92% [17]. More recently, Uslu-Beşli et al. confirmed the superiority of choline PET when compared with inconclusive traditional imaging methods [41]. Michaud et al. were one of the first to investigate the role of [ 18 F]FCH PET/CT in HPT as second line imaging. In the two studies performed by the same group, a hetero geneous population of patients suffering from both primary and secondary HPT wa considered, reaching a lesion-based sensitivity of 89% but a detection rate on a per pa tient level of 92% [17]. More recently, Uslu-Beşli et al. confirmed the superiority of cho line PET when compared with inconclusive traditional imaging methods [41].
Although used as a second line investigation when [ 99m Tc]Tc-MIBI and US imaging were negative or inconclusive, the idea of using choline PET/CT as first line imaging i now supported by promising data.
In  [18], up to 98% of sensitivity in other studies [38,42].  Although used as a second line investigation when [ 99m Tc]Tc-MIBI and US imaging were negative or inconclusive, the idea of using choline PET/CT as first line imaging is now supported by promising data.
In most articles of the available literature, choline is [ 18 F]labeled, probably due to its favorable characteristics, with only four studies employing [ 11 C]labeled choline.
In terms of sensitivity, both [ 11 C]labeled and [ 18 F]labeled choline showed higher performance compared with traditional imaging methods. The results from a large sample of patients studied by Orevi et al. showed a higher sensitivity of [ 11 C]Choline PET/CT (92.3%) when compared with [ 99m Tc]Tc-MIBI (88.5%) in parathyroid adenoma detection [18], up to 98% of sensitivity in other studies [38,42]. First of all, the higher spatial resolution of PET/CT allows the identification of even smaller lesions [16]. Cuderman et al. showed that lesions with significantly smaller volumes and weights were correctly identified only on PET/CT imaging. The authors found that lesions with a median volume of 4.1cm 3 and a median weight of 0.5g resulted as a true positive on PET/CT but as a false negative on conventional scintigraphy even if integrated with SPECT/CT [40]. Beheshti et al. confirmed this fact highlighting that the sizes of the parathyroid adenomas were significantly different between negative and positive [ 99m Tc]Tc-MIBI/tetrofosmin SPECT/CT imaging [11].
A noteworthy aspect, directly correlated with a higher spatial resolution, concerns the ability of the [ 18 F]FCH PET/CT to detect multiglandular pathology or parathyroid hyperplasia as well as adenomas. In fact, it is known that the diagnostic performance of [ 99m Tc]Tc-MIBI scintigraphy is limited in the case of hyperplasia. The reason is not fully understood but a different mutation in the mitochondrial DNA of the adenoma might be the cause [40,43,44].
Among the collected data, there were no studies that exclusively considered patients with multiglandular pathology or parathyroid hyperplasia. However, in a subanalysis conducted by Cuderman et al., [ 18 F]FCH PET/CT showed a sensitivity of 88% in the detection of multiple parathyroid adenoma/hyperplasia compared with 34%, 22%, 22% and 44% of SPECT/CT, subtraction, dual-phase and combined protocols, respectively [40]. In another study, [ 18 F]FCH PET/CT reached a sensitivity of 91% compared with 50% of conventional imaging [16].
Although there are no studies that specifically focused on ectopic parathyroid detection, [ 18 F]FCH PET/CT has been shown to be superior to conventional imaging. Interestingly, Thanseer et al. reported in their series a case of an ectopic localization in a woman with recurrent HPT who had suffered an accidental implantation of parathyroid tissue in the left breast region (~0.8-0.4 cm) following an endoscopic parathyroidectomy through a periareolar approach five years earlier. Only [ 18 F]FCH PET/CT successfully localized the parathyroid adenoma in this rather unusual location [22].
In patients with persistent or recurrent primary HPT after surgery [ 18 F]FCH yielded a 95.8% sensitivity when compared with [ 99m Tc]Tc-MIBI (50.0%), US (54.2%) and 4DCeCT (75.0%) [7]. Sensitive preoperative imaging becomes essential for this setting of patients as well as in patients with a history of significant neck surgery such as a total thyroidectomy. In fact, neck surgery often leads to scar formation and/or anatomical modification, increases the risk of recurrent laryngeal nerve injury and is related to post-operative permanent hypoparathyroidism in up to 20% of cases, thus complicating the detection of abnormal parathyroid gland(s) [33,45,46].
Currently, primary HPT has been the most studied condition because surgical treatment is the standard of care and preoperative localization is essential. For this reason, the literature is poor in studies that exclusively analyzed patients with secondary HPT as well as with tertiary HPT. F]FCH PET/CT seemed to be able not only to provide additional information for surgery than traditional imaging methods but also to detect any bone complications that often occur in the tertiary form [36,49].
Finally, two studies compared [ 18 F]FCH PET with 4DCeCT, a method that provides both functional and anatomical information thanks to the addition of the fourth dimension consisting of changes in contrast attenuation over time [39]. It was shown that [ 18 F]FCH PET/CT and 4DCeCT had quite similar detection rates (57% and 55%, respectively) and sensitivities (81% and 74%, respectively) and, consequently, integrated [ 18 F]FCH PET/4DCeCT could yield the highest reliability (73% detection rate and 100% sensitivity) [32].
The results of choline PET in combination with magnetic resonance (MR) imaging are promising (90% sensitivity, 100% PPV in the Kluijfhout et al. study [24]) but more research and a larger sample size are needed [28,35,50].

Pitfalls
The diagnostic performance of choline PET in HPT is affected by both false positive and false negative results.
False positives could be related to malignant disease such as differentiated thyroid carcinoma and/or metastatic lymph nodes as well as other benign conditions including thyroid nodules or inflammatory lymph nodes especially for ectopic locations [7,28].
In the setting of benign lesions, dual time scanning can help with a differential diagnosis; in the late phase, choline uptake decreases both in the thyroid gland parenchyma and inflammatory lymph nodes whereas it increases in parathyroid adenomas [25,26,31]. Similarly, Prabhu et al. reported that early dynamic imaging could also be helpful in differentiating between parathyroid adenomas and cervical reactive lymph nodes that might be confused with small adenomas/hyperplasic glands. The authors demonstrated that the adenoma/thyroid ratio was significantly higher than the lymph node/thyroid ratio (p = 0.0117) [30]. Not least, a high uptake of choline has been documented in oncocytic thyroid adenomas [51][52][53].
On the other hand, false negative findings could be related to very small parathyroid adenomas, particularly in a close vicinity to the thyroid gland [32], to the ectopic localization of the parathyroid gland, to multiglandular disease [7] or in cases of intrathyroidal atypical parathyroid adenoma, which is a rare entity with borderline pathological characteristics between adenoma and carcinoma [13,54].
In addition, in patients with autoimmune thyroid disease such as Hashimoto thyroiditis or Graves' disease a diffuse choline uptake in the thyroid gland can mask a possible parathyroid adenoma resulting in a false negative finding [3,[55][56][57]. However, another study documented that the presence of autoimmune thyroid disease did not compromise the use of choline PET/CT from an accurate localization of parathyroid adenoma [15,58].
All of these pitfalls could complicate the correct localization of enlarged parathyroid glands either by US or [ 99m Tc]Tc-MIBI scintigraphy [29,58,59]. However, thanks to the above-mentioned advantages of choline PET/CT in terms of a higher spatial resolution, a better lesion-to-background ratio as well as the added value of semi-quantitative analysis, it overcomes traditional imaging in facing a few of these important issues.

Semi-Quantitative Analysis and Biochemical Data Correlation
As mentioned above, different protocols were studied with special reference to the semi-quantitative parameters.
Prabhu et al. supported the use of dynamic acquisition instead of delayed static imaging, demonstrating that both the maximum standardized uptake value (SUVmax) and the parathyroid/thyroid ratio (P/T ratio) were comparable in the two acquisition time protocols [30].
When comparing early and delayed static acquisitions, several authors observed a better lesion-to-background and lesion-to-thyroid contrast on late images but all lesions were visible at both phases [16,17,21,31].
Other studies found an increase in the SUVmax of the lesion on late images, suggesting that a single late acquisition may be sufficient in patients with high PTH levels [22,23,27].
In the only study considering a triple time-point protocol, Rep et al. found that the optimal scan time was at 60 min after the tracer injection and had a higher accuracy than early images and the same accuracy of images obtained at 120 min (diagnostic accuracies of 94.1%, 96.5% and 96.5% at 5, 60 and 120 min after injection, respectively) [19].
Considering that some lesions could be missed in one-phase acquisition, the dual time-point protocol seemed to be the optimal parathyroid study [31]. More studies are necessary to establish the role of dynamic/early static acquisition not only for comparison with the late images but also for the possibility of performing it as a single scan, reducing examination times especially for uncooperative patients.
Data related to semi-quantitative parameters of choline PET/CT are not uniform and a standardized cut-off does not exist to differentiate towards physiological and pathological parathyroid uptake. In fact, the SUVmax was similar in the case of true positive and false positive uptakes (mean value 4.5 vs. 3.1, respectively) [7]. In a few articles, parathyroid adenoma showed a higher SUVmax than hyperplasic glands (mean value 6.80 vs. 4.53) but the difference was not significant [11]. However, additional information could be obtained by comparing the parathyroid uptake with the background and/or the thyroid uptake [29]. The P/T ratio ranged from 1.2 to 2.5 on average [11,20,21].
A few authors investigated whether the SUVmax may be predictive of disease severity by exploring a possible correlation between the SUVmax values of hyperfunctioning parathyroid gland(s) and serum PTH, Ca, phosphorus (P) levels and bone mineral densitometry (BMD) results. In 2018, a retrospective study of 52 patients highlighted a strong positive correlation between the SUVmax and the PTH level as well as between the adenoma-to-background ratio and the PTH level on both late and early phases [25].
Morland et al. identified a significant association between [ 18 F]FCH PET/CT and higher calcium levels; in their study, [ 18 F]FCH PET/CT was performed during Cinacalcet treatment in 23% of patients [32]. Cinacalcet is a modulator of the calcium-sensing receptor strongly expressed on the surface of parathyroid cells that reduces serum calcium and PTH levels in patients for whom surgery is not clinically indicated or contraindicated. The authors found that Cinacalcet did not negatively impact the [ 18 F]FCH PET/CT results. Rather, patients displaying positive [ 18 F]FCH PET/CT tended to be more frequently under Cinacalcet treatment presumably because of higher calcium levels [1,60,61]. Interestingly, calcium was the only parameter that showed a significant association with positive findings on integrated [ 18 F]FCH PET/4DCeCT [32].
The correlation between the SUVmax and densitometric results was also examined. Serum PTH levels were significantly higher and femoral/lumbar T values were lower in patients with an SUVmax greater than 4.4 [13]. Choline uptake seemed to be correlated with the molecular profile too; a significant correlation between the SUVratio and Ki67 expression and an inverse correlation between SUV parameters (SUVmax; SUVratio) and p53 expression was demonstrated [32].
At the time of writing this review, the phase III clinical trial APACH2 is ongoing [62]. As highlighted in this literature overview, the trial aims to demonstrate that the high diagnostic performance of [ 18 F]FCH PET/CT could improve the outcome of patients, resulting in a saving in terms of total clinical care costs. Preliminary results are waited for to confirm the direction in the near future.

Conclusions
In conclusion, due to the high diagnostic value demonstrated by the reported data, choline PET could represent the ideal method to replace traditional imaging in the preoperative localization of hyperfunctioning parathyroid gland(s), demonstrating it to be a significantly more sensitive method in different patient settings than any conventional scintigraphic approach used.

Conflicts of Interest:
The authors declare no conflict of interest.