Near-infrared fluorescent imaging techniques for the detection and preservation of parathyroid glands during endocrine surgery

Abstract Objectives In over 30% of all thyroid surgeries, complications arise from transient and definitive hypoparathyroidism, underscoring the need for real-time identification and preservation of parathyroid glands (PGs). Here, we evaluate the promising intraoperative optical technologies available for the identification, preservation, and functional assessment of PGs to enhance endocrine surgery. Methods We performed a review of the literature to identify published studies on fluorescence imaging in thyroid and parathyroid surgery. Results Fluorescence imaging is a well-demonstrated approach for both in vivo and in vitro localization of specific cells or tissues, and is gaining popularity as a technique to detect PGs during endocrine surgery. Autofluorescence (AF) imaging and indocyanine green (ICG) angiography are two emerging optical techniques to improve outcomes in thyroid and parathyroid surgeries. Near-infrared-guided technology has significantly contributed to the localization of PGs, through the detection of glandular AF. Perfusion through the PGs can be visualized with ICG, which can also reveal the blood supply after dissection. Conclusions Near infrared AF and ICG angiography, providing a valuable spatial and anatomical information, can decrease the incidence of complications in thyroid surgery.


Introduction
More than 80,000 thyroid and parathyroid surgeries are performed yearly in the United States for the treatment of both benign and malignant thyroid diseases [1,2]. Even for the most experienced surgeons, parathyroid glands (PGs) can be hard to distinguish from surrounding tissue due to their small size, variable position, and similarity in color to fat tissue [3,4]. Consequently, PGs can be damaged, removed, or devascularized during endocrine surgeries, leading to postoperative complications. The most common complication resulting from inadequate PG identification and preservation is transient or definitive postoperative hypoparathyroidism, affecting 20-35 and 1-10%, respectively, of thyroidectomy patients [5,6]. The risk of hypoparathyroidism increases in cases of total thyroidectomies, central neck dissections, and reoperative procedures [7,8]. The low plasma levels of parathormone (PTH) associated with hypoparathyroidism can cause hypocalcemia, adversely affecting clinical outcomes and the quality of life of the patient [9]. To reduce the incidence of postoperative hypocalcemia, proper intraoperative PG identification and preservation are necessary [10,11].
PG preservation techniques vary from basic visual inspection to highly developed optical techniques. Visual inspection is the standard procedure to identify and preserve PGs, but as the technique depends on the surgeon's experience, it is not reliable [12].
Magnification glasses with a meticulous parathyroid dissection have been proposed to prevent definitive hypoparathyroidism and also to decrease the postoperative incidence of transient hypocalcemia. However, this has not been clearly demonstrated in the literature. A systematic review and meta-analysis by Spadalis et al. state that the use of magnification techniques to identify recurrent laryngeal nerves and PGs appears to be as safe as direct vision. However, magnification techniques do not decrease the risk of recurrent laryngeal nerve injury and transient hypocalcemia after thyroid surgery compared with direct vision [13]. Moreover, in a recent article, André et al. [14] demonstrated that an active and systematic search of the parathyroid gland is not recommended due to the increased risk of gland lesion, mainly by devascularization. This has been largely studied in recent years, and the conclusion seems to be that it is best not to seek out PGs, as there is a risk of devascularization during dissection. No correlation has been shown between the number of glands detected and the risk of hypocalcemia; notably, the risks of hypocalcemia and hypoparathyroidism have been shown to increase with the number of glands visualized [10,[15][16][17][18][19][20]. Surgeons can also intraoperatively identify PGs by sacrificing a slice of tissue for frozen section analysis or use the "float or sink" method for tissue density analysis. Both visual inspection methods are invasive, but newer identification techniques have been developed to identify PGs without damaging tissue. Near-infrared autofluorescence imaging (NIRAF) and indocyanine green (ICG) fluorescence imaging are promising optical methods for guiding PG identification and preservation [21].
Here, we present an overview of innovative nearinfrared (NIR) imaging methods for PG identification during neck endocrine surgery, evaluate their clinical results, and suggest directions for future research.

Early techniques for PG identification
PGs are difficult to locate with the naked eye, making disruption of the PG vasculature a common complication during thyroidectomies. Disruption of the PG vasculature can cause temporary and definitive postoperative hypoparathyroidism [10,11,22,23]. Therefore, there is a great demand for a reliable technique to rapidly identify PGs during surgery. Although numerous localization techniques have been evaluated, only a few intraoperative imaging modalities have been deemed effective [24].
The use of a dye or a fluorophore to visualize PGs is not new. In 1971, Dudley described the colorimetric localization of PGs with the use of intravenous methylene blue (MB) as an exogenous contrast agent [25]. Prosst et al. [26] proposed the use of aminolevulinic acid (ALA) as an alternative contrast agent, but ALA was not generally accepted due to difficulties in its application to clinical practice [27].
Rubinstein et al. introduced the first noninvasive highresolution imaging technique, optical coherence tomography (OCT), which provides a microarchitectural characterization of structures up to 2 mm thick [28]. OCT images show characteristic features of the thyroid, parathyroid, lymph nodes, and adipose tissue, facilitating consistent identification. However, in vivo application trials did not achieve similar favorable results, due to technical problems and the difficulty of handling the OCT probe covered with a sterile sheath [29].
Van der Vorst et al. [21] explored the use of NIR fluorescence imaging using small, intravenous doses of MB. However, no reliable improvements in PG detection were reported. In addition, MB is potentially toxic, causing serious adverse neurological events. Therefore, no prospective randomized study has been performed with MB, and the literature has discouraged its use [30][31][32]. Thus, the clinical approach to parathyroid identification remained stagnant until the discovery of parathyroid autofluorescence (AF) in 2008. This discovery renewed interest in developing intraoperative PG imaging techniques that rely on the autofluorescence of parathyroid tissue in the NIR spectrum [33].

PG identification and exogenous fluorescence imaging
Fluorescence imaging is a well-established technique used in the biomedical sciences for both in vitro and in vivo visualization of cells and tissues but has only recently been applied to surgery [34]. Fluorescence imaging leverages the property of certain substances and molecules to absorb light at a given wavelength, which briefly raises the energy of the molecule to a higher excited state. The molecule then emits light at a higher wavelength with lower energy, which can be detected and measured [34,35]. Fluorescence imaging is conducted with exogenously administered contrast agents, such as fluorescent MB or 5-ALA [25,32], but there is also a label-free optical method that relies only on the intrinsic autofluorescence of certain tissues.
Hillary et al. and Tummers et al. found that fluorescent MB enabled the localization of normal PGs for up to 145 min after administrating the dye [24]. The fluorescent signal was found to be 2.6-fold higher in PGs than in thyroid tissue and 4.3-fold higher than in muscle tissue [36]. However, this approach was not further developed due to emerging concerns regarding the adverse neurotoxic effects of MB [32].
Two case reports [37,38] and three case series [27,39,40] have reported the use of 5-ALA for intraoperative PG identification. ALA, a metabolic-targeting contrast agent, is the precursor of the fluorescent molecule porphyrin, which is an intermediate in the heme synthesis pathway [41]. ALA uptake increases as the number of mitochondria in PG cells increases [27]. Oral administration of 20-30 mg/kg fluorescent ALA allows detection of PG for 1-8 h. However, exposure to direct light up to 24-48 h after ALA administration causes photobleaching and many other phototoxic effects on both the skin and eyes. Additionally, the fluorescent signal is absent in some pathological PGs [27,39].

PG identification and autofluorescence
Until recently, fluorescence imaging techniques relied on the interaction between serum proteins and fluorescent dyes. Recently, however, the autofluorescence of multiple tissue samples has been reported. PG tissue, in particular, exhibits a unique autofluorescence signature. When the PG tissue is excited with a NIR laser at a wavelength of 785 nm, there is a spontaneous and immediate emission of fluorescent light at 820-830 nm from PGs with a 2-to 11-fold signal enhancement over that of the surrounding tissue [2]. An NIR camera or spectroscope can be used to detect autofluorescence. The detection of autofluorescence is a dye-free technique that allows noninvasive, real-time identification and precise localization of PGs. The intrinsic fluorophore responsible for this optical effect in PGs is still unknown; however, evidence suggests that it could be a calcium-sensing or a vitamin D receptor [2,42].
Subsequent studies of thyroid surgery and PG diseases have further characterized PG autofluorescence, and have evaluated the applicability and practicality of this technique. McWade et al. and others have reported excellent PG detection rates, with a specificity of more than 80% [2,29,42,43]. However, the autofluorescence signal is influenced by variables such as disease state, preoperative vitamin D levels, serum calcium levels, and body mass index [44]. To accurately identify the PGs tissue, multiple measurements are taken on each gland. A spectroscopy probe must be in contact with the tissue to make point-by-point measurements, while an optical NIR fluorescence imaging camera is contactless and provides a larger field of view. Another limitation of this technique is the difficulty of distinguishing between thyroid and parathyroid tissues. In some disease states, such as thyroiditis, the contrast in autofluorescence intensities of the two tissues may be diminished. Furthermore, brown fat, colloidal nodules, and metastatic lymph nodes can exhibit autofluorescence which may overlap with the autofluorescence from PG tissue, resulting in false positives [4,45]. Moreover, without a standardized quantification method, the detection of an autofluorescence signal in PGs is based on the qualitative interpretation of the surgeon. Therefore, the development of devices to quantitatively detect autofluorescence in the NIR spectrum was necessary.

NIRAF devices
Currently, there are two commercially available NIRAF devices approved by the Food and Drug Administration (FDA) which are suitable for performing the real-time identification of PG tissue during surgery: Fluobeam ® (Fluoptics©, Grenoble, France) and the PTeye ™ Parathyroid Detection System (Medtronic, Inc. Minneapolis, USA previously AIBiomed Inc., Santa Barbara, CA, USA). The systems differ in the exact emitted and detected wavelength, the type of handpiece provided for detection, and the type of display used [46]. The Fluobeam ® system is used to explore the surgical field directly, reproducing a real-time grayscale image through the enhancement and detection of autofluorescing tissues. Conversely, the PTeye System includes a sterile probe, which must make contact with the tissue to analyze its optical properties, producing distinct audio and visual signal.
Karl Storz and Stryker (previously Novadaq) have also developed fluorescence devices for exogenous dyes, such as ICG, but these devices are unable to detect autofluorescence (Table 1).

NIRAF in clinical practice
The introduction of NIRAF has vastly improved intraoperative imaging [47,48]. Many recent studies (Table 2) have demonstrated that the use of NIRAF during thyroid surgery can improve surgical outcomes by facilitating PG identification (Figures 1-3), which reduces the incidence of postoperative hypoparathyroidism. In particular, Benmilloud et al. showed that NIRAF improved PG identification and helped to reduce the rate of temporary postoperative hypocalcemia, parathyroid autotransplantation, and inadvertent parathyroid resection [48,66]. In the literature, NIR  devices have been shown to facilitate parathyroid gland identification by detecting their AF before conventional, visual recognition by the surgeon, in 37-67% of cases [49]. Additionally, these devices enable the early identification of PGs before surgical dissection, helping to preserve their vasculature [50]. A systematic review and meta-analysis by Barbieri et al. found that NIR fluorescence imaging reduced short and medium-term hypocalcemia compared to conventional surgery [51]. Moreover, NIRAF can detect subcapsular/intrathyroidal PGs or PGs that have been accidentally removed, which can then be resected from the thyroid specimen and auto planted back into the patient (Figures 4 and 5). In addition to the identification of normal PGs, NIRAF can help to identify a pathological PG. Parathyroid adenomas exhibit a more heterogeneous and less intense autofluorescence signature than that of normal PGs (Figure 6), enabling the differentiation between a normally functioning and a pathological PG [52,53]. Furthermore, the fluorophore is resistant to freezing, heating, and fixing with formalin, meaning that the autofluorescence properties are preserved after gland resection [4,29,42,43]. However, the perfusion status and viability of the PG tissues cannot be assessed on the basis of this single parameter. Considering the importance of PG vitality in thyroid surgery, a method for the confirmation of vascular integrity is also needed.

PG identification and ICG angiography
ICG is an FDA-approved fluorescent contrast agent which has been used since 1956, and, when systemically administered, has few side effects and low toxicity [43]. ICG fluorescence imaging enables the real-time direct imaging and assessment of tissue perfusion and vascularization, and as a result, it has become one of the most common imaging techniques. ICG is excited at 780-805 nm and emits a maximum NIR signal at 830-835 nm [34,54]. IGC is an amphiphilic tricarbocyanine dye that travels through the circulatory system, has a half-life of approximately 3-5 min, and is excreted in 15-20 min by the biliary system [55]. For visualization, a dose of 0.2-0.5 mg/kg is recommended and can be repeated as required, as long as the daily dose does not exceed 5 mg/kg [55,56]. Approximately 30 s to 2 min after injection, NIR fluorescence camera images can capture the distribution of ICG in the explored tissue. ICG was first used in ophthalmology for the detection of macular degeneration and retinal angiography [34,35]. Soon after, this technique was applied to colorectal surgeries for the evaluation of intestinal anastomosis by imaging the intestinal microvasculature [57][58][59].
Most recently, ICG has been proposed to be the most suitable agent for the intraoperative assessment of the PG vascularization, which is closely correlated with parathyroid function (Figures 7 and 8) [56,[60][61][62][63][64]. Many studies have evaluated the use of ICG angiography to identify normal PGs and assess their perfusion in thyroidectomies (Table 3). These studies suggest that ICG angiography is a safe, feasible, effective, and easy technique to identify and preserve PGs [65]. Owing to the contribution of ICG angiography to intraoperative decision-making, the technique may be widely used in the future, especially as new applications are developed. For example, the technique is being developed for mini-invasive and robotic surgery applications, including the transoral endoscopic thyroidectomy vestibular approach and the robotic bilateral axillo-breast approach [66], although studies are limited.  Studies suggest that using ICG angiography can help preserve at least one well-vascularized PG after total thyroidectomy. Normal PTH levels have been recorded on the first postoperative day, which indicates a 100% of positive predictive value for excluding postoperative hypoparathyroidism [63,65,[67][68][69]. However, some studies have shown less positive results for the prediction of parathyroid function [62,70]. This difference is likely related to the individual visual interpretation of the grayscale images produced during angiography, which lack standardized numerical criteria [56]. In clinical practice, the vascularization of PGs can be evaluated either by quantitative analysis of the ICG fluorescence signal [67] or by qualitative estimation of dye uptake, which depends on the surgeon's judgment [56,61,71]. To date, quantification using ICG imaging is only possible in post-processing and not in real-time. Still, the available literature suggests that PG function resumes in patients with moderately wellvascularized PGs.
One of the challenges of this technique is that even the most experienced surgeon may confuse the PGs with other anatomical structures, such as the thyroid, thymus nodules, or lymph nodes [72]. In such cases, the assessment of good vascularization of what is actually non-PG tissue could lead to false assumptions about the risk of developing postoperative hypoparathyroidism. Thus, ICG angiography is currently only used to evaluate PGs after surgical resection. While there are many promising clinical trials aimed at reducing the extent of PG resection and the incidence of post-thyroidectomy hypoparathyroidism using ICG angiography [64], further studies are needed.

Conclusion
Although most thyroid and parathyroid operations are not complicated, difficulties in accurately locating PGs during thyroid surgery may result in the inadvertent resection of PG and the disruption of the PG vasculature, causing postoperative hypoparathyroidism. It is too early to assume that NIRAF imaging and ICG angiography will have a major role in the intraoperative identification and preservation of PGs, but both modalities have provided valuable spatial and anatomical information. The techniques can be applied easily during surgery, enabling the real-time detection of both healthy and diseased PGs, and have reliably and substantially decreased the incidence of complications in thyroid surgery. However, it is important to understand the limitations of these techniques and to improve their surgical applications. Therefore, further studies, including randomized controlled trials and meta-analyses, are needed.