**4. Discussion**

An increasing number of scientific reports have investigated the utility of fluorescenceguided surgery in preserving the PGs and preventing hypoparathyroidism as a complication of thyroid surgery. However, only three studies have addressed thyroid cancer patients alone. In contrast, in the majority of reports, the patient cohorts are mixed, including both benign and malignant indications and thyroid as well as parathyroid diseases. According to the available literature, NIRAF imaging improves the visualization and preservation of the PGs during thyroidectomy. This technique is feasible and safe, as NIRAF imaging is non-invasive. Moreover, no NIR imaging-related complications were reported in any of the studies, including the studies with ICG injection.

Among the included reports, NIRAF was used in 11 studies, and an exogenous fluorophore was employed in 14 studies. Because NIRAF exploits the endogenous fluorophore of the PGs, no additional time is required to administer an exogenous dye [42]. Consequently, the only equipment required for NIRAF imaging is a fluorescence camera system. A comparative study of NIRAF and ICG imaging found only a minimal difference in the usefulness of both techniques for identifying the PGs [54]. However, preserved perfusion can be visualized only by the introduction of a contrast agent (e.g., ICG) via the blood stream. Therefore, NIRAF is technically inappropriate for assessing the integrity of arterial supply and venous outflow. In contrast to ICG, NIRAF cannot determine the vascular integrity of the PG. ICG has a short half-life of 3–5 min and is eliminated by the hepatic system after approximately 15 min [19]. Hence, repeated doses can be required and safely administered [50,60,61]. The toxic dose in adults is 5 mg/kg circulating per time point [19]. All doses administered in the included studies were well below this threshold. ICG was administered intravenously in all cases, and a duration of 1–3 min was required for the PG to take up the ICG [44,48,52,72]. Upon excitation with NIR light, the fluorescence produced by the gland was scored as 0, 1, or 2 if the gland appeared black (not vascular), grey (partially vascular), or white (vascular), respectively [48,60,61].

The prevailing causes of hypoparathyroidism after thyroidectomy are disruption of the blood supply to the glands and their incidental removal. These complications are more frequently seen in thyroidectomy for thyroid cancer [19], especially with accompanying lymph node resection. Because of the ability of NIR imaging to visualize the PGs and to detect impaired blood supply to the glands during thyroid cancer surgery, this technique is invaluable in averting hypoparathyroidism. In a randomized controlled trial, there was a

significant difference between the number of PGs identified with NIRAF compared with conventional surgery. The percentage of patients with four identified PGs was higher in the NIRAF group than in the standard-care group (47.1% vs. 19.2%, respectively; *p* < 0.001). Consequently, the number of incidentally resected PGs was significantly lower in the NIRAF group [45]. In another controlled study, visual and ICGA assessment of vascularity showed agreement in 87% of cases. ICGA prevented autotransplantation of 19 (6.8%) glands that were shown to be adequately perfused on ICGA, but would have been transplanted if visual inspection alone had been used. Interestingly, autotransplantation of the PGs was significantly more common among the ICGA group than the control group (36% vs. 12%, *p* = 0.0001) [49].

In a retrospective study [51], the use of white light and the use of ICG and NIR imaging were compared for identifying the PGs at an early dissection stage. The surgeons documented the number of PGs that were visualized with the naked eye using white light. Next, the operating room's light was turned off, and the surgical field was illuminated with NIR light. The number of PGs visualized with NIR light was significantly higher than in white light. Furthermore, the integrity of the blood supply to the PGs was assessed via ICGA in cases of ambiguity. Perfusion was intact in all cases; hence, no autotransplantation was performed. While the majority of the studies in this review reported improved visualization of the PGs in favor of NIR fluorescence imaging, one study [50] reported slightly better visualization of the glands in white light. It is possible that the bright background signal of highly vascularized malignant tissue may have interfered with visualization on fluorescence imaging. In the same study, the majority of surgeons considered NIR fluorescence imaging to be advantageous. The reduction in both the autotransplantation rate and the inadvertent parathyroid resection rate, which are both risk factors for postoperative hypocalcemia, reported by studies in this review is corroborated by Fanaropoulou et al. [19].

The largest series reported in the literature included 542 patients [66] and focused exclusively on thyroid cancer. This study found no significant differences in the number of visualized PGs between the NIRAF group and the conventional group. This finding could be due to the substantial experience of the surgeon. The same article still reported a significantly lower incidence of transient hypoparathyroidism in the NIRAF group compared with the control group. Based on the large series included in this study and the quality of the study, this statement seems to be well supported.

Visualization of the PGs alone may not be sufficient to prevent postoperative sequelae. Dissection may lead to inadvertent damage to the end arteries that supply the PGs. Hence, some studies addressed the assessment of PGs after resection of the thyroid gland to ensure its viability and/or the need for autotransplantation [47,48,52]. Enny et al. [47] found that the identification of PG viability by NIR imaging visualization of intact vascularity at the end of the procedure is associated with a reduced need for routine postoperative calcium and vitamin D supplementation, along with reductions in accompanying costs and duration of hospital stay. However, some authors [48] recommend that surgeons interpret these results with caution, as their study on the use of ICGA did not show a significant improvement in thyroid surgery outcomes. In particular, poor vascular perfusion on ICGA did not correlate with a postoperative reduction in PTH levels or transient hypocalcemia. The ability of ICGA to detect impaired perfusion and to induce a change in surgical strategy was demonstrated in colorectal surgery with a positive outcome [23].

Thirteen studies reported on postoperative PTH levels, whereas 17 studies reported serum calcium levels. Jin et al. [73] performed a small study assessing PG perfusion via ICGA after resection of the first thyroid lobe, with serum PTH and calcium levels assayed on postoperative day (POD) 1, 7, and 14, and 6 months after operation. PGs that had been accidentally devascularized during the surgery were identified using ICGA and were autotransplanted. Their findings showed that the ICGA score was predictive of postoperative PTH levels, as all of the 22 patients with at least one PG having an ICG score of 2 (84.6%) had normal postoperative PTH levels, while half of those with poor vascularization on ICGA developed transient hypoparathyroidism. Similar findings were reported in another study [60], where all patients with at least one well-vascularized gland as detected by ICGA (83.3%) had normal serum PTH levels. One of these patients had asymptomatic hypocalcemia at POD 10. In contrast, two out of six patients with poor vascularization on ICGA developed transient hypoparathyroidism. Overall, postoperative supplementation with calcium or vitamin D was averted in all patients. Thus, ICGA can be used intraoperatively to determine whether at least one PG remains vascularized and functional, in order to predict the absence of postoperative hypoparathyroidism.

The use of fluorescein dye to visualize the PGs was found to be superior to naked eye assessment with regard to preventing hypocalcemia. Among patients with >3 PGs visualized via fluorescein dye, none suffered clinical hypocalcemia, as opposed to more than 20% with clinical hypocalcemia when >3 PGs were visualized with the naked eye only [47].

Several comparative studies [45,53,64] reported better postoperative PTH and calcium levels among NIR imaging groups compared with conventional surgery groups, but the opposite trend was observed by others [48,49]. Among the latter, Razavi et al. suggested that intraoperative monitoring of PTH levels should remain, which they suggest as a better method than ICGA for predicting PG vascularization. Rudin et al. opined that the potential of ICGA to quickly identify PGs and to positively alter surgical strategy cannot be ignored, but that more research should be conducted to determine whether this practice reduces the risk of postoperative hypoparathyroidism.

This controversy is fueled by a lack of standardization among NIR fluorescence techniques, such as in the timing of assessment and the quantification of fluorescent signal intensity. PG visualization can be enhanced with NIR imaging prior to any dissection or during thyroid dissection. Early identification can prevent inadvertent resection or impairment of vascular supply. In this way, NIR imaging can contribute to the preservation of all PGs. The distinction between the more intense fluorescence of the PGs compared with that of the lymph nodes can provide guidance during lymphadenectomy. Late detection, such as that performed on the specimen, can allow functional recovery of the PG with autotransplantation. Contrast-enhanced fluorescence imaging with intravenous ICG or fluorescein injection can visualize preserved versus impaired perfusion (e.g., congruent results shown intraoperatively for adrenal perfusion) [9]. However, except for quantitative contrast-enhanced NIR imaging research protocols [74], fluorescent signals and ICG scores are generally assessed subjectively. For ICG injections, the rate at which the fluorescence signal intensity increases, as well as the added signal intensity produced by reinjections, can only be objectively assessed with a quantitative approach, which has not yet been integrated into commercially available NIR imaging systems.

Furthermore, it can be concluded from this review that the use of NIR imaging is safe and feasible. No study reported any complication arising from this technique. Rare allergic reactions to ICG have been previously reported [75]. These reactions are particularly attributable to iodine or sodium iodide, which is often added to improve fluorophore solubility [19]. Therefore, patients with iodine allergy are not eligible for ICG administration. However, ICG solutions without iodine are commercially available (e.g., INFRACYANINE, SERB S.à.r.l.); thus, this categorical exclusion of patients for iodine allergy or hyperthyroid status might be lifted in the future.

Several studies in this review reported minimal added time (3–10 min) for NIR imaging but this was shown as not statistically significative [45,54,62]; anyway, this added time would be negligible, considering the benefits of the technique. Similar findings were reported by van den Bos et al. [50]. Interestingly, and in contrast to the observation of a longer duration with the use of fluorescence imaging, two studies reported that more time was spent during conventional surgery [45,64]. Although this trend was not elucidated, it may have been due to a more rapid identification of the glands using autofluorescence, without the added duration needed for the administration and uptake of an exogenous fluorophore.

Concerning ICGA, among the authors who analyzed this factor, the procedure required an additional operative time of 2–5 min; however, this approach has the benefit of eliminating any doubt on the viability of the explored PGs. Moreover, according to some authors [47,61], ICGA can also prevent systematical postoperative calcium supplementation, thus eventually reducing the hospital stay. The use of fluorescence imaging in surgery has been reported as cost-effective [44,76]. However, no study in the present review stated the cost implication of NIR imaging, which is mainly associated with equipment purchase. The camera systems required for NIR imaging can be expensive. However, once a fluorescence system is installed, autofluorescence creates no further cost. Hence, the additional cost is limited to the price of the exogenous fluorophore used per patient [77] and the sterile cover. When assessed in light of the deleterious health effects and economic burden of hypoparathyroidism and hypocalcemia, NIR imaging can be viewed as very cost-effective, as demonstrated by its positive impact.

In a cost–benefit analysis on the usefulness of routine intraoperative intact parathyroid hormone (IOPTH) assay in parathyroid surgery, it was recommended to reserve the use of IOPTH assays for select cases due to its high cost (EUR 170 for five rapid IOPTH assays vs. EUR 125 for five delayed PTH assays, plus added OR time amounting to EUR 15/min in the authors' institution) [78].

The NIR fluorescence equipment cost is amortized more rapidly when it is used by several disciplines within a hospital environment, as abovementioned NIRAF use then generates no additional cost, and a 25 mg bottle of ICG costs around USD 80 and could, in appropriate sterile conditions, be used for 3 to 5 patients in the same day.

The majority of the analyzed articles appear to agree that NIRAF imaging allows a better identification of PGs but that its impact on autotransplantation rate is difficult to determine. Notably, some authors state that NIRAF imaging allows the identification and autotransplantation of accidentally resected glands (e.g., found on the specimen or in central neck dissection) and that the better early identification of PGs in an expert surgeon's hands will be able to spare these PGs from vascular disruption, thereby reducing the need for autotransplantation [45,70].

ICG angiography is commonly accepted to assist in decision-making on autotransplantation. However, it appears to lead to an increase in the autotransplantation rate, and some authors have mentioned the risk of unnecessary parathyroid autotransplantation on the basis of low-flow ICG patterns [48]. They suggested limiting autotransplantation only to clearly devascularized glands that cannot be preserved.

Moreover, metastatic lymph nodes from papillary thyroid cancer are known false positives presenting an high autofluorescence pattern similar to the one of PGs [55], and reimplantation of an autofluorescent nodule found in the central neck dissection should be done with caution if not clearly a PG or confirmed with an intraoperative frozen biopsy analysis [79].
