Postoperative hypoparathyroidism remains an unpleasant complication for patients undergoing total thyroidectomy. Technical tools to assist the intraoperative identification of PGs are limited. Binocular loupes rely on the surgeon’s experience and do not provide additional imaging information [
12], which can be acquired using fluorescence-based imaging techniques. Autofluorescence-guided intra-operative visualization of PGs is a groundbreaking technique shown to achieve excellent detection rates of both healthy and diseased PGs [
6]. As no tracer is required, this real-time, repeatable method appears ideal for use in thyroid surgery to prevent inadvertent parathyroid injury or removal. In this randomized controlled trial, we examined the utility of NIR-AF in preventing postoperative hypoparathyroidism after total thyroidectomy.
Parathyroid detection
This study did not show a higher mean detection rate of the PGs using NIR-AF but fewer patients with less than 2 detected PGs (Table
1). One single PG was detected by NIR-AF rather than visualization by the surgeon. In contrast, other authors demonstrated an increased number of PGs detected with NIR-AF [
11,
13]. A known problem is the inability of NIR light to penetrate tissues beyond a depth of a few millimeters. As PGs are often covered by adipose tissue or blood, they elude stimulation by NIR light; however, Kahramangil et al. observed that 37–67% of all PGs could be recognized using NIR-AF before detection by the surgeons, even if they were covered by tissue [
14]. Although a study from Cleveland, Ohio, demonstrated that once the overlying tissues are removed, NIR-AF identified PGs with a sensitivity of 98.5% [
15], these glands are also likely to be identified by an experienced surgeon. Thus, this technique is more a real-time confirmation than a “navigation system” to the PGs. Although we were not able to detect an increased number of PGs by NIR-AF compared to visual identification alone, as surgeons we have all previously been confronted with the uncertainty as to whether a nodule corresponds to a PG and could have benefited from NIR-AF.
The technical system we used was image-based, and therefore, AF intensity was not quantified. As PGs display different AF intensities, influenced by external factors such as stray light, experience in interpreting the images is necessary. Spectroscopy-based systems may support less experienced surgeons by providing quantitative information, but these require physical contact with the presumed PG [
16].
Postoperative outcome
As shown in Table
2, we could not detect a significant difference in the occurrence of hypoparathyroidism at skin closure or POD1. Using NIR-AF, incidence reduced by 23.6% at skin closure, and by 13.3% on POD1. Postoperative parathyroid hormone and calcium blood levels were also not significantly different. Likewise, there was no difference in hypocalcemia-associated symptoms, medication at discharge, or length of hospital stay.
A key factor in preserving PG vascularization is careful vessel preparation along the thyroid capsule. Our intention was to identify PGs as early as possible during the operation. The Karl Storz® system used throughout the whole study requires partial dissection of the covering tissue to achieve sufficient AF intensity, as its ability to penetrate tissue is low. Especially in fatty conditions, this may increase the risk of accidental devascularization prior to identification. Laser-based systems are reported to have higher tissue penetrance [
17,
18]. Benmiloud et al. reported a significant reduction in postoperative hypocalcemia using an integrated laser light-based system (Fluobeam® 800, Fluoptics, Grenoble, France). They started their trial just after the recruiting period of this study. With the mentioned system, significantly more PGs could be detected by NIR-AF [
11]. An enhanced tissue penetrance may contribute to establishing NIR-AF as a real-time “mapping tool” for PGs in the early stages of thyroid surgery.
In both this and previous studies, we observed persistent AF activity in excised PGs. DiMarco et al. scanned thyroid specimens in 106 patients with an image-based system (Fluobeam® 800, Fluoptics, Grenoble, France) to detect inadvertently removed PGs but could find no difference in the number of PGs, early hypocalcemia or late hypoparathyroidism compared to 163 controls [
18]. Therefore, the benefit of NIR-AF in the late operative phase remains unclear.
AF activity persists after PGs devascularization which compromises the ability of NIR-AF to serve as a decision tool as to whether to replant a PG. ICG angiography can facilitate this decision but requires an intravenous dye [
19]. It can be surmised that NIF-AF-based identification together with ICG-based angiography has a high potential to preserve PGs. While a combined NIR-AF and ICG system with FDA and CE clearance already exists (Fluobeam LX®, Fluoptics, Grenoble, France), comparative studies are still lacking to prove its effectiveness.
In addition, one should not underestimate the learning curve both in visual and NIR-AF-based PG identification. We experienced a wide range of AF intensity depending on various conditions such as background lights, distance between the camera and tissue, and tissue penetrance. In our opinion, novice surgeons should undergo a training program before relying on NIR-AF.
Limitations and future prospects
Due to the single-blinded study design, it is difficult to estimate how much the observer effect contributed to the results. In both the intervention and control groups, the participating surgeons reported a training effect in recognizing PGs. Thus, NIR could serve as a practice tool for both novice and experienced surgeons.
The participating six surgeons were not randomized to the two groups. Based on their experience (>30 thyroid procedures per year, >200 procedures in total) and a standardized operation technique, we assumed a low influence on the outcome.
The replantation rate of 27–37% in this study was high compared to 14.3% in 746 consecutive thyroidectomies prior to this trial. One explanation could be a forced search of the parathyroid glands by the surgeon (observer effect). Due to the system’s low tissue penetrance, this may have increased the risk of inadvertent devascularization.
The image system used was developed for ICG fluoroscopy but was modified with additional filters to be able to visualize NIR-AF. To ensure comparability of all study patients, we used the existing system and made no technical changes. This could be a disadvantage compared to systems originally designed for NIR-AF-based PG detection like the Fluobeam system (Fluoptics®, Grenoble, France). One advantage of the Storz® system is that it can also be used as a laparoscopic unit. Investment costs could thus be put into perspective.
The main limitation of this study is the small sample size based on the assumption of a 25% reduction in the incidence of postoperative hypoparathyroidism. A larger case series based on the effect demonstrated in our study could have substantially increased the power of this work.
The focus of our studies was on transient postoperative hypoparathyroidism. Therefore, no follow-up was performed. Since permanent hypoparathyroidism in particular represents a considerable burden for patients and is cost-intensive, a follow-up in a larger number of cases would be interesting.