Fluorescent Imaging in Visceral Surgery: Current Opportunities and Future Perspectives

Indocyanine green (ICG) was first used in the medical context in the 1950s for the cardiological diagnostics field [1]. ICG is a water-soluble dye and can be administered via a peripheral or central vein catheter. After administration, about 98% of the dye binds to plasma proteins, mainly to albumin. ICG is biliary excreted after hepatocellular uptake. Since the half-life of ICG is only about 3–4 min, repetitive measurements are easily possible. For many years, ICG was primarily used by ophthalmologists to analyze retinal perfusion of the eye before it experienced a renaissance [2]. But even after extended use, significant side effects have been rarely described [3].

Recently, a wide variety of systems have been available, ranging from systems for open operations, laparoscopic approaches, and, most recently, for robotic surgery. This allowed the technology to be tested in extensive areas of visceral surgery ranging from perfusion evaluation of colorectal anastomosis to fluorescence-guided lymphadenectomy in esophageal or gastric surgery.

But at the same time, the problem lies in the large variety of systems that make individual studies not comparable. Additionally, the administered dosage of ICG and its time point are essential and differ between the various approaches. Moreover, the technology is mainly used subjectively since only a few objective evaluation parameters are available.

The following sections should give an overview of current opportunities for fluorescent imaging using indocyanine green (FI-ICG) in different areas of visceral surgery. Furthermore, possible future perspectives are to be examined.

The incidence of esophageal cancer continues to rise, so esophagectomy is still a common operation as part of a multimodal therapy concept [4]. Irrespective of medical progress, anastomotic leakage after esophagectomy and reconstruction using a gastric tube is a frequent and life-threatening complication [5]. Since impaired microperfusion represents an independent risk factor for anastomotic leakage, mostly due to the vascular supply of the gastric tube, intraoperative assessment of gastric tube perfusion is essential.

In recent years, the FI-ICG has proven to be a promising tool for intraoperative perfusion assessment of the gastric tube after esophagectomy in several heterogeneous studies. With encouraging results, three literature reviews with meta-analyses have demonstrated a significant reduction of anastomotic leakage by up to 70 percent using FI-ICG for perfusion assessment [6, 8].

There are several different recommendations regarding the dosage and timing of performing the FI-ICG. To prevent overexposure leading to the unusability of the angiography, the applied dose should not be too high. We recommend 0.1 mg/kg body weight dosage administered as a bolus.

When subjectively evaluating gastric tube perfusion, several studies describe areas with good perfusion and those with impaired perfusion, suggesting placing the anastomosis in well-perfused areas as it is associated with a reduction of anastomotic leakage. In 2015, Zehetner et al. [9] published promising results with an anastomotic leakage rate of 2% if the placement of the anastomosis in the areas with good perfusion was possible. Karampinis et al. [10] could support their results with a low leakage rate of 3%. Figure 1 shows the monochromatic fluorescent image of a gastric tube with a demarcation line between well-perfused areas and those with impaired perfusion.

In a recently published clinical study, we combined subjective evaluation of the gastric tube and its marking of the demarcation line with the quantification of FI-ICG. Previously, we were able to validate the quantification of the FI-ICG in animal experiments. The most minor differences in perfusion could be adequately displayed [11, 13]. In the clinical setting, a perfusion reduction of 32% was not associated with leakage. In addition, we could confirm the previously published data regarding the placement of anastomosis in well-perfused areas, which resulted in no occurrence of anastomotic leakage in our cohort [14].

Due to the diverse therapeutic approaches, colorectal surgery offers many possible applications of FI-ICG, even in the advanced stages of the disease. In addition to perfusion assessment, which represents the main application, FI-ICG can be used in locally advanced or metastatic colorectal cancer.

As in esophageal surgery, anastomotic leakage is a frequently observed complication in colorectal surgery that needs to be addressed. However, the published data regarding FI-ICG in colorectal surgery are more solid. The first case-controlled study was already published in 2010 by Kudszus et al. [15]. The authors included 201 patients undergoing colorectal resection with intraoperative FI-ICG between 2003 and 2008. As a control group, they matched 201 patients using their cohort between 1998 and 2003. In a subgroup analysis of all elective operations, they demonstrated a significant reduction of anastomotic leakage needing operative revision (3.3 vs. 7.7%, p = 0.04). Several retrospective single-center studies could demonstrate similar results with a reduction in anastomotic leakage rate, strengthening the evidence [16, 19]. Moreover, in a meta-analysis including 1,302 patients undergoing colorectal resection, the rate of anastomotic leakage was especially reduced in the rectal surgery subgroup. The odds ratio for leakage was reduced by 81% [20].

In addition, several prospective studies have been published in recent years. In 2015, Jafari et al. [21] published the results of the single-arm, prospective, multicenter PILLAR II trial. Only two cases of leakage (1.4%) occurred in 139 patients undergoing laparoscopic left-side or anterior colorectal resection. Based on intraoperative FI-ICG, the location of the anastomosis was changed due to impaired perfusion, resulting in no case of insufficiency. Another multicenter, prospective phase 2 trial published by Ris et al. [22] including 504 patients reported an anastomotic leakage rate of 2.4%.

In 2020, de Nardi et al. [23] published the results of a multicenter, prospective, and randomized trial. A total of 240 patients were included and randomized into an intervention group with intraoperative FI-ICG and a control group with subjective perfusion evaluation. Anastomotic leakage occurred in 5% and 9%, respectively. An additional resection due to a perfusion disorder detected by FI-ICG was performed in 11%. Lately, Jafari et al. [24] published the results of the PILLAR III study, which could not confirm the expected results. This was mainly due to the underpowering of the study cohort with 1,000 planned but only 347 included patients.

The dosage differs between the studies, while the application is uniform. All authors describe an administration as a bolus immediately before perfusion evaluation. Regarding the dosage, some authors use fixed dosages between 3.75 and 12.5 mg [16, 17, 21, 22], while others prefer weight-adjusted dosages ranging from 0.1 to 0.3 mg/kg body weight [15, 18, 19, 23]. The decisive factor is a continuous, uniform approach so that an FI-ICG can be interpreted repetitively. Our department uses a standardized dosage of 0.1 mg/kg body weight administered as a bolus before perfusion evaluation (Table 1). To avoid overexposure, the dosage should not be chosen too high. An example of colorectal anastomosis is shown in Figure 2.

Cytoreductive surgery, mostly in combination with hyperthermic intraperitoneal chemotherapy, remains a therapeutic option in patients with peritoneal carcinomatosis from colorectal cancer, resulting in prolonged overall survival [25]. To increase the detection rate of peritoneal lesions, FI-ICG has been investigated in some small patient collectives with encouraging results. Liberale et al. [26] used 0.25 mg/kg ICG 24 h preoperatively for intraoperative visualization. They could detect several additional lesions due to their hyper- or isofluorescent pattern. Likewise, Lieto et al. [27] identified 16 additional lesions using the same dosage but applied 60 min preoperatively. It is important to note that subsequent ICG administration after the start of the resection is difficult to assess due to contamination of the operation site.

Due to the primary biliary excretion of ICG, there are special approaches, which can only be performed in hepatobiliary surgery. Already in 2009, Ishizawa and colleagues investigated and published a classification of the fluorescence pattern of the hepatic lesion after injection of ICG. They identified three patterns: a homogenous hyperfluorescent pattern, a partial fluorescent type with inhomogeneous fluorescence, and a rim fluorescent type. In addition, the different entities of the lesion were associated with varying patterns of fluorescence. Colorectal liver metastasis appeared as the rim type, whereas the appearance of hepatocellular carcinomas depends on the differentiation of the tumor [28]. This is mainly due to the previously described pharmacological kinetics of ICG. Dysfunction of hepatocellular carriers leads to delayed excretion of ICG, resulting in persisting fluorescence despite the short half-life. In addition, the compression of surrounding liver tissue caused by the tumor causes adjacent cholestasis. Due to the biliary excretion of ICG, this local cholestasis appears as a rim of fluorescence [2, 28]. Regarding the dosage and timing of preoperative injection of ICG, there is no concrete recommendation available since the available literature varies from 24 h to 14 days before surgery and from 0.1 mg/kg body weight to 50 mg ICG, respectively [29, 30].

Next to tumor identification, FI-ICG can be used for fluorescence-guided surgery in an overlaid or combined white light mode, which is now available in some devices. Tumors can be detected, and resection margins can be defined.

The main advantage of intraoperative FI-ICG when resecting liver metastasis is the detection of additional lesions. Even lesions smaller than 10 mm could be detected. However, the detection depth is limited to 8–10 mm which is a limitation [29]. Therefore, FI-ICG should always be combined with intraoperative ultrasound. In addition, two studies could show a higher rate of R0 resection using fluorescence-guided surgery. The rim could be used as the resection margin [31, 32]. In addition, the fluorescent rim could be confirmed as a valid resection margin using fluorescence microscopy without malignant cells remaining [33].

Currently, the use of FI-ICG for assessing perfusion is mostly subjective. Thus, the interpretation of fluorescence angiography is not reliable and reproducible. To address this issue, different approaches to quantifying FI-ICG have been investigated. Kumagai et al. [34] proposed the 90-s rule defined as a homogenous fluorescent pattern 90 s after injection of ICG. Placement of the anastomosis in these areas was associated with lower leakage rates. With a time of 98 s after injection which was associated with anastomotic leakage, Slooter et al. [35] showed similar results.

Despite the measurement of time to the beginning of fluorescence, other approaches using a time-dependent fluorescence intensity curve have proven suitable for the quantification of ICG. From the fluorescence intensity curve, several parameters can be calculated. de Groot et al. [36] combined the time measurement with an intensity curve in robotic esophagectomy. Detter et al. [37] described the quantification of FI-ICG using the slope of fluorescence intensity (SFI) and the background-subtracted fluorescence intensity. Using these parameters and adding the time to slope defined as the first fluorescence signal after injection of ICG, we transferred this approach to the visceral microperfusion quantification. In several animal experiments, we could validate perfusion measurement with FI-ICG quantified by SFI, background-subtracted fluorescence intensity, and time to slope [11, 13]. Figure 3 shows a schematic representation of the fluorescence intensity curve and the calculated parameters.

Finally, the experimental animal data could be used in the human gastric tube. Quantification was feasible, and anastomotic leakage could be sufficiently predicted. In addition, cutoff values could be established. A perfusion reduction of up to 32% using SFI was not associated with anastomotic leakage [14].

Presently, there is a lack of valid data regarding the quantification of FI-ICG regarding perfusion evaluation. Further studies are necessary to exploit the full potential of this technology in this regard and to generate reference parameters.

Marking and thereby detecting sentinel lymph nodes represents another possible application of FI-ICG. This technique is already established for endometrial carcinoma and is included in the German guideline [38]. In several feasibility studies, FI-ICG has proven suitability for intraoperative lymph node targeting and mapping in colorectal, esophageal, and gastric surgery [39, 42]. Regarding dosage and administration, there are a variety of techniques. Most studies describe an endoscopic approach with peritumoral, submucosal, or mucosal injection up to 24 h before surgery using a dosage between 0.5 and 2 mL of diluted ICG solution. However, the clinical benefit of imaging the lymph nodes regarding the oncological outcome remains unclear.

Lately, few innovative approaches for tumor targeting have been described using monoclonal antibodies bound to ICG. CEA was used as a target structure in colorectal cancer in one case series. Using the modified fluorescent dye, metastasis could be detected [43]. In a similar approach, VGEF was used as a target structure using bevacizumab as the conjugate to the fluorescent dye [44]. Modified fluorescent dyes could play an important role in navigated surgery in the future.

In summary, FI-ICG is a promising tool for visceral surgeons, especially regarding intraoperative perfusion evaluation. Despite the mostly subjective use of the technology, important complications have already been addressed in prospective studies. Since the FI-ICG is inexpensive and does not significantly increase the operation time, the technology should also be used in macroscopically unimpaired areas since FI-ICG can unmask limited perfusion in some cases. Based on the existing evidence, the routine use of FI-ICG should be extended in clinical practice. In addition, the objective areas of application such as quantification or navigated operations are encouraging. Since the existing literature is quite heterogeneous, further standardization and prospective studies investigating FI-ICG in different fields of use are needed.

The authors have nothing to declare.

This work received no funding.

Philipp H. von Kroge and Anna Duprée jointly wrote the manuscript and approved the final version.