A practical guide for the use of indocyanine green and methylene blue in fluorescence-guided abdominal surgery.

Near-infrared (NIR) fluorescence imaging is gaining clinical acceptance over the last years and has been used for detection of lymph nodes, several tumor types, vital structures and tissue perfusion. This review focuses on NIR fluorescence imaging with indocyanine green and methylene blue for different clinical applications in abdominal surgery with an emphasis on oncology, based on a systematic literature search. Furthermore, practical information on doses, injection times, and intraoperative use are provided.

INTRODUCTION

Cancer is one of the leading causes of death worldwide, resulting in 8.2 million deaths annually.1 Treatments for cancer include surgery, radiation therapy, chemotherapy, and targeted therapies. Surgery is often the cornerstone in the treatment of solid cancers in the abdominal cavity. Especially during oncologic surgery, it is pivotal to remove as much tumor as possible and simultaneously to prevent unnecessary damage to surrounding healthy tissue. Despite that the quality of preoperative imaging modalities, such as MRI and CT, has been improved during the last decades, intraoperative navigation using CT or MRI is only performed in specialized hospitals, mostly confined to the field of neurosurgery, due to the complexity, high costs, tissue deformation, and long acquisition times.2 Today, for intraoperative navigation, the surgeon has to rely in the majority of the cases only on visual and tactile feedback to distinguish between different kind of tissue structures. Additionally, (laparoscopic) intraoperative ultrasound or sometimes a gamma probe can be used during oncologic surgery. However, this is not always sufficient. Tumor‐positive (R1 or R2) resection margins in colorectal cancer, for example, occur in approximately 10% of the operated patients.3, 4 Together with the increasing rate of minimal invasive (laparoscopic and robotic) surgery and therefore the lack of tactile feedback, there is a demand of improving visibility of different kind of tissue types, especially for distinguishing malignant and benign structures.

In the past years, new intraoperative imaging systems that exploit the near‐infrared (NIR) light spectrum, have been evaluated (pre)clinically for different applications.5, 6, 7, 8 For intraoperative purposes, NIR light (700‐900 nm) is more advantageous than visible light due to its capability to penetrate deeper into tissue, up to 10 mm. In addition, in the NIR spectrum tissue exhibits almost no autofluorescence. NIR fluorescent contrast agents lead therefore to maximized signal‐to‐background ratios and enhance the contrast of different tissue types.9, 10 NIR fluorescence imaging requires a NIR fluorescent agent (ie, fluorophore) combined with an imaging system that is able to both excite and detect the fluorophore. The fluorescence signal can be visualized after milliseconds, which is advantageous above other emerging imaging techniques, such as Raman spectroscopy or Optical Coherence Tomography, which require more time to visualize the same field of view.11 Moreover, most systems are capable of merging fluorescence signals with normal RGB color videos, allowing direct anatomical orientation.

Different fluorophores have been evaluated (pre)clinically in NIR fluorescence imaging. These fluorophores have to be injected locally or systematically before surgery. Today, only two of them, indocyanine green (ICG) and methylene blue (MB), are approved for clinical use by the Food and Drug Administration and the European Medicines Agency. 5‐aminolevulinic acid (5‐ALA) is also clinically used as a fluorophore, especially in the field of neurosurgery field. However, this dye is fluorescent at 510 nm, which is outside the NIR spectrum.12 Consequently, the penetration capacity of 5‐ALA is negligible. ICG and MB are nonspecific contrast agents in the near‐infrared region and they do not bind to tumor ligands. ICG has a half‐time of 150–180 s and is solely cleared by the liver.13 Its excitation peaks around 800 nm. ICG is very safe; adverse events are reported in less than 1 in 40 000 patients and mostly comprise hypersensitivity reactions.14 MB in low doses (<2 mg/kg) is safe, however it can induce severe adverse effects such as arrhythmias, coronary vasoconstriction, and hemolytic anemia in patients with renal insufficiency or after administration of higher doses.15 MB is partially renal cleared and has an excitation peak of approximately 700 nm. Therefore it has a less tissue penetration capacity and background tissue shows more autofluorescence. Due to their differences in clearance patterns ICG and MB have been used for various applications. In short, the indications can be used for lymph node mapping, detection of vital structures (eg, ureters, bile ducts), identification of tumors, and assessment of tissue perfusion. This review provides a practical approach for NIR fluorescence imaging with ICG and MB for different clinical applications in abdominal surgery, with an emphasis on oncology.

METHODS

We searched in PubMed for in human (in vivo) trials using ICG and MB in NIR fluorescence guided abdominal surgery, published before January 2018. The search was based on the following search items: “Indocyanine green” or “ICG,” “Methylene Blue” or “MB,” “Near‐infrared” or “NIR,” and “Fluorescence” or “fluorescent.” More specific search terms were added per application (SLN mapping, tumor detection, tissue perfusion, detection of vital structures). Only articles written in English were included. Case reports and (systematic) reviews were excluded from the analysis. After selection of relevant articles, clinically relevant information focussing on how to perform NIR fluorescence imaging in daily practice was extracted from the articles to set‐up a review. In this article, the different applications of NIR fluorescence in abdominal surgery, SLN mapping, tumor detection, visualization of vital structures and tissue perfusion, are discussed separately.

RESULTS

Our search revealed a total of 891 articles, which are reported in a flow diagram (Figure 1). After exclusion of 720 articles, who did not meet our eligibility criteria, a total of 171 original articles remained, which formed the basis of this review.

Figure 1

Flow‐chart of the literature search strategy

SENTINEL LYMPH NODE MAPPING

Sentinel lymph node mapping is only part of standard care in melanoma and breast cancer.16, 17 Current standard‐of‐care is a combination of blue dye and gamma probe for Technetium‐99m (Tc‐99m) detection, which has certain disadvantages. Blue dye has no penetration capacity whatsoever and a gamma probe does not allow visual identification. SLN mapping is not part of standard clinical care in abdominal oncological surgery.18, 19. Today, NIR fluorescence imaging of SLN(s) as a new and additional modality was evaluated for SLN mapping in malignancies of the oesophagus, stomach, colon, bladder, prostate, cervix, endometrium, and ovarium.

Clinical applications

Esophageal cancer

Esophageal cancer is a disease with a dismal prognosis, showing an overall 5‐year survival rate around 14%.20 Extensive lymphadenectomy improves the prognosis, especially at early stages.21, 22, 23 To visualize the sentinel lymph nodes, ICG was evaluated as a NIR fluorescent marker in three clinical studies. All studies were feasibility studies and therefore using a small study population: 1‐20 patients.24, 25, 26 Ninety‐five percent of the sentinel lymph nodes could be detected in the cohort of 20 patients, as shown by Yuasa et al.25 However, further research is necessary before introduction into clinical practice.

Gastric cancer

Gastric cancer is one of the most common cancers in the world, especially in Eastern Asia with an incidence of 24 and 9.8 per 100 000 in men and women, respectively. We found ten clinical studies using ICG during NIR fluorescence imaging during laparoscopic or open gastric surgery.27, 28, 29, 30, 31, 32, 33, 34, 35, 36 The SLN identification rates ranges from 90% to 100%, however it decreases to 0% for T4 tumors.35 It also has to be mentioned that most studies did not include these higher stages in their study, which could explain the high resection rates. In Japan and Korea, T1N0M0 tumors are standard treated with endoscopic resection, which creates new opportunities for clinical use of NIRF for SLN mapping in selected cases, as demonstrated by Bok et al.,37 who showed that a combination of endoscopic submucosal dissection and sentinel node navigation surgery (laparoscopic) was feasible in 12 out of 13 patients.

Colorectal cancer

SLN mapping in colorectal cancer is still a controversial procedure, due to the relatively low visibility of the lymphatic system after injection of blue dye.38 In the past years NIR fluorescent imaging with other available dyes, such as ICG, was used in nine clinical studies during surgery.29, 39, 40, 41, 42, 43, 44, 45, 46 All studies were performed in a small population, ranges from 5 to 30 patients, who underwent laparoscopic or open surgery due to colorectal cancer. The SLN detection rates by NIR fluorescence imaging vary between 65.5% and 100%, which could probably explained by the experience in the participating centers.

Bladder cancer

A cystectomy with pelvic lymphadenectomy is the treatment of choice for patients with bladder cancer. However, accurate intraoperative detection of SLNs is difficult and lacks sensitivity.47 Moreover, there is a great inter‐patient variability in lymph drainage patterns, which makes the location of the SLN unpredictable. Recently, NIR fluorescence imaging using ICG could detect a sentinel lymph node in at least 90%, which indicated that this technique could contribute to improved intraoperative staging of bladder cancer.48, 49, 50

Prostate cancer

Prostate cancer is one of the most common cancers in men and treatment consists of diverse interventions, including surgical resection.51 (Extended) lymph node dissection is sometimes performed during radical prostatectomy, depending on the preoperative risk assessment.52, 53 Ten studies evaluated the use of ICG during prostatectomy, which showed a high detection rate of SLNs, however due to the low prevalence of SLN metastases a low positive predictive value was found.54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64 This supports the findings that NIR fluorescence imaging could be useful in high risk groups.

Cervical cancer

Early stage cervical cancer, which is often detected by screening, is normally treated with radical hysterectomy and pelvic lymph node dissection.65 Lymph node status is one of the independent prognostic factors for the patient's survival.66 However, if the resected lymph nodes were not involved, which is often the case in early stage cervical cancer, patients underwent unnecessarily lymphadenectomy. This could be prevented by performing SLN mapping, for instance with NIR fluorescence imaging, as shown in different studies.67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83 These studies, performed in early stage cervical cancer, showed detection rates varying from 60% to 100%. Recently, three studies compared ICG to the current golden standard for SLN mapping, blue dye and Tc‐99m, in study populations varying from 58 to 144 patients.73, 74, 76 It was concluded, that detection rates did not differ significantly between blue dye with Tc‐99m and ICG. However, all studies showed significantly higher rates of bilateral lymph node detection, especially in larger tumor sizes (>2 cm). Hence, it indicates that intraoperative lymph node mapping with fluorescence imaging using ICG is at least comparable to the current golden standard.

Endometrial cancer

Endometrial cancer is one of the most common cancers in women and patients often presents with an early stage disease.51, 84 Currently, there is no agreement about performing SLN mapping to avoid extensive lymphadenectomy during surgery in early stage endometrial cancer.19 However, in some centers fluorescence imaging is used for SLN mapping and incorporated in standard care.85 Nineteen studies showed that fluorescence imaging is useful, indicating high detection rates of lymph nodes, varying from 68 to 100%.69, 71, 72, 73, 78, 80, 83, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97 Laios et al69 showed the lowest detection rate (68%), which was probably caused by the learning process as visualized by a learning curve. Rossi et al87 showed that cervical stromal injection of ICG is preferable over hysteroscopic endometrial injection for SLN detection, as illustrated by the significant difference in detection rates (82% vs 33% respectively, P = 0.027).87 These results indicate that SLN mapping with NIR fluorescence imaging is feasible and may be worth considering for use in clinical practice.

Ovarian cancer

One study evaluated the implementation of NIR fluorescence imaging of SLNs during laparoscopic surgery in seven patients with early‐stage ovarian cancer.98 In all patients, at least one SLN was detected after intraoperatively injection with ICG in both sides of the proper ovarian and suspensory ligament. Nevertheless, more studies have to be performed to determine the added value of fluorescence imaging for intraoperative lymph node staging, which is essential before implementation into clinical practice.

Summary of findings

Sentinel lymph node mapping using ICG has been evaluated for several applications in abdominal surgery (gastroenterology, urology, and gynecology) with different imaging systems.24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 39, 40, 41, 42, 43, 44, 45, 46, 48, 49, 50, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98 In these studies a high diversity in doses has been used, varying from 100 μg to 25 mg. In general, based on the evaluated clinical studies a dose of (at least) 2.5 mg is recommended for good visualization of sentinel lymph nodes.

It was shown, that ICG bound to human serum albumin (ICG:HSA) could also be used for SLN mapping.26, 79 Moreover, Hachey et al26 showed that ICG:HSA is preferable above using ICG alone in esophageal cancer, which could probably be explained by the rapid lymphatic clearance and poor retention of ICG alone. However, due to the limited evidence we prefer an injection of ICG alone. As part of standard care, which consists of peritumoral blue dye injection, ICG could also be injected around the tumor. Most of the studies used a four quadrant injection in the submucosa around the tumor. The timing of the injection is also an important issue. One study in gastric cancer prefers injection (endoscopically) 1 day before surgery.36 They showed lower false negative results than intraoperatively subserosal injection, probably because of frequent leakage from damaged lymphatic vessels, caused by mobilising the stomach. On the other hand, most studies showed good results by performing injection, just after applying general anesthesia.25, 29, 31, 33, 41, 46, 48, 49, 50, 54, 57, 68, 69, 70, 71, 72, 73, 74, 76, 78, 80, 81, 83, 87, 88, 89, 92, 93, 95, 96 For good visibility of the tumor and consequently a good injection place, a minimal‐invasive (endoscopic, cystoscopic) approach is recommended, which also prevents damaging of lymphatic vessels. For cervical and endometrial cancer a cervical stromal injection of ICG is preferable above a hysteroscopic injection.87 As shown in Figure 2, approximately 15–30 min after injection, the sentinel lymph nodes will be visualized.33, 39, 43, 50, 62, 89 In the mean‐time the surgical procedure can be started, and the surgeon will therefore lose minimal time. After identification of the SLN(s) the additional oncological resection can be performed.

Figure 2

Example of lymph node mapping in bladder cancer during surgery and ex vivo. Upper panel: arrowheads indicate the NIR fluorescent lymph nodes along the left external iliac vein during surgery. Lower panel: fluorescent lymph nodes after excision. Reprinted by permission from John Wiley and Sons: Journal of Surgical Oncology49 © 2014

How to use it

Taken together, optimal use of ICG for SLN mapping could be achieved by peritumorally injection of 2.5 mg ICG. Intraoperative injection, just after applying anesthesia, in a minimal invasive manner seems optimal. The SLNs could be visualized 15–30 min after injection (Table 1).

Table 1

Summary of known clinical applications of near‐infrared fluorescence imaging in abdominal surgery and recommendations for intraoperative use

Application	Tissue type	Imaging system used in different studies	Preferred contrast agent	Recommended dose (mg)	Preferred injection site	Timing of injection	Time to visualize structures (min)
SLN mapping
	Esophagus	Laparoscopic	Indocyanine green	2.5	Endoscopic: 4 quadrants injection	Just before surgery	15‐30
	Gastric	Laparoscopic, open	Indocyanine green	2.5	Endoscopic: 4 quadrants injection	Just before surgery	15‐30
	Colorectal	Laparoscopic, open	Indocyanine green	2.5	Subserosal: 4 quadrants injection	Intraoperatively	15‐30
	Bladder	Robotic, open	Indocyanine green	2.5	Cystoscopic: bladder mucosa	Intraoperatively	15‐30
	Prostate	Robotic, laparoscopic	Indocyanine green	2.5	Transrectal: under US guidance into prostate lobes	Intraoperatively	15‐30
	Cervix	Laparoscopic, robotic	Indocyanine green	2.5	Transvaginal: submucosa cervix 4 quadrants injection	Intraoperatively	15‐30
	Endometrium	Laparoscopic, robotic	Indocyanine green	2.5	Transvaginal: submucosa cervix 4 quadrants injection	Intraoperatively	15‐30
	Ovarium	Laparoscopic	Indocyanine green	2.5	Laparoscopic: dorsal and ventral side of proper ovarian and the suspensory ligament	Intraoperatively	15‐30
Tumor imaging
	Liver	Open, laparoscopic	Indocyanine green	10	Intravenously	24 h before surgery	Directly
	Adrenal	Laparoscopic, robotic	Indocyanine green	2.5	Intravenously	Intraoperatively	1
	Peritoneal metastases	Open	Indocyanine green	0.25 (mg/kg)	Intravenously	Intraoperatively	>5
Vital structures
	Bile duct	Laparoscopic	Indocyanine green	5	Intravenously	3‐7 h before surgery	Directly
					Into gallbladder	Intraoperatively	Directly
	Ureter	Open, laparoscopic, robotic	Methylene blue	0.25 (mg/kg)	Intravenously	Intraoperatively	>10
Perfusion
	Esophagogastric anastomoses	Laparoscopic, robotic	Indocyanine green	2.5	Intravenously	After mobilization and selecting of anastomotic site	1
	Colorectal anastomoses	Open, laparoscopic, robotic	Indocyanine green	2.5	Intravenously	After mobilization and selecting of anastomotic site	1
	Liver segments	Open	Indocyanine green	2.5	Portal vein	During surgery	2
		Laparoscopic	Indocyanine green	2.5	Intravenously	During surgery	2

TUMOR IMAGING

ICG has been used as a contrast agent for intraoperative detection of diverse tumor types, which is based on two characteristics of ICG: hepatic clearance and the enhanced permeability and retention (EPR) effect. Solid tumors (other than hepatobiliary tract) could be visualized by the EPR effect, which is based on the visualization with ICG of the increased permeability and reduced drainage in tumor tissue after tumor‐induced angiogenesis.99, 100

Liver tumors (primary & secondary)

Due to its hepatic clearance, ICG can be used for imaging of hepatobiliary tumors, both metastases and primary liver cancers (cholangiocarcinoma and hepatocellular carcinoma). Healthy liver tissue clears ICG within a couple of hours, whereas tumor tissue retains ICG by compression of bile ducts. However, recent studies suggested also a molecular mechanism.101, 102 Immature hepatocytes, which are located in the transition zone between the tumor and normal liver parenchyma, were not able to excrete ICG into the biliary ducts due to down‐regulation of anion transporters. This results in a fluorescent rim around the tumor or metastasis, making it easy to identify them. Furthermore, ICG accumulates into well‐differentiated hepatocellular carcinoma (HCC), resulting in complete fluorescence signal rather than a dark center. In a retrospective cohort study in our institute it was shown that NIR fluorescence imaging of colorectal liver metastases was able to identify more and smaller tumors, resulting in reduced recurrences in a subset of patients.103 Twenty‐one studies evaluated the use of NIR fluorescence imaging with ICG for liver tumor detection, both primary tumors as metastatic spread of colorectal, uveal, breast, and pancreatic cancer.101, 102, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122

Summary of findings

To establish the optimal dose and injection time our group performed a comparison study with 10 and 20 mg ICG injection. However, no significant difference in optimal tumor‐to‐background contrast between respectively a 24 or 48 h administration and 10 or 20 mg injection was found.122 Therefore, we recommend a bolus of 10 mg of injected intravenously 24 h before surgery based on clinical and logistical preferences. Tumor tissue compresses small bile ducts and consequently causes local inflammation. In these bile ducts and inflamed liver tissue ICG is retained for some days. Therefore, other groups showed also good results with ICG injection up to 14 days before surgery, although using much higher doses of ICG, up to 0.5 mg/kg.101, 106, 108, 109, 111, 112, 113, 117, 118, 119, 120 Today, no randomized trial has been performed to compare these high differences in dosing and injection timing. After administering ICG, a fluorescent rim can be identified around tumors during surgery, as illustrated in Figure 3. Several studies showed that this technique not only identifies known tumors, but also occult, otherwise undetectable submillimeter tumors. However, only subcapsular lesions up to around 1 cm from the liver surface are expected to be visible, due to the limited penetration depth of NIR fluorescence.123 Nevertheless, it could therefore be used as a complementary intraoperative tool, as CT/MRI and intraoperative ultrasound still showed relatively high rates of false negatives for superficially located (subcapsular) lesions.124, 125

Figure 3

Example of two colorectal liver metastases detected by NIRF imaging. White arrow: a fluorescent lesion, which was not detected by preoperative imaging. Dashed arrow: a preoperative suspected lesion could be recognised by its characterizing fluorescent rim. Reprinted by permission from Elsevier: European Journal of Surgical Oncology103 © 2017

Peritoneal metastases

Preoperative detection of peritoneal metastases is often difficult with the current image modalities.126 Adequate pre‐ and intra‐operative staging is therefore an important issue for offering the best treatment. In the last years, more evidence is coming up, which showed that hyperthermic intraperitoneal chemotherapy (HIPEC) combined with cytoreductive surgery gave a better survival in patients with limited peritoneal metastases from colorectal or ovarian cancer.127, 128 Four (pilot) studies evaluated the intraoperative visibility of peritoneal spread in liver, colorectal, and ovarian cancer with NIR fluorescence imaging.129, 130, 131, 132 All studies showed a good visibility of peritoneal metastases, however in patients with ovarian cancer treated with HIPEC, distinguishing benign scars and malignant lesions was difficult.131 Moreover, Tummers et al132 showed a high false positive rate, which could be explained by the EPR effect, which underscores the need for tumor‐selective targets. As HCC tissues often retain bile juice productivity even in metastatic sites, fluorescence imaging following preoperative intravenous injection of ICG can also be applied to identification of extrahepatic metastases of HCC.130

All studies started with ICG injection 24 h before surgery. Unfortunately, no fluorescence signal was visible during exploration. However, intraoperative injection was feasible and showed good results. Hence, we recommend intravenously injection of ICG after abdominal exposition. A dose of 0.25 mg/kg of ICG could therefore be used. As shown above, detection of peritoneal metastases could be useful during intraoperative staging and cytoreductive surgery (combined with HIPEC). Clinical application of NIRF imaging during cytoreductive surgery should be reserved for patients with limited peritoneal spread (peritoneal cancer index <8).129, 133 If advanced peritoneal spread has been established preoperatively, the added value of fluorescence imaging with ICG is limited. After injection of ICG, peritoneal nodules become fluorescent from at least 5 min.129 Tumor positive nodules become hyperfluorescent. Due to physiological accumulation of ICG, detection of small peritoneal metastases is hampered in some abdominal regions (such as in the liver and visceral peritoneum).

Adrenal tumors

Adrenal masses are often resected by performing minimal invasive surgery (laparoscopic or robotic).134, 135 Identification of the adrenal glands and determining the resection margins could be difficult, especially when it is surrounded by retroperitoneal fat. The added value of NIR fluorescence imaging using ICG during minimal invasive (partial) adrenalectomy has been determined in four studies.136, 137, 138, 139 Based on the difference in perfusion between the adrenals and the surrounding tissues, the adrenal glands and different types of adrenal tumors could be detected. Colvin et al139 showed that NIR fluorescence imaging was superior and equivalent in 46.5% and 25.6% respectively, to conventional robotic view in determining the borders of the adrenal glands in 40 patients. Other mentioned studies showed similar results in determining the adrenal gland borders, however they were performed in much smaller populations.

Three out of four studies used a 5 mg dose of ICG, which turned out to be the optimal dose for visibility of the adrenal glands.136, 138, 139 After exposure of the retroperitoneum and acquiring a good view on the adrenal glands and its surrounding tissue, ICG could be intravenously injected. One minute after injection of ICG, the fluorescence signal could be detected.139 First of all, the arterial vasculature of the adrenal glands and kidneys became fluorescent. After a few seconds, the fluorescence signal highlights the adrenal parenchyma, however it is slightly less fluorescent as the adjacent kidneys.138 Adrenocortical tumors could be easily recognized by the hyperfluorescent signal, although medullary tumors (such as pheochromocytomas) were hypofluorescent during NIRF imaging.138, 139 In the background, the liver could show more intense fluorescence, in case of right adrenal surgery. The optimal tumor‐to‐background contrast is achieved after 5 min.139 If the fluorescent signal washes out (>20 min after injection), a new injection of 5 mg ICG could be administered.

Other solid tumors

ICG has also been evaluated in other solid tumors in the abdominal region, although the added value has not been proven. Our group studied the role of ICG for intraoperative visualization of pancreatic tumors during pancreaticoduodenectomy in eight patients.140 After injection of ICG, only in one patients a clear tumor‐to‐pancreas contrast was observed. The use NIR fluorescence imaging during nephrectomy has also been a topic of research, as reviewed by Bjurlin et al.141 Based on the angiographic properties of ICG, it could play a role in determining vascular clamping during nephrectomy.142, 143 However, for tumor imaging NIR fluorescence imaging with ICG showed inconsistent results. One study compared the fluorescent signal with histologic findings in 100 patients undergoing partial nephrectomies, and showed that the ICG pattern (hypofluorescent, isofluorescent, or afluorescent) was insufficient to distinguish benign from malignant renal lesions.144

As summarized in Table 1, we recommend 10 mg, 2.5 mg and 0.25 mg/kg ICG injection intravenously for good intraoperative visibility of liver tumors, adrenal tumors and peritoneal metastases, respectively. Intraoperative injection is optimal for detection of peritoneal metastases and adrenal tumors, although for detection of liver tumors it is preferable to inject ICG 24 h before surgery.

VISUALIZATION OF VITAL STRUCTURES

Currently, no good intraoperative imaging techniques are available for both biliary duct and ureter visualization. Intraoperative cholangiography is sometimes used for enhanced bile duct identification. Two disadvantages of this technique are the exposition of radiation to the patients and health care personnel and the chance of causing bile duct injuries by bile duct cannulation.145, 146 Other techniques such as ureteral stenting is sometimes performed to visualize the ureter. However, the benefit is currently questionable and it is also associated with a risk of ureteral injury.147, 148 NIR fluorescence imaging with ICG and MB is used for both biliary and ureter mapping, based on their clearance characteristics.7

Biliary duct

During laparoscopic cholecystectomy, nowadays one of the most performed surgeries, it is important to visualize the so‐called critical view of safety. This ensures that the cystic duct has been identified correctly. Still, due to anatomical variations, the cystic duct could be misidentified by the common hepatic duct or the common bile duct. Misidentification, with a reported incidence ranging from 0.08% to 1.5%, is accomplished with a morbidity up to 21% and high healthcare costs.149, 150, 151, 152 The use of intravenously ICG injection for identification of the bile ducts with NIR fluorescence cholangiography was evaluated in 30 studies, which showed a pooled cystic duct identification in 1019/1050 (97%) patients.140, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181 Although, the biliary ducts could be very well visualized, the diminished visualization of the bile ducts in minority of the patients, could be caused by inflammatory tissue or fatty peritoneum (in obese patients) covering the biliary ducts, which hampered the penetration of NIR light.174, 176 In a pilot study it was recently discovered that directly injection of ICG into the gallbladder shows low fluorescence signals in the liver and seems also feasible, especially in case of severe cholecystitis.182

Before starting with NIR fluorescence cholangiography, 5 mg ICG should be prepared according to a recent optimal dose study of our group.183 Administering ICG directly prior to surgery results in highly fluorescent liver tissue, which may hamper identification of bile ducts. Therefore, ICG has to be injected intravenously 3‐7 h before surgery for optimal bile duct to liver contrast, resulting in a better bile duct discrimination. For elective cholecystectomy, which is sometimes part of oncologic surgery (in case of cholangio‐ or pancreatic‐carcinoma), this approach is feasible. Five studies showed that with this dose the cystic duct beginning to became fluorescent around 30 to 40 min after intravenous ICG injection.158, 159, 169, 170, 176

In general, during cholecystectomy it is important to obtain the critical view of safety: visibility of the cystic duct and cystic artery entering the gallbladder. After preparation of the gallbladder hilum both medially and laterally, critical view could be achieved by retracting the gallbladder caudo‐medially and exposing the triangle of Calot (Figure 4). If the cystic duct was detected, ICG could also be administrated again to visualize the cystic artery. It is important to realize that after a second ICG administration, the liver will become again highly fluorescent. After detection of both the cystic duct and cystic artery, these structures could be transected and the gallbladder could be removed.

Figure 4

Example of bile duct imaging using ICG 24 h after injection. The position of the common bile duct was indicated by the arrow; the liver by L; and surrounding adipose tissue by Ad. Adapted and reprinted by permission from Springer Nature: Surgical Endoscopy171 © 2014

Ureter

Ureter injuries were seen in minority of the cases during colorectal surgery, with a reported incidence varying from 0.15% to 0.66% for open and laparoscopic procedures, respectively.184 Although the incidence seems low, ureteral injuries are associated with a significant morbidity. The use of intraoperative ureter visualization with NIR fluorescence imaging was evaluated in six studies, of which four studies used ICG and two used MB as contrast agents.185, 186, 187, 188, 189, 190 ICG was used in three studies during robotic surgery to detect the level of the ureteral stenosis in a population varying from 7 to 25 patients. In these studies ICG was injected both through the ureteral catheter, of which only the tip was inserted into the ureteral orifice, or in the renal pelvis.

We prefer MB above ICG as a NIR fluorescent contrast agent for detection of the ureter, because MB is cleared via the kidneys and could therefore intravenously injected easily. Verbeek et al190 showed that in 12 patients both ureters could be detected 10 min after infusion of MB, whereas Al Taher et al185 detected both ureters in only a half of the patients (5 out of 10). Verbeek et al190 injected MB intraoperatively after visualization of the ureters. However, this approach is not comparable to clinical practice. Therefore we recommend to inject MB intravenously after first incision was made (open surgery) or after introduction of the first trocars (laparoscopic surgery). Based on both studies, 0.25 mg/kg MB could be injected, after which the ureters will become visible from 10 up to 60 min, as illustrated in Figure 5. A disadvantage of MB is that it is fluorescent at approximately 700 nm, which is subject to higher absorption and more nonspecific background fluorescence. A novel dye, ZW800‐1, was therefore developed. This dye is fluorescent at 800 nm, exclusively cleared via the kidneys.

Figure 5

Example of ureter imaging using MB. Upper panel: NIRF image of right ureter, 45 min after MB administration. Lower panel: NIRF image of right ureter, covered by blood and tissue. Reprinted by permission from Elsevier: The Journal of Urology190 © 2013

To summarize, ICG could be used for biliary duct imaging, whereas MB is preferable for ureter mapping. Optimal doses and injection times are 5 mg 3‐7 h before surgery, and 0.25 mg/kg intraoperatively, respectively (Table 1).

TISSUE PERFUSION

Restoring normal functioning and tissue healing after surgical intervention is, among others, critically dependent on tissue perfusion. Insufficient perfusion with oxygenated blood could result in ischemia and subsequent tissue damage. An important clinical problem is anastomotic leakage, which is one of the most serious complication after gastrointestinal surgery.191, 192 Large retrospective cohort studies showed leakage rates up to 19% in case of colorectal surgery.192 Today, surgeons have to evaluate the anastomotic perfusion by inspection, which is a subjective method and difficult to quantify.193 Different techniques have been used to evaluate the anastomotic perfusion, however none of them are currently common used in clinical practice, due to their disadvantages.194

Intraoperative evaluation of liver perfusion is also a clinical relevant issue, especially for determining the effect of the resection on the liver function, however currently no good intraoperative tools are available. Some studies suggests the use of contrast enhanced ultrasound, as an additional tool for resection planning.195, 196 Due to the characteristic of ICG binding to plasma proteins, it will mainly remain intravascular after intravenously injection, which makes it a perfect candidate for perfusion imaging.

Anastomotic bowel perfusion

As the introduction of NIR fluorescence imaging, 11 clinical studies reported the use of ICG as a NIR fluorescent agent in evaluation of the esophagogastric anastomosis.24, 194, 197, 198, 199, 200, 201, 202, 203, 204, 205 One study compared the anastomotic leakage rate in perfusion detection both with and without the use of ICG fluorescence imaging, which resulted in reducing the rate from 20% to 0%.205 Twenty‐one clinical trials reported the use of ICG as a NIR fluorescence agent with different imaging systems in intraoperative evaluation of colorectal anastomoses.180, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224, 225 ICG was often injected after anastomotic site selection to evaluate the effect of ICG in selecting the correct anastomotic site. This resulted in the change of surgical plan between 3.7% and 40%. Some studies also evaluated the rate of the AL with using ICG and it was shown in a recent systematic review that the rate of AL decreases from 8.5% to 3.3% in the group which used ICG.226 On the other hand, a case‐matched retrospective study in 346 patients did not show a significant decline in AL (7.5% vs 6.4%), although the anastomotic site was changed in 5% of the patients.214

Based on the above mentioned studies we recommend the following approach. First of all, normal surgical procedure could be started according to standard of care. Before performing the anastomosis, 2.5 mg of ICG dye can be injected intravenously, which is based on the doses used in most of the clinical trials.197, 198, 199, 200, 202, 203 However, most studies in colorectal surgery used ICG injection doses varying from 0.2 to 0.5 mg/kg.206, 213, 221, 223, 224 Nevertheless, it was shown that the colorectal anastomotic perfusion could also be visualized by using a much lower and less toxic dose (2.5 mg), which is in concordance with the optimal dose used for evaluation gastroesophageal anastomotic perfusion.210, 212 Hence, we recommend an intravenously injection of 2.5 mg ICG for both types of anastomoses. Within 60 s the vasculature will be visible and the anastomotic site can be visualized with NIR fluorescence imaging.209, 210, 212, 215, 223 If the fluorescence signal starts to fade, another bolus of 2.5 mg ICG can be injected to evaluate the perfusion for a second look (>15 min after first injection). If a good perfusion was detected with NIR fluorescence imaging, the procedure could be finished as usual.

Hepatic perfusion during liver segmentectomy

Due to the complex liver anatomy, liver surgery is only performed in high volume centers. Especially, during oncologic surgery, it is difficult to determine the resection margins and to estimate the remnant liver volume in order to achieve the optimal balance between cancer curability and postoperative hepatic function. The portal and hepatic veins are the important anatomical landmarks used for resection planning. Seven studies evaluated the use of ICG for determining the resection planes during liver surgery.104, 110, 117, 164, 227, 228, 229 In 89% to 100% of the performed surgeries, the resection lines could be easily visualized.

After clamping of the selected portal branch, ICG could be injected, resulting in highlighted liver, except of the less‐perfused part of the liver. Two methods of injection (portal vein and peripheral vein) were evaluated. The used doses of ICG varied from 1.25 mg to 0.5 mg/kg.104, 110, 117, 164, 227, 228, 229 However, Inoue et al228 showed comparable results of portal vein and peripheral vein injection with 2.5 mg ICG, which we therefore recommend as the optimal dose. Intraportal ICG injection could be performed under ultrasonic guidance after clamping the proximal portal pedicle. An alternative for portal injection is an intravenously (in a peripheral vein) ICG injection after clamping of the selected portal vein, which we prefer during laparoscopic surgery, due to the technical difficulties with portal injection. A disadvantage of intravenously ICG injection is the lower concentration of ICG available than using the intraportal injection method.228 The liver segments become fluorescent after 1‐2 min.104, 227

For optimal evaluation of anastomotic bowel and liver segment perfusion a 2.5 mg intravenously injection of ICG is recommended. After intraoperative injection, the vasculature becomes visible after approximately one to 2 min (Table 1).

DISCUSSION AND FUTURE PERSPECTIVES

NIR fluorescence imaging has been widely evaluated during the past decades. Several clinical studies have been performed with different conditions and therapeutic doses of contrast agents. Hence, it may be difficult for surgeons to directly apply these new techniques in daily practice, although the technique itself is fairly straightforward and has a limited learning curve. This review focussed on its application with the currently only available contrast agents, ICG and MB, during oncologic abdominal surgery, of which a summary was given in Table 1. We evaluated the timing of the injection, the injection locations, and the dose of contrast agents, which could guide the surgeon for intraoperative use of the different applications of NIR fluorescence imaging in abdominal oncologic surgery.

The use of ICG, contributes to a better visibility of sentinel lymph nodes, biliary ducts and tissue perfusion, than current image modalities or surgeons visual/tactical feedback. Furthermore, liver tumors could be easily identified with ICG, because of its clearing by the liver, whereas it is not specific for other types of solid abdominal tumors in the abdomen, such as pancreatic and renal tumors.140, 144 Moreover, we have shown that intraoperative visualization of peritoneal metastases and adrenal tumors using ICG is feasible and could therefore be used as an additional intraoperative image modality until specific tumor targets are available. Today, no evidence is available for intraoperative visualization of nerves, which is important during colorectal surgery, using ICG due to its non‐specificity.230, 231, 232 Therefore, diverse other nerve specific contrast agents are currently being investigated.233, 234, 235, 236, 237

MB was evaluated for ureter detection, however its clinical use during abdominal surgery, would be limited subject due to higher absorption and more nonspecific background fluorescence. Hence, a specific antibody binding to a fluorophore, ZW800‐1, has been evaluated and showed better ureter visibility than MB in preclinically testing.238 Moreover, it was also described that MB could be useful in abdominal surgery for detection of neuroendocrine tumors. However, the evidence is limited to one preclinical mice study and a case report of a patient with a paraganglioma, and should therefore be further validated in new studies.239, 240

To improve the tumor visibility during oncologic surgery, (new) NIR fluorophores, such as IRDye‐800CW, were combined with tumor‐specific targets as described by Haque et al.241 Diverse specific tumor‐specific fluorophores have been investigated during the last years, for instance in patients with colorectal cancer (during endoscopy), ovarian cancer, renal cell carcinoma, and peritonitis carcinomatosa.242, 243, 244, 245, 246 Currently, more clinical trials in patients with pancreatic, colorectal and renal cancer are underway (NCT02743975, NCT02317705, NCT01778933, NCT02973672).

Although we have shown the great progress made using NIR fluorescence imaging, it also has his limitations. The most important drawback is the lack of quantitative comparison of fluorescence signals in the currently published studies. Especially, for evaluation of anastomotic bowel perfusion, most of them did not take into account the over‐time diffusion of the used NIR fluorophore, resulting in an overestimation of the fluorescence signal in vascularized areas. Diana et al247, 248, 249, 250 developed a quantitative software based analysis which calculates the stiffness of the slope of the fluorescence signal to reach the peak intensity, given the fact this is the most important perfusion marker. Based on those parameters, a virtual bowel perfusion cartography is created, which could be superimposed on the white light images, resulting in a nearly real‐time quantitative perfusion image, as shown in Figure 6.

Figure 6

The concept of quantitative fluorescence measurements (Fluorescence‐based Enhanced Reality = FLER) for determining the bowel perfusion. Upper panel: The fluorescence signal is analyzed during 40 s after intravenous administration of ICG. Using specific software (VR‐PERFUSION, IRCAD; France) the slope of the fluorescence time‐to‐peak is computed and converted to color codes, resulting in a virtual perfusion cartography. Lower panel: The white light image is combined with the perfusion cartography, creating an augmented reality view of the bowel perfusion at the resection site

Today, NIR fluorescence imaging is performed in the classic first NIR window, between 650 and 950 nm. The penetration depth of NIR light in the first NIR window is up to 1 cm and therefore limited for detection of superficial lesions, however it could even be useful for intraoperative margin assessment.123 NIR fluorescence imaging systems, capable of imaging in the second NIR or shortwave infrared window (1000‐1700 nm), will allow more in depth visibility of structures.

In conclusion, ICG and MB are relatively safe and sensitive non‐specific fluorophores, which have been widely evaluated in NIR fluorescence imaging. In the nearby future, more specific tumor targets will be evaluated and used in clinical practice, which will provide new opportunities in the field of abdominal surgery.

CONFLICTS OF INTEREST

All authors declare that there is no conflict of interest.

SYNOPSIS

A practical guide for the use of NIR fluorescence imaging with indocyanine green and methylene blue during abdominal (oncologic) surgery.