The investigators also aim to use our previously described quantifying method (q-ICG), to evaluate changes in gastric perfusion, during resection of cancer in the gastroesophageal junction (GEJ) and if a change in the operative technique influences microvascular flow in the gastroesophageal (GE) anastomosis.

2. Introduction

2.1 Adenocarcinoma of the gastroesophageal tract Adenocarcinoma of the gastroesophageal tract, including the esophagus, stomach, and gastroesophageal junction (GEJ) is a growing clinical challenge worldwide due to the increasing incidence and poor prognosis. Current management of resectable GEJ cancer involves comprehensive diagnostic procedures followed by standardized treatment with surgery and pre- and postoperative chemotherapy.

At our department, 90 patients are annually treated for resectable cancer of the GEJ. The surgery comprises of an Ivor Lewis procedure with either open or robot-assisted mobilization of the stomach with limited (D1+) lymphadenectomy and creation of the gastric conduit and pyloroplasty (2). Here, the creation of the gastric conduit involves partial preservation of the right gastric artery. The right gastric artery is ligated approximately half way up the lesser curvature, and the remaining conduit stapled (figure 1). Theoretically, the preservation of both the gastro-epiploic and right gastric artery should result in improved vascular patency and blood supply to the anastomosis. The theoretical advantages of preserving the right gastric artery may have some oncological drawbacks as placing the line of resection half way up the lesser curvature allows for a limited lymph node (LN) dissection, leaving behind some tissue comprising of LN station number 3b and 5 (figure 1). Lymph node metastasis is a significant prognostic factor in gastroesophageal cancer, and the American Joint Committee on Cancer recommends an extensive lymphadenectomy to achieve accurate N staging.

In contrast, the creation of a gastric tube by serial applications of a linear-cutting stapling device parallel to the greater curvature (ligating the right gastric artery) could generate a larger LN harvest, but this could be at the expense of anastomotic and conduit perfusion. A retrospective series noted an increase in gastric tip necrosis and neck leaks with intrathoracic extension when a narrower gastric tube (3 to 4 cm) was used. Contrary, several randomized trials, and cohort studies have been conducted regarding the size of the gastric tube with no differences in anastomotic leakage rates being reported. However, the consensus is that a gastric tube of five cm width in both open and minimal invasive oesophagostomy is the optimal size. No difference in the rates of anastomotic leakage, anastomotic stenosis, or delayed gastric emptying has been seen between ligating or preserving the right gastric artery. Thus, both ligating and preserving the right gastric artery are considered viable surgical options.

Figure 1. A) Schematic image of the gastric blood vessels. B) Resection when ligating the right gastric artery, C) Resektion when preserving the right gastric artery. D) Schematic image and partial lymphatic overview, circled are lymph node stations 3b. and 5. resected in "B" and left behind in "C".

2.2 Fluorescence guided surgery with Indocyanine Green. Fluorescence guided surgery enables visualization of structures that are otherwise hidden from the naked human eye. A fluorescent contrast agent is used, often Indocyanine Green (ICG), and by illuminating the tissue with near-infrared light, the excited ICG can be detected by a camera with an optical filter. ICG is a tricarbocyanine dye with extremely few adverse events, a short half-life, it is metabolized exclusively in the liver, and excreted unchanged in the bile. The safety of ICG is well established, and it is used routinely in surgical clinics worldwide and has also in recent years gained popularity with regards to oncological surgery.

ICG binds to plasma proteins after intravenous injection, and by illuminating the tissue with near-infrared red light the fluorescence signal over time during first passage past tissue is considered proportional to microvascular flow, i.e., perfusion. Assessing gastrointestinal blood supply is a challenging task, even for experienced surgeons. Of specific concern is the vascular supply of the anastomosis since poor blood supply is a significant risk factor for anastomotic leakage. An esophagectomy is associated with an anastomotic leak rate of approximately 5 to 25 %. Anastomotic leakage after esophagectomy is a severe complication and associated with up to 17 % 30-day mortality. As ICG angiography allows for real-time visualization of visceral perfusion, an inadequate vascularisation of the anastomosis (despite satisfactory macroscopic appearance) can be detected and addressed intraoperatively. Thus, the use of perioperative ICG angiography has been shown to lower the risk of anastomotic leakage.

A study from our group evaluated intraoperative changes in the gastric microcirculation with laser speckle contrast imaging (LSCI) during Ivor-Lewis esophagectomy. A decreased microcirculation at the gastric corpus by 25% during mobilization of the stomach to the thorax was found. Also, ICG fluorescent intensity, as a sign of reduced perfusion, has been shown to drop after the formation of the gastric conduit and yet again before anastomosis formation compared with baseline values. This decrease was not detectable macroscopically in normal white light. However, macroscopic ICG assessments are useful as the speed of blood flow along the gastric conduit wall (visually assessed with ICG) could help predict the risk of anastomotic leakage after esophagectomy. Still, cut-off values for quantitative anastomotic perfusion associated with anastomotic leakage after resection of GEJ cancer is lacking. In conclusion, assuring adequate conduit perfusion is mandatory, and ICG angiography has been proven to be a viable aid in this regard.

Indocyanine Green retention in LNs has been shown to predict malignant content. In patients with esophageal- and GEJ cancer, a submucosal endoscopic ICG injection allowed the visualization of the tumor's lymphatic drainage. This visualization, could advance the diagnostics of malignant spread, increase lymph node harvest, and significantly affecting the choice of surgical resection.

The investigators have developed an ICG quantification algorithm which has been validated and described earlier. This algorithm has now been incorporated into a touch screen tablet, allowing for live perioperative quantitative perfusion assessments with ICG (q-ICG). A color-coded map of perfusion intensity is provided as an overlay on the white light visualized tissue (Figure 2). Previously, the quantification measurements had to be performed via analyses of video-recordings after the procedure. The investigators believe that q-ICG has the potential to improve intraoperative evaluation of tissue viability further and as such, improve the surgical plan and outcome in patients with GEJ cancer. In a feasibility study of ten patients undergoing GEJ-resection, significant alterations of optimal perfusion points selected by surgeons were found when comparing points selected in white light, ICG, and q-ICG.

In this trial, intravenous ICG injections will be used to evaluate conduit, and anastomotic perfusion and peritumoral intramucosal ICG injections will be used to evaluate a potential improvement in LN detection and harvest. Lymph node quantitative fluorescents will be compared with histological findings.

Figure 2. Gastric conduit after pull-up to the thorax viewed by white light, Near-Infrared Light (ICG) and with q-ICG overlay.

2.3 Aim The investigators aim to compare quantitative gastric conduit and anastomotic perfusion when preserving or ligating the right gastric artery. Further, the investigators wish to investigate weather peritumoral ICG injection will allow for additional lymph node visualization and harvest and examine the correlation between quantitative ICG fluorescents and histological findings.

2.4 Hypothesis The investigators hypothesize that ligating the right gastric artery will result in lesser perfusion of the gastric conduit and anastomosis. Further, the investigators hypothesize that peritumoral ICG injection will improve the identification of regional lymph nodes and potentially improve staging and the extent of lymphadenectomy.

3. Methods 3.1 Study design A randomized clinical trial comparing the effect of two different robotic-assisted surgical techniques for gastric conduit construction on conduit and anastomotic perfusion in patients with GEJ cancer. Patients will be allocated either to preserving the right gastric artery or to an extensive lymphatic resection (ligating the right gastric artery). Before and after the creation of the gastric conduit, and after the creation of the GE anastomosis, perfusion will be assessed with q-ICG.

The sample size calculation is based on estimating prediction models describing the relationship between the vascular ligation and gastric perfusion using linear regression. The investigators expect a reduced perfusion by 17%. The desired sample size is therefore n=70 (35 + 35), (effect 10%, SD 15%, Power 80%, alpha 0.05). Prospective evaluations of the surgical outcome will be recorded, by means of anastomotic leakage rates (verified endoscopically or radiologically), delayed gastric emptying (verified with barium swallow examination and/or endoscopy), and other postoperative complications including mortality. Also, the relationship between the quantified fluorescence of LNs and the pathology result will be examined in vivo and in vitro. The trial will be conducted in accordance with the CONSORT guidelines.

3.4 Recruitment Patients will be approached in the out-patient clinic. They will be informed orally, and written information will be provided. Written informed consent will be obtained from all patients before the procedure. All patients are informed of the possibility of withdrawal of their consent at any time during the trial without any consequence for the operation of their treatment. First, after written consent will inclusion in the study be possible.

3.5 Randomization The investigators wish to include 70 patients; 70 envelopes will be prepared each containing a "ligation" (n=35) or "preservation" (n=35) note. These seventy envelopes will be prepared and sealed by the trial investigator. A trial assistant will then mark these envelopes 1-70 at random using a number generator. Once the surgical set up is in place, one envelope will be chosen for each patient using a number generator, and the group allocation will be acquired. This patient CRP and Trial number will be saved on a secure server (L:drev). Separately the trial number and clinically relevant data will be stored on a secure server (V:drev).

3.6 Indocyanine green Indocyanine green is a well-described nontoxic tricarbocyanine dye used for decades in ophthalmology, cardiology, and hepatology. Very few mild adverse reactions have been reported, but caution in patients with thyrotoxicosis, allergy towards iodine or indocyanine green should be made. When injected, it binds to blood lipoproteins and is solely metabolized by the liver and excreted in the bile, with a short half-life of approximately 4-5 minutes. Indocyanine green is already used as a routine in many specialties; here among surgery, ophthalmology, hepatology, and cardiology.

3.7 Surgical procedure and perfusion assessment

3.7.1 Resection of the gastroesophageal junction The patient is placed under general anesthesia and, using a standard adult endoscope, 2 mL of ICG doubly diluted in sterile water (5 mg/mL) is injected into the submucosa at the four quadrants around the tumor (0.5 mL for each injection). The robotic setup is after that completed. Based on previous studies, the time from injection to lymph node visualization is approximated to 10-30 min.

After the initial presentation of the stomach is acquired, a bolus of ICG (0.2 mg/kg body weight) will be injected intravenously and flushed with 5 mL of saline. The ROIs will then be assessed with q-ICG. The anesthetic protocol will to this point match that of the setting during the screening laparoscopy. The q-ICG assessments then repeated after the dissection, division of the feeding blood vessels and the creation of the conduit. Finally, once the gastroesophageal anastomosis is completed a q-ICG assessment of anastomotic perfusion will be performed.

3.7.2 Fluorescence angiography The GEJ resection will be carried out with the da Vinci® robot, (Intuitive Surgical Systems, Sunnyvale, CA, USA) and the fluorescent angiography will be performed using the infrared endoscopic equipment on the da Vinci® robotic system (Firefly® in the DaVinci Xi).

3.8 Statistics Comparison of the gastric and anastomotic perfusion between groups will be performed using the Mann Whitney U test or a paired sample t-test depending on either a non- or parametric nature of the data. Between-group differences in continuous variables were assessed using analysis of variance, while those in categorical variable were assessed using pooled relative risk with 95% confidence interval (CI). A P-value < 0.05 will be considered significant. Statistic evaluation will be performed using IBM SPSS Statistics © (v 22.0 SPSS Inc. Chicago, IL, USA).

4. Adverse events, risks, and disadvantages Adverse events associated with the use of ICG are rare, and severe anaphylactic reactions are extremely rare. However, safety precautions will be made by excluding patients with allergy to iodine, ICG, or shellfish. Also, the investigators will not include patients with severe kidney and liver insufficiency, thyrotoxicosis, and pregnant or lactating women. The total dose of dye injected will be kept well below the recommended 2 mg/kg. There are no data available describing the signs, symptoms, or laboratory findings accompanying overdosage. The LD50 after I.V. administration ranges between 60 and 80 mg/kg in mice, 50 and 70 mg/kg in rats and 50 and 80 mg/kg in rabbits.

No difference in the rates of anastomotic leakage, anastomotic stenosis, or delayed gastric emptying has been seen between ligating or preserving the right gastric artery, and both are considered viable surgical options.

5. Patient data Upon meeting in the out-patient clinic, the patients will be informed about an ongoing study and asked if they wish further information. If accepted the principle investigator will be called and to briefly inform the patient about the project, and ask about allergy towards; iodine, indocyanine green or shellfish, liver insufficiency, thyrotoxicosis, pregnancy or lactation. If no allergies exist, the patient will then receive written, and oral information about the project and a follow-up telephone call regarding inclusion will be held. If interested in participation, the patient will supply the principle investigator with written consent. The written consent provides with access to obtain information in the patient's electronic journal. This, to see information about the subject's health and the course during hospitalization, necessary as part of the research project. Including self-control, quality control, and monitoring, which are obligatory. Data obtained from electronic patient records are registered prospectively and pseudo-anonymized in a database and kept at the secure server provided by Region Hovedstaden (V:drev). Thus, data in this database may not be identified back to the patient without access to a patient participation list linking patient identification number and the pseudo anonymization number. This participation list will be stored in another folder on the secure server provided by the Region Hovedstaden (L:drev). Only the principle investigator, supervisors and will have access to these folders and files. The law on the processing of personal data (Persondataloven) will, of course, be followed. No data will be sent to other countries. Permission from the national data protection (Datatilsynet) agency will be obtained.

6. Ethical considerations Approval from the regional ethical committee and the national data protection agency will be in place before initiating the study.

ICG is a safe and non-toxic fluorescent dye as described and has been used for decades in ophthalmology, cardiology, and hepatology. Very few adverse reactions are reported. Patients may withdraw from the study at any time. The patients participating in the study will not be paid to participate. Both preserving and ligation the right gastric artery are considered safe and viable surgical approaches.

7. Publication The investigators aim to publish negative, positive, and inconclusive results in international reputed surgical journals, possibly as open access with Jens Osterkamp as the primary author. Also, the investigators will present the work at national and/or international surgical conferences and meetings to share the findings with a large crowd of clinicians and researchers.

8. Timeframe The study will be initiated as soon as the investigators have approval from the ethical committee and the national data protection agency. Hopefully, starting in early 2020, and completed ultimo 2023.

9. Economy This study is a part of the principle investigators Ph.-thesis. Salary and experimental cost will be cover by a grant provided by Medtronic. With regards to study design, realization, and presentation, the research group has complete autonomy. No remuneration will be given to patients participating in the trial. The principal investigator has in collaboration with the project group initiated the project. There are no financial interests for the project group in this project. Equipment (tablet, video-capture device, laparoscopic equipment, surgical robot) is already present in the department. If additional support is granted, the name of the foundation(s), size of the amount, any relation between the grant provider and investigators, as well as any requirements of the grants will be announced to the committee and project participants (patients).

10. Insurance Patients will be covered by the national Danish patient insurance.

Inclusion Criteria:

• Patients (above 18 years) scheduled for planned open or robot-assisted resection of gastroesophageal junction cancer.

Exclusion Criteria:

• Allergy towards; iodine, indocyanine green or shellfish
• Severe liver insufficiency
• Thyrotoxicosis
• Nephropathy requiring dialysis
• Pregnancy or lactation
• Legally incompetent for any reason
• Withdrawal of inclusion consent at any time