Accurate intraoperative real‐time blood flow assessment of the remnant stomach during robot‐assisted distal pancreatectomy with celiac axis resection using indocyanine green fluorescence imaging and da Vinci Firefly technology

Ischemic gastropathy is one of the unique postoperative complications associated with distal pancreatectomy with celiac axis resection for locally advanced pancreatic cancer. Therefore, it is essential to evaluate blood flow to the stomach following a resection; however, no intraoperative procedures have been established to assess this issue. Herein we describe two cases in which intraoperative evaluation of real‐time blood flow in the residual stomach was performed using indocyanine green fluorescence and da Vinci Firefly technology during a robot‐assisted distal pancreatectomy with celiac axis resection.

K E Y W O R D S distal pancreatectomy with celiac axis resection, firefly, intraoperative real-time blood flow assessment 1 | BACKGROUND Pancreatic cancer is the most common malignancy worldwide. The 5-year survival rate is approximately 10%. 1 Approximately 20% of pancreatic cancers are resectable at the time of diagnosis. 2 In recent years, perioperative chemotherapy and radiation therapy regimens have been improved. Thus, indications for surgery in patients with locally advanced pancreatic cancer have been expanded. 3 For locally advanced pancreatic cancer with celiac axis and/or common hepatic artery invasion, a distal pancreatectomy with celiac axis resection (DP-CAR) is the procedure of choice. 4 DP-CAR, however, has a high overall morbidity rate. 5 Indeed, one of the unique complications of DP-CAR is ischemia. 6 Because the celiac axis is dissected during a DP-CAR, it is important to ensure hepatic and residual gastric blood flow. Poor blood flow to the stomach may lead to ischemic gastropathy and gastric necrosis. Preoperative evaluation of vascular anatomy by computed tomography (CT) or angiography is important. The intraoperative real-time evaluation of blood flow is needed to improve safety, but the clinical value and efficacy have not been established.
Recently, robot-assisted surgery has been increasingly indicated in patients with pancreatic cancer. The da Vinci surgical system has a function (Firefly) in which nearinfrared light emitted from excited indocyanine green (ICG) fluorescence is processed by a computer and displayed in green. In this study we performed intraoperative real-time blood flow assessment of the remnant stomach in two patients during a robot-assisted DP-CAR using ICG fluorescence imaging and da Vinci Firefly technology. One patient underwent a total gastric preservation and the other patient underwent a proximal F I G U R E 1 Computed tomography scan showed the tumor location (A), three-dimensional image (B), the soft tissue around the tumor surrounded the celiac artery, but no invasion to the proper hepatic and gastroduodenal arteries (C,D). gastrectomy for gastric invasion of pancreatic cancer. Both patients had good blood flow to the stomach and did not require revascularization or an additional gastrectomy. The postoperative course was uncomplicated. In this report we describe a real-time blood flow assessment of the stomach during robot-assisted DP-CAR using ICG fluorescence imaging and da Vinci Firefly technology.

| METHODS
Robot-assisted DP-CAR with celiac axis resection was performed using a da Vinci Xi surgical system (Intuitive Surgical, Sunnyvale, CA, USA) on two patients with locally advanced pancreatic cancer and suspected invasion of the celiac artery. ICG (0.5 mg/kg) was injected intravenously after resection to evaluate real-time blood flow of the stomach using a da Vinci Firefly system. Blood flow of the stomach was evaluated 60 seconds after the intravenous injection of ICG.
Case 1 was a 74-year-old male with pancreatic cancer (25 mm in diameter) located in the body of the pancreas.
A CT scan showed that the tumor was located near the celiac artery and the soft tissue around the tumor surrounded the bifurcation from the celiac artery to the splenic artery / common hepatic artery ( Figure 1). Because there were no findings of invasion to the proper hepatic and gastroduodenal arteries, we decided to perform a DP-CAR.
First, the suprapancreatic lymph node dissection was performed and the common hepatic artery and pancreatic body were taped. The area around the superior mesenteric artery (SMA) was dissected as much as possible from the ventral view. The right dorsal view was then obtained, the Kocher maneuver was performed, and the left renal vein was taped. In addition, the roots of the celiac artery and SMA were identified in the same field of view, and the surrounding area was dissected. Switching again to the ventral view, the common hepatic artery (CHA) was test-clamped, intrahepatic blood flow was confirmed, and the CHA was dissected. After the pancreas was dissected with an automatic suture, the dissection around the SMA proceeded to the root, and the dissected layer from the left ventral side was made continuous with the dissected layer from the right dorsal aspect, which was already dissected, and the celiac artery was taped. After test clamping again and confirming intrahepatic blood flow, the celiac and left gastric arteries were dissected. The Gerota fascia was dissected, the pancreatic tail and spleen were dissected, and the specimen was removed. After specimen removal, ICG was intravenously injected to evaluate gastric blood flow in real time. It was confirmed that there was sufficient blood flow within 60 seconds (Figure 2), even in the region of the left gastric artery, a favorite site of poor blood flow. Therefore, reconstruction of the left gastric artery was not performed, and the surgery was completed with preservation of the entire stomach.
Case 2 was a 76-year-old male with pancreatic cancer (35 mm in diameter) located in the body of the pancreas. A CT scan showed that the tumor was located near the celiac artery; tumor invasion into the lesser curvature of the stomach was suspected (Figure 3). Because the vascular invasion morphology was similar to Case 1, we decided to perform a DP-CAR. In this case, the tumor directly involved the gastric lesser curvature and the left gastric artery; therefore, we decided to perform a proximal gastrectomy.
First, a proximal gastrectomy was performed. The stomach was separated at the lower part of the gastric body, where no invasion was observed. A DP-CAR was then performed using the same procedure as in Case 1. Finally, the peri-esophageal area was trimmed, and the specimen was removed. Next, ICG was injected intravenously, and after confirming good blood flow in the residual stomach and stomach stump (Figure 4), an esophageal residual gastric anastomosis was performed.

| RESULTS
The operative time for Case 1 was 628 min and the estimated blood loss was 766 mL; however, a blood transfusion was not performed. There was no evidence of ischemic gastropathy or delayed gastric emptying; only diarrhea was observed (Clavien-Dindo classification grade 2). The hospital length of stay was 38 days. The operative time for Case 2 was 662 min and the estimated blood loss was 628 mL; however, a blood transfusion was not performed. The postoperative course was unremarkable; there was no evidence of ischemic gastropathy or delayed gastric emptying. The patient developed a chyle fistula, which improved with conservative treatment (Clavien-Dindo classification grade 2) and was discharged to home on postoperative day 23.

| DISCUSSION
It is important to achieve an R0 resection in the surgical treatment of pancreatic cancer. DP-CAR has been performed to achieve an R0 resection for pancreatic cancer with invasion of the celiac or common hepatic artery; however, ischemic complications have been reported as a serious complication following dissection of the celiac artery. 7,8 Specifically, ischemic gastritis and gastric necrosis have been reported when the stomach is preserved. 7,8 When the celiac artery is dissected, blood flow to the stomach is maintained by blood flow from the right gastric artery via the pancreatic head arcade and the right gastroepiploic artery. Therefore, to avoid ischemic complications, blood flow should be carefully evaluated, and if poor blood flow of the stomach is observed, revascularization of the left gastric artery and/or a gastrectomy should be performed. Reconstruction of the left gastric artery have been reported that it is a feasible option and would enhance safety. 9 Accurate evaluation of blood flow to the stomach is essential when deciding to perform additional operative procedures. In addition to preoperative examinations, such as a preoperative CT scan for vascular anatomy and angiography, real-time intraoperative evaluation of blood flow is extremely important, but a method for this evaluation has not yet been established.
ICG fluorescence methods have been applied clinically in various specialties, and in hepatobiliary pancreatic surgery to visualize bile anatomy and the hepatic regions. [10][11][12] ICG fluorescence has been used in esophageal and colorectal surgery. [13][14][15] ICG fluorescence is less commonly used for gastric blood flow evaluation during DP-CAR. Currently, the da Vinci Xi system is frequently used for robotic surgery. This system has a function (Firefly) in which near-infrared light emitted from excited ICG fluorescence is processed by a computer and displayed in green. From our institution, we have previously reported on the usefulness of assessing blood flow at the anastomosis in robotic distal gastrectomy using the da Vinci Firefly system. 16 In this study we performed real-time blood flow evaluation of the stomach in two patients during robotassisted DP-CAR using ICG and Firefly.
The celiac artery was transected at the root in Case 1, but the entire stomach was preserved. The evaluation based on Firefly showed good gastric blood flow in the region of the left gastric artery, which is the most common site for ischemia, and good blood flow of arteries in the entire gastric wall were also confirmed. Therefore, no additional left gastric artery reconstruction or gastrectomy was performed, and the postoperative course was good. In Case 2, gastric invasion by pancreatic cancer was noted, so we decided to perform a resection involving part of the stomach. We performed a robot-assisted DP-CAR and proximal gastrectomy. Before reconstruction, blood flow in the residual stomach was evaluated in real time. Because blood flow was good, a residual gastroesophageal anastomosis was performed and a left gastric artery reconstruction was not indicated. No postoperative gastric necrosis or ischemic gastroenteropathy was observed, and the patient was discharged from the hospital after an unremarkable postoperative course. The ability to evaluate blood flow in real time intraoperatively is a great advantage of this method, which makes it possible to make an accurate and timely decision with respect to the need for additional procedures.
Real-time blood flow assessment using ICG and Firefly are very useful; however, there are some drawbacks, such as the lack of quantification and the indeterminate evaluation time. In the future, it will become a better decision-making tool if we can set an appropriate time for evaluation and quantify the results based on more cases.
In conclusion, intraoperative real-time blood flow assessment of the remnant stomach using ICG fluorescence imaging and da Vinci Firefly technology was shown to be useful in determining the need for reconstruction of the left gastric artery and/or additional resection of the remnant stomach during robot-assisted DP-CAR.