Understanding fluorescence time curves during ileal pouch-anal anastomosis with or without vascular ligation

Background Intraoperative indocyanine green fluorescence angiography (ICG-FA) may be of added value during pouch surgery, in particular after vascular ligations as lengthening maneuver. The aim was to determine quantitative perfusion parameters within the efferent/afferent loop and explore the impact of vascular ligation. Perfusion parameters were also compared in patients with and without anastomotic leakage (AL). Methods All consenting patients that underwent FA-guided ileal pouch-anal anastomosis (IPAA) between July 2020 and December 2021 were included. After intravenous bolus injection of 0.1 mg/kg ICG, the near-infrared camera (Stryker Aim 1688) registered the fluorescence intensity over time. Quantitative analysis of ICG-FA from standardized regions of interests on the pouch was performed using software. Fluorescence parameters were extracted for inflow (T0, Tmax, Fmax, slope, Time-to-peak) and outflow (T90% and T80%). Change of management related to FA findings and AL rates were recorded. Results Twenty-one patients were included, three patients (14%) required vascular ligation to obtain additional length, by ligating terminal ileal branches in two and the ileocolic artery (ICA) in one patient. In nine patients the ICA was already ligated during subtotal colectomy. ICG-FA triggered a change of management in 19% of patients (n = 4/21), all of them had impaired vascular supply (ligated ileocolic/ terminal ileal branches). Overall, patients with intact vascular supply had similar perfusion patterns for the afferent and efferent loop. Pouches with ICA ligation had longer Tmax in both afferent as efferent loop than pouches with intact ICA (afferent 51 and efferent 53 versus 41 and 43 s respectively). Mean slope of the efferent loop diminished in ICA ligated patients 1.5(IQR 0.8–4.4) versus 2.2 (1.3–3.6) in ICA intact patients. Conclusion Quantitative analysis of ICG-FA perfusion during IPAA is feasible and reflects the ligation of the supplying vessels. Supplementary Information The online version contains supplementary material available at 10.1007/s00464-023-09921-y.

Ileal pouch-anal anastomosis (IPAA) helps to restore continuity after proctocolectomy for patients with ulcerative colitis, familial adenomatous polyposis and well selected patients with Crohn's colitis [1]. After pouch surgery, a feared complication is anastomotic leakage (AL) which can occur in up to 15% of patients [2,3]. A traction-free, well-vascularized anastomosis is essential for anastomotic healing, and in pouch surgery these two factors need to be carefully balanced. In order to have a tension free anastomosis, lengthening maneuvers may require vascular ligation [4] of the ileocolic artery(ICA) or ileal arterial branches. In most patients the ICA is already compromised at initial subtotal colectomy, and further vascular ligations may even have a bigger impact.
Intraoperative fluorescence angiography using indocyanine green (ICG-FA) is widely applied to assess tissue perfusion and could contribute to the prevention of AL secondary to perfusion restriction [5,6]. During IPAA, ICG-FA may be of added value as it can guide safe vascular ligations without increasing the incidence of AL [7,8]. However, the interpretation of ICG-FA remains mostly subjective, and quantification of ICG-FA remains a challenge [9,10]. First efforts in quantifying the fluorescence signal during IPAA surgery support that time from ICG administration to fluorescent enhancement of the afferent and efferent loop may be prolonged in patients with vascular ligation [7]. Outflow on the other hand may also provide valuable information, as outflow problems are correlated to venous congestion contributing to ischemia [11]. Taking this into account, it is important to focus on both inflow as outflow parameters.
The objective of the present pilot study is to determine quantitative fluorescent parameters to assess in-and outflow in relation to vascular ligation during IPAA surgery. Intraoperative change of management related to ICG-FA and AL were recorded.

Study design
This case series has been reported in line with the PROCESS Guideline [12]. In this single center pilot study, we included all consenting patients that underwent FA-guided IPAA in Amsterdam UMC from July 2020 until December 2021.
Patients were included when they met the following criteria: 18 years or older, proctocolectomy or completion proctectomy with (redo) IPAA for inflammatory bowel disease or inherited colorectal cancer disorders. Patients were excluded in case of allergy to ICG, iodide or sodium iodide. ICG-FA data were recorded in a prospectively maintained database, along with patient data from the electronic patient system.
The Institutional Review Board of the Amsterdam University Medical Centres (UMC), location Academic Medical Centre (AMC), approved the study protocol and confirmed that the Medical Research lnvolving Human Subjects Act (WMO) did not apply.

Surgical procedure
Different strategies for restorative proctocolectomy with IPAA were applied, including 1 and modified-2-stage procedures [13]. The modified 2-stage procedure was the standard approach after prior subtotal colectomy for refractory disease. In patients who had their subtotal colectomy at Amsterdam Medical Centres, the ileocolic arcade was preserved as a routine. In patients referred from other units, the ileocolic arcade was mostly not preserved, and this was verified through the operative notes and postoperative abdominal CT if available.
Completion proctectomy and IPAA were performed by a combined abdominal and transanal minimally invasive approach as described previously [11]. Standard lengthening maneuvers consisted of dividing all adhesions, mobilization of the terminal ileal mesentery to the level of the duodenum and transverse peritoneal incisions on both sides of the small bowel mesentery. The yardstick for sufficient length was that the apex of the pouch should reach 1-2 cm below the pubic bone. If length was considered insufficient, the ICA and/or interconnecting terminal ileal branches were ligated (Fig. 1). After securing sufficient length, the J-pouch was constructed by a side-to-side ileal anastomosis of 10 cm length using a linear stapler. A double purse string single staple anastomosis was performed as a routine with suture reinforcement. A diverting ileostomy was only created in case of technical problems (e.g., positive reverse leak test or subjective feeling of tension on the anastomosis) or aberrant ICG-FA findings (i.e., delayed fluorescence enhancement based on subjective interpretation of the surgeon).

Standardized fluorescent assessment
Perfusion of the pouch was assessed extracorporeally before connection of anvil and stapler for the anastomosis. The operating table was placed in a neutral position, the laparoscopic Stryker 1688 camera system (Stryker, Kalamazoo, MI, U.S.A.), 30 degree optic was placed in a fixed position 9 cm perpendicular from the anvil in the top of the pouch. All light in the operation room was switched off to minimize external light reflection. ICG (Verdye, Diagnostic Green; 0.1 mg/kg/bolus) was injected peripherally as a bolus directly in the iv catheter of the left arm without iv extension. Starting with the moment of ICG injection, perfusion was captured in a continuous video recording for 200 s.

Quantification of fluorescent imaging
In order to achieve objective quantification, the raw ICG-FA video data were analyzed post hoc by tailor made software written in Python (COPYRIGHT). After loading the video into the software, size was calibrated using a measuring tape which was placed in the frame. Subsequently, a circular region of interests (ROI) with a diameter of 1 cm was placed in the middle of the pouch body, both on the afferent and efferent loop (Fig. 2). The software subsequently extracted the mean intensity within the ROI for every frame and plotted the ICG in-and out-flow in a fluorescence-time curve. A slightly modified version of the arterial input function reported by Elliott et al. was fitted to the curve to reduce the influence of noise on the calculated parameters [14]. From this fit, the following parameters were extracted ( Fig. 3): Influx time point (t 0 ): the time point at which the fit started, thereby increasing to above baseline intensity, F max : maximal intensity in arbitrary units (AU), T max : time in seconds from ICG administration until F max has been reached., time-to-peak (ttp); time in seconds from τo until F max has been reached, mean slope: mean rate at which the fluorescence intensity increased between t 0 and T max (AU/s), T 90% : time in seconds after F max until fluorescence intensity has dropped to 90% of F max and T 80% : time in seconds after F max until fluorescence intensity has dropped to 80% of F max .

Outcomes
The primary outcome was to determine quantitative perfusion parameters during IPAA within the efferent/afferent loop and explore the impact of lengthening measures requiring vascular ligation or previous inadvertent interruption of the ileocolic arcade. Perfusion parameters in patients with AL were also compared to those without.
Secondary outcomes included change in management due to FA, AL within 90 days and reinterventions due to AL. Change in management by FA was defined as every measure taken based on the results of FA only, i.e., suture reinforcement of hypo-perfused regions, additional resections or preserving the pouch and creating an ileostomy. Interpretation

Statistics
Patient baseline characteristics and imaging characteristics are summarized using simple descriptive statistics. All categorical data will be presented as number of cases and percentages, while continuous data will be shown as either mean ± standard deviation or as median and interquartile range (IQR), depending on the data distribution. Data was analyzed using the Statistical Package for Social Sciences (SPSS) of IBM Statistics, version 26.0, or the latest version.

Baseline characteristics
In total, 21 patients were included in this study with a mean age of 40.5 ± 11.3 years at time of IPAA. Half of the patients were male (52.4%). Patient characteristics are outlined in Table 1.

Operative characteristics
Operative characteristics are outlined in Table 2. The abdominal approach was in all except two patients via laparoscopy, requiring conversion into a laparotomy in one patient due to the extensiveness of dense adhesions. All patients underwent transanal minimally invasive surgery for the proctectomy, rectal cuff mobilization, and double purse string single stapled anastomosis. Influx time point (t 0 ): the time point at which the fluorescence intensity in the ROI was statistically significantly larger than the background, F max : maximal intensity in arbitrary units (AU), t max : time in at which the background-corrected fluorescence intensity reached F max , time-to-peak (ttp): T maxt 0 , the green line represents the mean slope: rate at which the fluorescence intensity increased (AU/s), t 90% : time in seconds after F max until 90% of F max has been reached, t 80% : time in seconds after F max until 80% of F max has been reached

Vascular ligation
In almost half of the patients the ICA had already been ligated during the previous subtotal colectomy. During IPAA, intraoperative vessel ligation was performed in three patients to obtain sufficient reach for anastomosis, consisting of ligation of the ICA in one patient and interconnecting terminal ileal branches in two patients. Both patients with ligation of interconnecting terminal ileal branches had intact ICA.

Vascular ligation group
In Table 3  Both patients in whom the interconnecting terminal ileal branches were ligated had macroscopic blue colorization of the efferent loop during the surgery. In one patient, T max in both the afferent as efferent loop was prolonged (see Table 2 Operative characteristics Data is shown in n (%) or n/n), unless otherwise stated, *redo pouch due to FAP overgrowth in the initial pouch

Change of management due to ICG-FA assessment
Based on qualitative assessment of the ICG-FA, change of management was opted in four out of 21 patients (19%). Of these, two patients had prior ICA ligation and two patients necessitated intraoperative ligation of the interconnecting terminal ileal branches to obtain more length. One patient with prior ICA ligation had delayed fluorescent enhancement of the entire pouch, with no fluorescent enhancement in the apex of the pouch. This region was inverted by sutures. In retrospect, flat slopes in both the afferent as efferent loop (0.75 and 0.64) were observed. This patient developed an AL on postoperative day 16. In the second patient with prior ICA ligation, there was a delayed fluorescent enhancement of the efferent loop of the pouch. The anastomosis was constructed, and a protective ileostomy was created, switching from a planned modified 2-stage into a 3-stage procedure. In one patient after ligation of the interconnecting terminal branches, a delayed enhancement during ICG-FA was observed in the apex of the pouch, which was not observed by white light assessment, therefore this hypoperfused area was inverted by sutures. In the other patient after ligation of the interconnecting terminal ileal branches, the pouch body discolored gradually in white light, but subsequent ICG-FA showed uniform and rapid enhancement. The operative plan was changes from a 1-stage into a 2-stage procedure. The latter three patients had an uncomplicated postoperative course.

Postoperative outcomes
Fluorescence parameters are summarized in Table 4, in which the results are also shown for patients with or without AL separately. AL was observed in three out of 21 patients (14%), all located at the circular anastomosis and diagnosed on postoperative day 5, 12, and 16. All patients did not have an ileostomy at moment of diagnosis. Two of the AL were diagnosed by endoscopic examination and the remaining one by CT scan. All patients were treated by creating an ileostomy and subsequent Endo-SPONGE® (B. Braun Surgical S.A., Barcelona, Spain) vacuum assisted closure (EVAC) procedures [15]. All ileostomies were reversed within 1 year, without long term sequelae.
Longer time values were observed in these few patients for both in-as out-flow parameters with respect to the afferent as efferent loop (see Table 4). The mean slope in the afferent loop was less steep in the AL group (1.4 versus 2.4 AU/s).

Discussion
By quantifying ICG-FA of 21 patients undergoing IPAA surgery, we were able to determine quantitative in-and outflow parameters within the pouch and explore the impact of lengthening measures requiring vascular ligation or previous inadvertent interruption of the ICA. In patients without vascular ligation of the ICA perfusion parameters between afferent and efferent loops were similar. However in patients with vascular ligation of the ICA both in-and out-flow time values were prolonged and the mean slope was less steep in the efferent loop of the pouch.
This study supports prior results, patients without vascular ligation show rapid in and outflow similarly in both afferent and efferent loop. In case of ligation, changes are observed predominantly in the efferent loop with in-and out-flow. This might be due to redirection of arterial flow through the arcade or potential venous obstruction. By contrast in case of AL, changes are also seen, but particularly in the afferent loop. ICG-FA is a promising tool to demonstrate adequate perfusion in gastrointestinal surgery [16]. This technique is potentially of added value during IPAA surgery, especially after vascular ligation [7,8,11,17]. However, it remains challenging to differentiate between inflow (arterial) and outflow (venous) problems. Prior work indicates that it might be categorized in inflow problem (no ICG fluorescence), an outflow problem (delayed, but intact ICG flow), and adequate perfusion (rapid ICG flow) [7]. However, the use and interpretation of ICG-FA enhancement depends on subjective interpretation and in-and out-flow might affect each other.
This study shows that changes are visible in the shape of the curve and quantification parameters in the event of vascular ligation. Despite the fact that the curves in this study were not generated in real time, the clinical application for employing ICG-FA during IPAA might lie there. It can be challenging to determine intraoperatively whether the ICA is still intact. Particularly in patients with visceral obesity and in those cases where the arcade has been interrupted by ligation the descending branch of the ICA. In these patients ICG-FA can be applied to explore the integrity of the ICA and if needed to assess whether additional ligation of the interconnecting terminal ileal branches is possible. It is important to pay attention to keep the ICA intact when performing the colectomy since this might compromise perfusion and endanger the pouch's (specifically the efferent loop's) perfusion. This is the first study to report on quantitative parameters of fluorescent time curves in IPAA surgery. All measurements were performed in a standardized manner and similar conditions. This produced valuable in and outflow ICG fluorescence data not reported before. The current study created new knowledge on pouch perfusion in general and explored the effect of lengthening measures on quantitative pouch perfusion patterns. The described technique is reproducible and will lead to a more objective interpretation of fluorescence angiography. A limitation of the study is that the cohort is too small to draw any firm conclusions regarding quantitative parameters and anastomotic leakage. However trends  (11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22) appear to be evident in the enhancement curve of the mainly the afferent loop of the pouch. If a quantitative threshold can be determined, this may select high risk patients for AL. Larger prospective trials should be carried out enabling multivariate analysis to identify a fluorescent threshold that may predict anastomotic leakage. This might influence clinical decision making intra-or post-operatively. For example changing a modified 2-stage into a 3-stage procedure or by taking pre-emptive measures postoperatively; for instance performing an early pouchoscopy for anastomotic inspection or preemptive endoSPONGE placement [11]. Besides, one fluorescence imaging system with standard settings was used. It is unknown how these findings relate to a setting with a different imaging system. In future studies, calibration of imaging systems is needed to identify differences in fluorescence read-out. Furthermore, in this study the fluorescent signal was still interpreted subjectively for intraoperative decisionmaking.
In conclusion, this study explored the effect of vascular ligation on pouch perfusion using quantitative ICG-FA. Evident changes in perfusion were found after ligation, mainly in the efferent loop of the pouch. ICG-FA can detect previous vascular ligation and might guide surgeons during clinical decision making intra-or post-operatively in IPAA.