GS3 - The Grass Is Not Always Greener:ICG Use During Colorectal Surgeries May Not Be Associated with Improved Anastomotic or Oncologic Outcomes

Purpose/Background: Indocyanine green (ICG) near-infrared visualization (NIR) is being increasingly used in colorectal surgery to evaluate perfusion of anastomotic limbs and in aiding lymph node harvest in oncologic surgery. ICG-NIR has become more readily available with increased adoption of laparoscopic and robotic platforms. Proponents of this adjunct claim improved anastomotic and oncologic outcomes, but no large-scale studies are reported in the literature. This study aims to evaluate the effect of ICG-NIR on outcomes of elective colorectal surgeries on a nationwide scale.

Methods/Interventions: This was a retrospective cohort analysis of 2016-2020 ACS-NSQIP colorectal specific database. Patients who underwent elective laparoscopic/robotic colorectal surgery (low anterior resection [LAR], partial colectomy [PC] with colo-colonic anastomosis, or right colectomy with ileocolonic anastomosis [RC]) were identified using CPT codes. Patients were stratified into two groups based on intraoperative ICG-NIR. Outcomes measured were operative duration, anastomotic leak, unplanned conversion to open, return to the OR, perioperative bleeding requiring blood transfusions, survivor-only hospital length of stay (LOS), mortality, and unplanned readmissions. We also assessed lymph node harvest and margin-positive status for patients who underwent oncologic surgery. Binary logistic regression analysis was performed to identify the independent effect of ICG-NIR use on anastomotic leak rates, adjusting for demographics, comorbidities, pre-operative lab-work, bowel preparation, and operative characteristics

Results/Outcomes: 95,179 patients underwent elective laparoscopic/robotic colorectal surgeries (ICG-NIR: 1,101 [1.2%]; No ICG-NIR: 94,078 [98.8%]). Mean age was 61±14 years, 48.2% were male, 74% were white, mean BMI 29±6. 36.8% of patients underwent LAR, 36.1% underwent PC, and 27.1% underwent RC. The most common surgical indication was colorectal cancer (48.9%), followed by chronic diverticular disease (19.2%), non-malignant polyp (11.5%), Crohn’s (5.3%), and acute diverticulitis (5.0%). 12.1% of surgeries were robotic, 54.0% laparoscopic, 0.2% hybrid, 33.7% were otherwise unspecified. Results of the univariate analysis are outlined in Table. There were no differences between groups in terms of 30-day rates of anastomotic leak, conversion, bleeding events, unplanned return to OR, mortality, unplanned readmissions or mortality (p>0.05). However, ICG-NIR use was on average associated with 32 minutes longer operative times (p < 0.001). After adjusting for potential confounders, ICG-NIR use was not independently associated with reduced adjusted odds of anastomotic leak. For oncologic surgeries, there were no differences between groups in terms of lymph node harvest (19 [15-26] vs 19 [14-26], p=0.322) or resection margin positive status (0% vs 0%, p=1.0).

Conclusion/Discussion: Major limitations of previous studies on ICG-NIR use include small sample size, and historical cohort comparisons, without accounting for improvements in surgical technique and colorectal ERAS protocols. This is the first large-scale study on ICG-NIR during colorectal surgery, showing that leak rates were not significantly lower, and in fact operative duration increased. Similarly, oncologic outcomes were not shown to improve. The added utility of ICG-NIR over pre-operative bowel preparation, good surgical technique, and gross intra-operative visualization may be overstated in this population and warrants further study.