Optimizing Indocyanine Green Dosage for Near-Infrared Fluorescence Perfusion Assessment in Bowel Anastomosis: A Prospective, Systematic Dose-Ranging Study

Background: Indocyanine green (ICG) near-infrared fluorescence (NIRF) has emerged as a promising technique for visualizing tissue perfusion. However, within the wide range of dosages and imaging conditions currently being applied, the optimal dosage of ICG remains unclear. This study aimed to investigate the feasibility and implications of implementing lower dosages of ICG than commonly used for visual and quantitative perfusion assessment in a standardized setting. Methods: A prospective single-center cohort study was conducted on patients undergoing ileostomy reversal by hand-sewn anastomosis. ICG-NIRF visualization was performed before (T1) and after (T2) anastomosis with one of four different dosages of ICG (5 mg, 2.5 mg, 1.25 mg, or 0.625 mg) and recorded. Postoperatively, each visualization was evaluated for signal strength, completeness, and homogeneity of fluorescence. Additionally, perfusion graphs were generated by a software-based quantitative perfusion assessment, allowing an analysis of perfusion parameters. Statistical analysis comparing the effect of the investigated dosages on these parameters was performed. Results: In total, 40 patients were investigated. Visual evaluation demonstrated strong, complete, and homogeneous fluorescence signals across all dosages. Perfusion graph assessment revealed a consistent shape for all dosages (ingress followed by egress phase). While the average signal intensity decreased with dosage, it was sufficient to enable perfusion assessment even at the lowest dosages of 1.25 mg and 0.625 mg of ICG. The baseline intensity at T2 (the second intraoperative visualization) significantly decreased with dosage. The slope of the egress phase steepened with decreasing dosage. Conclusions: Lower dosages of ICG were sufficient for intraoperative perfusion assessment, while causing lower residual fluorescence and quicker egress in subsequent visualizations.