Intraoperative fluorescence quantification of organ perfusion: Insights for future applications

Indocyanine green (ICG) fluorescence imaging is increasingly used intraoperatively to assess tissue perfusion and guide surgical decision-making. While its qualitative use has been associated with improved outcomes, including reduced anastomotic leak rates, interpretation remains subjective and lacks standardization. Quantitative fluorescence analysis may improve reproducibility, precision, and broader clinical applicability.

Handheld intraoperative ICG fluorescence video from a representative bowel perfusion assessment was analyzed using three quantitative approaches: (1) a static region-of-interest (ROI) "steady lasso," (2) a manually adjusted "moving lasso," and (3) a semi-automated computational method. ROIs were applied to well- and poorly-perfused bowel segments, and mean pixel intensity over time was measured.

The steady lasso method demonstrated substantial signal misalignment due to tissue and camera motion, limiting reliability. The moving lasso improved spatial accuracy and demonstrated earlier signal rise and higher peak intensity in well-perfused bowel compared with poorly perfused tissue. The semi-automated method enabled high-throughput analysis and detailed characterization of fluorescence signal dynamics. Large ROIs demonstrated lower signal variability compared with small ROIs, particularly in poorly perfused tissue. Across methods, visually well-perfused tissue demonstrated faster time-to-peak and higher peak fluorescence.

Quantitative analysis of ICG fluorescence in handheld surgical video is feasible and reveals measurable differences in fluorescence signal behavior. While manual techniques improve accuracy over static approaches, semi-automated methods offer a higher degree of scalability. These findings highlight the need for automated ROI tracking and standardized acquisition protocols to support reproducible, real-time quantitative perfusion assessment in surgery.