P402 - Does White Light Imaging Contain Sufficient Information to Differentiate Perfusion-Ischemia Margins Like Indocyanine Green-Near Infrared Fluorescence in Colorectal Surgery?

Sunday, May 10, 2026

12:51 PM - 12:59 PM ET

Abstract Podium Presenter(s)

PS

Philip J. Seger, MD (he/him/his)

Research Fellow
University at Buffalo
University at Buffalo, Jacobs School of Medicine and Biomedical Sciences
Buffalo, NY, United States

All disclosed relationships have been reviewed and managed prior to the start of this activity. Any relevant relationships will be listed below. This individual has no financial relationships to disclose

Author(s)

Jessica Martinolich, MD

Assistant Clinical Professor
University at Buffalo
Buffalo, New York, United States

Disclosure information not submitted.
Joseph Sciacca, DO

Colorectal Fellow
University at Buffalo
University at Buffalo
Buffalo, NY, United States

Disclosure information not submitted.
JM

Joseph D. Mills, MD

Associate Program Director
University at Buffalo
West Seneca, New York, United States

Disclosure information not submitted.
MF

Mark Falvo, MD

Assistant Professor Department of Surgery, Jacobs School of Medicine & Biomedical Sciences
State University of New York at Buffalo
EAST AMHERST, New York, United States

Disclosure information not submitted.
JV

Jeffrey Visco, MD

Assistant Professor of Surgery Jacobs School of medicine university at buffalo
Ubmd/ Kaleida health
Clarence Center, New York, United States

Disclosure information not submitted.
GC

Giles Carr-Locke, BA

Pre-medical student
N/A
Saint Louis, Missouri, United States

Disclosure information not submitted.
BF

Brendan T. Fox, BS

Medical Student
University at Buffalo Jacobs School of Medicine and Biomedical Sciences
Buffalo, New York, United States

Disclosure information not submitted.
SS

Steven D. Schwaitzberg, MD

Professor and Chair, Department of surgery
University at Buffalo, Jacobs School of Medicine
Buffalo, New York, United States

Disclosure information not submitted.

Senior Author(s)

PK

Peter Kim, MD

Professor of Surgery
SUNY at Buffalo
Buffalo, New York, United States

Disclosure information not submitted.

Purpose/Background: Accurate intra-operative assessment of the perfusion-ischemia margin (PIM) is a critical determinant of anastomotic outcome in colorectal surgery. Visual assessment of perfusion in white light imaging (WLI) is subjective and affected by surgeon's ability to distinguish minute differences in shades of red present in oxygenated and de-oxygenated tissue. Indocyanine Green-Near Infrared Fluorescence (ICG-NIRF) has been shown to be superior to WLI in visualizing the PIM. However, the use of ICG-NIRF is hampered by cost, workflow disruption, and seemingly marginal superiority. We hypothesize that standard WLI contains sufficient perfusion information to discern PIM equivalent to ICG-NIRF. Herein, we demonstrate the feasibility that enhanced WLI (e-WLI), via computer vision (CV) color enhancement, can detect and display PIM signal-to-background ratio (SBR) comparable to ICG-NIRF in left-sided colorectal surgery.

Methods/Interventions: Surgical videos from left sided colorectal surgeries were prospectively collected. Intra-operatively surgeons identified the PIM sequentially in WLI and ICG-NIRF after de-vascularization of the colon, prior to resection (n=10 assessments). e-WLI was then constructed using a three-step color augmentation in DaVinci Resolve (Blackmagic Designs): (1) white balance, (2) increased contrast magnitude, (3) hue-saturation color shift. Post hoc survey of intra-operative images to identify the PIM in WLI, e-WLI, and ICG-NIRF was conducted with general and colorectal surgeons (n=5) blinded to surgery. PIM of WLI and e-WLI were compared against ICG-NIRF PIM +/- 5 mm concordance zone and concordance was measured. The visual impact of CV enhancement was analyzed using imageJ for quantification of color intensity of ischemic and perfused bowel and SBR difference was calculated.

Results/Outcomes: The SBR difference in WLI of ischemic to perfused colon was 0.34 +/- 0.17. e-WLI (Figure 1B) demonstrated significant visually perceptible improvement in SBR difference, 1.13 +/- 0.89 compared to WLI (p=0.017). No significant difference in ischemic to perfused SBR was noted between e-WLI and ICG-NIRF. Ischemic bowel in e-WLI demonstrated a significant increase in blue intensity (p < 0.05) and decrease in red intensity (p < 0.05) when compared to perfused. (Figure 2) The intra-operative surgeon perception of concordance rate of the PIM in WLI to ICG-NIRF was 0.4 with an average discordant distance of 1.96 +/- 1.22 cm. Post hoc surveys had a concordance rate of 0.32 and average discordant distance of 2.33 +/- 0.98 cm (p= 0.21 vs. intra-operative). e-WLI had a greater concordance rate to ICG-NIRF, 0.46, compared to both the intra-operative and blinded surgeons. The average discordant distance for e-WLI to ICG-NIRF was 1.59 +/- 0.68 cm, a 0.73 cm improvement over baseline WLI (p=0.0016).

Conclusion/Discussion: This work shows for the first time that e-WLI contains sufficient information even in the early ischemic period (30 +/- 10 minutes) to detect the PIM comparable to ICG-NIRF. The CV e-WLI detected and displayed a significant change in red and blue color intensities, enhancing the SBR difference of ischemic to perfused colon, previously imperceptible in WLI. Further, we demonstrate that visual perception of perfusion SBR in e-WLI and ICG-NIRF were similar. This CV approach of enhancing pixel level differences and quantifying color intensity at the PIM allows for a more objective assessment of perfusion with the potential for scalable machine-learning based decision support in the operating room.

Figure 1. Ischemic to Perfused Signal-to-Background Ratio in Left Colectomy

I. This panel demonstrates the colonic perfusion-ischemia margin in baseline WLI (A) and eWLI (B). It highlights the significant improvement in ischemic (signal) to perfusion (background) ratio for eWLI compared to baseline WLI (p=0.017). There was no statistically significant difference in eWLI SBR compared to ICG-NIRF

II. This panel demonstrates the quantitative changes of red and blue color intensity in ischemic (diamond) and perfused (circle) colon. Ischemic colon in eWLI demonstrated a statistically significant decrease in red (p <0.05) and increase in blue (p <0.05) intensity compared to perfused colon.
Figure 2. Perfusion-Ischemia Margins in White Light Imaging (WLI), Enhanced WLI (eWLI), and Indocyanine Green-Near Infrared Fluorescence

I. This panel demonstrates the intra-operative appearance of the perfusion-ischemia margin (PIM) in white light imaging (WLI), computer vision generated enhanced WLI (eWLI), and Indocyanine Green-Near Infrared Fluorescence (ICG-NIRF). The PIM (solid line) and surrounding 5 mm concordance zones (dotted lines) proximally and distally are overlayed on each respective image.

II. This panel demonstrates the concordance rate of WLI and eWLI to ICG in the post-hoc surgeon survey. There was a statistically significant improvement in average discordance distance, 0.73 cm (p=0.0016) in eWLI compared to WLI.