Shanglei Liu, MD1, David Owens2, Domingo Montalvo, MD1, Toshiaki Suzuki, MD1, Garth R Jacobsen, MD1, Jonathan C DeLong, MD1, Jessica J Reynolds, MD1, Ran B Luo, MD1, Bryan J Sandler, MD1, Santiago Horgan, MD1. 1UCSD Healthcare Systems, 2UCSD Department of Computer Science
Introduction: Indocyanine Green (ICG) fluorescence imaging has gained popularity in recent years to evaluate intraoperative tissue perfusion. However there are no standards across platforms and almost no relevant numerical measurements. As a result, surgeons are often unable to decipher the clinical significance of intra-operative fluorescent images. In this study, we investigate the possibility of using time based relative intensity scale to obtain quantitative data in measuring ischemia.
Method: For this study, we developed a novel computer algorism for filtering and analyzing focal ICG signal as a function of time. We tested the algorism in pig models by serially ligating blood supply from the stomach in 20-minute intervals followed by ICG angiography. This data is compared to pig undergoing sham surgery with the same ICG intervals. Imaging data is collected on the greater curve of the 15cm below the GE junction. Ligation patterns are: the gastroepiploic arteries (1st Ligation), short gastric arteries (2nd ligation), left gastric artery (3rd ligation), and right gastric artery via distal gastrectomy (4th ligation). Then we retrospectively examined ICG angiography footage from patients undergone sleeve gastrectomy to evaluate the fluorescent signal at the resection staple line.
Results: The ICG signal intensity of pig stomach after serial blood vessel ligation is seen in the following graph after polynomial fit. Descriptive values of the curve are in the following table.
These distinct numerical trends are reproducible when compared to the sham surgery pig at each 20 minute mark.
In human, we evaluated the before and after ICG footage of a patient undergoing sleeve gastrectomy using the same software. The tissue near the staple line had a 45% increase in time to maximum (9.5 to 13.8 seconds), a 35% decrease in max rate of signal increase (17.5 to 11.2), and a 22% (5.1 to 6.3) increase in time to max rate of signal increase, similar to findings of mild ischemia in the pig model.
Conclusion: We present a reliable method for analyzing relative ICG signal to produce quantitative measurements of tissue perfusion. This program can be used on any digital image file and is therefore not limited by commercial devices. Larger population study will be required to find out the exact numeric cutoffs for clinical significant outcomes.
Presented at the SAGES 2017 Annual Meeting in Houston, TX.
Abstract ID: 79957
Program Number: P425
Presentation Session: Poster (Non CME)
Presentation Type: Poster
