Extrahepatic cholangiography in near-infrared II window with the clinically approved fluorescence agent indocyanine green: a promising imaging technology for intraoperative diagnosis

Rationale: Biliary tract injury remains the most dreaded complication during laparoscopic cholecystectomy. New intraoperative guidance technologies, including near-infrared (NIR) fluorescence cholangiography with indocyanine green (ICG), are under comprehensive evaluation. Previous studies had shown the limitations of traditional NIR light (NIR-I, 700-900 nm) in visualizing the biliary tract structures in specific clinical situations. The aim of this study was to evaluate the feasibility of performing the extrahepatic cholangiography in the second NIR window (NIR-II, 900-1700 nm) and compare it to the conventional NIR-I imaging. Methods: The absorption and emission spectra, as well as fluorescence intensity and photostability of ICG-bile solution in the NIR-II window were recorded and measured. In vitro intralipid® phantom imaging was performed to evaluate tissue penetrating depth in NIR-I and NIR-II window. Different clinical scenarios were modeled by broadening the penetration distance or generating bile duct injuries, and bile duct visualization and lesion site diagnosis in the NIR-II window were evaluated and compared with NIR-I imaging. Results: The fluorescence spectrum of ICG-bile solution extends well into the NIR-II region, exhibiting intense emission value and excellent photostability sufficient for NIR-II biliary tract imaging. Extrahepatic cholangiography using ICG in the NIR-II window obviously reduced background signal and enhanced penetration depth, providing more structural information and improved visualization of the bile duct or lesion location in simulated clinical scenarios, outperforming the NIR-I window imaging. Conclusions: The conventional clinically approved agent ICG is an excellent fluorophore for NIR-II bile duct imaging. Fluorescence cholangiography with ICG in the NIR-II window could provide adequate visualization of the biliary tract structures with increased resolution and penetration depth and might be a valid option to increase the safety of cholecystectomy in difficult cases.

performed. For the capillary tube submerged 3 mm below the surface of 1% Intralipid ® solution, images with 900-, 1000-, 1100-, 1200-, 1300-nm long-pass filter (ThorLabs, USA) were also taken for further analyses, using a 793-nm excitation (power intensity, 10 mW/cm 2 ). Coomassie Brilliant Blue staining assay Coomassie Brilliant Blue staining was performed according to the standard protocol as previously described. Briefly, fresh bile was collected and protein concentration was determined using Thermo Scientific Pierce BCA Protein Assay Kit. Bile samples (60 μg per well) were separated by 10% sodium dodecyl sulfate-polyacrylamide (SDS-PAGE) gel electrophoresis. The gels were carefully removed and immersed in distilled water, heating to 100℃ for 30 seconds. Then the gels were immersed in prepared staining buffer and heated to 100℃ for 30-60 seconds. After washing the stained gels with cold DI water, distilled water was added and heated to 100℃ for 30-60 seconds. Lastly, repeating washing step several times and acquiring gel images by Bio-Rad ChemiDoc XRS+ System (1708265, California, USA).

Animal experiments
All the animal experiments in this work were conducted strictly in compliance with the requirements and guidelines of the Institutional Ethical Committee of Animal Experimentation of Zhejiang University. Institute of Cancer Research (ICR) mice (6-8 weeks old, female) and Sprague-Dawley (SD) rats (6-8 weeks old, female) were provided from the SLAC laboratory Animal Corporation. (Shanghai, China) and housed in the Laboratory Animal Center of Zhejiang University (Hangzhou, China). The animal housing area was maintained at 24℃ with a 12 hours light/dark cycle, with free access to water and food. Before each imaging experiment, mice or rat was anesthetized via intraperitoneal injection of 2% pentobarbital (40-50 mg/kg), sufficient depth of anesthesia was maintained, redosing with one-fifth the original dose of pentobarbital only as necessary. The animal was secured to a platform in the supine position, with the ventral side facing the camera and laser source. Laparotomy was performed before imaging, fully exposing the surgical field of interest. Fluorescent agent ICG at a concentration of 0.5 mg/kg was intravenously injected through tail vein before imaging, using a 793-nm excitation. For dynamic NIR-II fluorescence cholangiography, lower dose of fluorophore (0.025 mg/kg) was injected via the inferior vena cava before imaging in order to obtain complete fluorescence curve of extrahepatic biliary tract. Acute peritonitis mouse model establishment The method to establish experimental peritonitis model using glacial acetic acid has previously been reported [1,2]. Adult female mice were randomly divided into two groups (n = 3/group). Mice of experimental group were intraperitoneally injected with 0.15 ml of 2% glacial acetic acid in normal saline. Animals of the control group were given the same volume of normal saline. The general state of health was observed. At 24 hours post glacial acetic acid or saline administration, the animals were anesthetized for collection of blood from eye socket, after which the mice were sacrificed by cervical dislocation and the abdominal tissue were removed for the following experiment. The proportion of neutrophils in the collected blood sample were analyzed by flow cytometry. Moreover, the histopathological changes by H&E (hematoxylin-eosin) staining and CD45 expression analysis (a cell surface antigen specially expressed in neutrophils) in the generated inflammatory tissue and normal tissue were examined for confirmation of establishment of acute peritonitis model.

Human bile collection
The human bile samples used in this study were obtained from three patients with permission whose bile had been continuously drained for at least 5 days after laparoscopic common bile duct exploration. The study had been approved by the Clinical Research Ethics Committee of Sir Run Run Shaw Hospital of Zhejiang University.

Supplementary figures
Figure S1 Schematic illustration for NIR-II fluorescence macroscopic imaging system.

Figure S3 NIR-II emission of ICG-human bile solution.
Absorbance spectra of (A) ICG-human bile solution and (B) ICG-water solution in 550-900 nm wavelength range at concentration of 5, 10, 15 μg/ml. Red dashed bars in each figure shows the absorbance peak. Fluorescence emission profile of ICG in human bile (magenta line) and water (blue line) normalized to equimolar concentration between (C) 900 and 1600 nm wavelength region, (E) 1000 and 1600 nm wavelength region. The absolute fluorescence quantum yield measurement of ICG-water solution (blue column) and ICG-human bile solution (magenta column) between (D) 900 and 1600 nm wavelength region, (F) 1000 and 1600 nm wavelength region.

Figure S6 Intralipid ® phantom study of ICG-water solution in NIR-I and NIR-II windows.
Fluorescence images in (A) NIR-I window and (B) NIR-II windows of glass capillary filled with ICG-water solution (31.25 μg/ml) at depths of 0, 2, 3, 4, 5 and 6 mm in 1% Intralipid ® solution. (C) The capillary filled with ICG-water solution was submerged in 3 mm of 1% intralipid ® solution and imaged again with 800-nm long-pass NIR-I detector and 900-, 1000-, 1100-, 1200-, 1300-nm long-pass NIR-II detector. FWHM was calculated for capillary glass tube filled with ICG-water solution (D) at varying depths of 1% Intralipid ® solution or (E) with different long-pass filters at the same depth of 1% Intralipid ® solution. (F) Capillary tube filled with fresh rat bile in the NIR-II window.

Figure S12 Comparison of SBR for NIR-I and NIR-II cholangiography in rats
Comparison of SBR for NIR-I and NIR-II cholangiography in rats. (A) Quantitative analysis of the SBR of NIR-I and NIR-II cholangiography in no tissue phantom models (n = 10). Quantitative analysis of the SBR of NIR-I and NIR-II sub-surface bile duct imaging in (B) normal tissue phantom models (n = 9), (C) adipose tissue phantom models (n = 10) and (D) acute inflammatory tissue phantom models (n = 9). Quantitative analysis of the SBR of NIR-I and NIR-II sub-surface biliary lesions imaging using normal tissue phantom in (E) acute bile duct obstruction models (n = 6) and (F) acute bile duct transection models (n = 6).
Movie S1 Dynamic imaging of extrahepatic bile duct peristalsis and bile excretion into the duodenum in NIR-II window with a 1000-nm long-pass filter in a mouse model.