Preclinical evaluation of a fluorine-18-labeled indocyanine green analog for PET imaging of hepatic function Authors: Khashayar Vakili, Alan Packard Published: 2026 ABSTRACT BackgroundAccurate assessment of liver function and biliary excretion is a key aspect in the medical and surgical management of various liver conditions. Dynamic nuclear imaging techniques such as PET (positron emission tomography) offer the potential to directly quantify hepatic clearance and hepatobiliary function. Indocyanine green (ICG) is a fluorescent dye that is used to indirectly evaluate liver function via its plasma clearance, but its optical properties preclude deep tissue imaging. Here, we report the synthesis and preclinical evaluation of a novel fluorine-18 labeled ICG derivative ([18F]SFB-ICG; SFB = N-succinimidyl-4-[18F]fluorobenzoate) as a potential PET radiotracer for noninvasive assessment of hepatobiliary function.ResultsICG was successfully labeled with 18F: The product was obtained in high chemical and radionuclidic purity, albeit in relatively low yield (0.25–1.0%, non-decay-corrected). PET imaging demonstrated rapid hepatic uptake with a peak standardized uptake value (SUV) of 8.6 ± 0.8 at 20 min post-injection. The tracer exhibited prompt biliary excretion as demonstrated by a gallbladder SUV of 22.3 ± 1.4 at 45 min post-injection. Time–activity curves confirmed hepatic clearance similar to unlabeled ICG. At 4 h PI, the SUVs of selected organs were: liver 4.7 ± 0.5; lungs 0.15 ± 0.02; small intestine 3.68 ± 0.76; spleen 2.25 ± 0.88; large intestine 4.9 ± 1.7. Renal uptake and clearance was minimal. The whole blood SUV at 1 and 4 h post-injection was 0.47 ± 0.04 and 0.20 ± 0.02, respectively. No overt adverse effects were observed during the 4-h study period.ConclusionsThis study presents [18F]SFB-ICG as a novel PET tracer for real-time, quantitative assessment of liver function and biliary clearance in vivo. The tracer exhibits favorable pharmacokinetics, high liver specificity, and excellent correlation with established hepatobiliary physiology. Thus, 18F-SFB-ICG represents a lead compound that warrants further development—including yield optimization, stability and toxicity characterization, and evaluation in animal models of liver disease—for potential applications in hepatic functional imaging, surgical planning, characterization of liver masses, and assessment of biliary pathology. Background Imaging modalities such as ultrasound, computed tomography (CT), technetium-99 m hepatobiliary iminodiacetic acid (99mTc-HIDA) scans, and magnetic resonance imaging (MRI) have substantially advanced our ability to detect structural and some functional abnormalities relating to the liver. These modalities play an important role in the medical and surgical management of various liver conditions. However, one of the limitations of these methods is their inability to quantify hepatic function and biliary excretion at high resolution (Duan et al. 2022). A complementary approach has emerged from radiotracers targeting the asialoglycoprotein receptor (ASGPR), a hepatocyte-specific transmembrane glycoprotein whose surface density correlates with functional hepatocyte mass and is reduced in acute and chronic liver injury. 99mTc-GSA SPECT, has been used clinically for decades—primarily in Japan—to estimate functional hepatic reserve prior to resection (Torizuka et al. 1991; Ha-Kawa et al. 1991). More recently, 68Ga-labeled PET analogues such as 68Ga-Dolacga (68Ga-NOTA-hexalactoside) and 68Ga-NOTA-GalNAc have been developed, with 68Ga-Dolacga having completed Phase II clinical evaluation against CT volumetry and the ICG retention test for measurement of liver reserve in surgical candidates, though no results have been reported. Additional trials in hepatocellular carcinoma have been registered but not yet recruiting based on the information available on ClinicalTrials.gov. These ASGPR-targeted probes report on hepatocyte receptor abundance, providing a static measure of functional hepatocyte mass. The ICG-based approach we describe here is mechanistically distinct: it interrogates the dynamic, transporter-mediated uptake and biliary excretion pathway that underpins the clinical ICG retention test, and is therefore expected to provide complementary information to receptor-density imaging. To address these limitations, we have developed an 18F-labeled ICG analog—[18F]SFB-ICG—for positron emission tomography (PET) imaging of hepatic function. In contrast to other modalities, PET offers superior sensitivity, superior tissue penetration, and quantifiable three-dimensional functional imaging. ICG is a water-soluble tricarbocyanine dye that is exclusively taken up by hepatocytes and excreted into bile without enterohepatic circulation or metabolic transformation (Reinhart et al. 2016). Clinically, ICG retention tests—such as ICG-R15 and plasma disappearance rate—are widely used to evaluate hepatic functional reserve prior to liver resection and transplantation (Reinhart et al. 2016; Olmedilla et al. 2009). Additionally, ICG-based fluorescence imaging has emerged as a valuable tool for intraoperative surgical guidance, offering real-time visualization of liver segments, biliary structures, and tumor margins (Lim et al. 2014). By developing an ICG-based PET radiotracer, we aim to preserve ICG’s hepatocyte-specific uptake and biliary excretion profile while enabling noninvasive, quantitative, and whole-body imaging of hepatic function and biliary clearance. This molecular imaging approach has the potential to improve the assessment of liver function, enhance preoperative planning, and facilitate individualized management of hepatobiliary disease. Methods Radiochemistry Fluorine-18-labeled ICG was prepared using an N-succinimidyl-4-[18F]fluorobenzoate (SFB) prosthetic group (Vaidyanathan and Zalutsky 2006; Wester et al. 1996) on a GE FX-FN synthesis module as outlined in Fig. 1.Fig. 1Synthesis of [18F]SFB-ICG Synthesis of [18F]SFB-ICG Materials Fluorine-18 was produced on a cyclotron using the 18O(p,n)18F reaction and purchased from PETNET. The 4-(ethoxycarbonyl)-N,N,N-trimethylbenzenaminium trifluoromethanesulfonate precursor was purchased from Asta Tech, the amine-functionalised ICG derivative (hereafter “ICG-amine”) was purchased from AAT Bioquest (catalogue no. 188), and the non-radiolabeled (19F) SFB-ICG standard was prepared by Sai Life Science. All other reagents were purchased from MilliporeSigma and used as received. Synthesis [18F]Fluoride was trapped on a Sep-Pak® Light Accell™ Plus QMA Carbonate cartridge (Waters) and eluted into the reaction vessel using K2CO3 (3.5 mg in 0.2 mL H2O) and Kryptofix 2.2.2 (15 mg in 1.0 mL acetonitrile). The solution containing the eluted activity was heated at 70 °C under nitrogen flow and vacuum for 2 min, and the temperature was then raised to 100 °C for 2 additional minutes. The nitrogen flow was then turned off, and the activity was dried under vacuum for an additional 8 min yielding anhydrous Kryptofix 2.2.2-K[18F]F. The SFB precursor (4-(ethoxycarbonyl)-N,N,N-trimethylbenzenaminium trifluoromethanesulfonate, 5.0 mg in 0.5 mL anhydrous DMSO) was added to the reaction vessel and heated at 125 °C for 10 min then cooled to 120 °C. Tetrapropylammonium hydroxide (TPAH, 20.0 μL 1 M dissolved in 1 mL anhydrous acetonitrile) was added and reaction vessel was heated at 120 °C for 3 min. The reaction vessel was then cooled to 90 °C under nitrogen flow for 2 min and vacuum for 3 min. TSTU (N,N,N′,N′-tetramethyl-O-(N-succinimidyl)uronium tetrafluoroborate; 15 mg in 1 mL anhydrous acetonitrile) was added and reaction vessel was heated at 90 °C for 5 min followed by cooling to 35 °C. The reaction mixture was then diluted with 1% acetic acid (15 mL) and passed through Sep-Pak® C-18 Plus Light cartridge which retained the [18F]SFB precursor. The collected [18F]SFB was eluted from the cartridge with a mixture of acetonitrile and DMSO (3:1, 1 mL) into a second reaction vessel. ICG-amine (2 mg in 0.2 mL anhydrous DMSO) and DIEA (diisopropylethylamine, 40 µL in 0.2 mL acetonitrile) were added to the second reaction vessel, and the reaction was incubated at room temperature for 30 min. The crude [18F]SFB-ICG was diluted with HPLC buffer (2 mL) and purified using the semi-preparative HPLC (Column: XBridge BEH C18, 130 Å, 5 μm, 250 × 10 mm; Eluent: Water (0.1% TFA):MeCN (50:50), Isocratic; Flow rate: 5 mL/min; Detector: UV 280 nm). The retention time of the product is 18–20 min. The fraction containing purified [18F]SFB-ICG was collected in a flask containing 50 mL of water. The diluted [18F]SFB-ICG solution was then passed through a Strata® C18-E cartridge, which retained [18F]SFB-ICG. The cartridge was washed with water (5.0 mL), and [18F]SFB-ICG eluted with ethanol (0.5 mL) followed by PBS (4.5 mL) into a formulation vial. The final product was filtered through a sterilizing filter (0.22 µm) into the final product vial, and a sample was removed for QC testing. The product was characterized by analytical HPLC (Column: XBridge BEH C18, 130 Å, 5 μm, 4.6 mm X 150 mm; Eluent: Water (0.1%TFA):MeCN (45:55), Isocratic; Flow Rate: 1 mL/min; Detector: UV 280 nm) using the non-radiolabeled (19F) compound as a reference. The retention time (min) of the product, [18F]SFB-ICG, was 8.5 min. The retention time of the SFB prosthetic group was 2.0 min, and the retention time of the ICG-amine precursor was 4.2 min. Quality control in this study was performed by radio-HPLC and UV-HPLC only; radio-TLC was not included. Although [18F]fluoride elutes near the solvent front under our HPLC conditions and the radiochemical purity of the final product was consistently > 90% by HPLC, radio-TLC will be added to the standard QC panel in ongoing optimization work as an independent confirmation of the absence of free [18F]fluoride. Animals Female C57BL/6 mice, 8—14 weeks old, were purchased from Charles River Laboratories (Wilmington, MA). The mice were anesthetized with isofluorane (4–5%) for injection and maintained at 1–2% isofluorane for imaging. A blood sample was collected via the submandibular vein immediately after each PET/CT acquisition (1 and 4 h post-injection) for gamma counting. Dose administration Each animal was weighed on the day of dose administration. The animals ranged in weight from 19.8 to 21.5 g. Single intravenous (IV) doses of the test compound were administered via the lateral tail vein using a 1 mL syringe to provide a injected volume of 5 mL/kg containing 6–12 MBq (180–320 µCi) of [18F]SFB-ICG. The syringe was assayed in a dose calibrator before and after injection, and the activity administered to each animal was calculated from the difference in the activity in the syringe pre- and post-injection. PET/CT imaging While under anesthesia, mice were placed prone in a 2-bed imaging hotel. Whole body PET data was collected in list mode using a Siemens Focus 220 camera. Imaging was initiated immediately prior to injection of the test compound for a total 60 min from time of injection (TOI). Static images were obtained for 30 min at 4 h post-injection. CT images (Mediso, nanoScan) were acquired immediately following PET image acquisition. Image reconstruction was performed using 2D Ordered Subset Expectation Maximization (OSEM 2D) with framing for dynamic imaging as follows: 10 × 30 s, 10 × 1 min, 6 × 5 min, and 1 × 15 min. Image processing Reconstructed images were generated in units of activity. Namely, the values assigned to the voxels (volume elements) comprising the 3D reconstructed PET images were in units of kBq or equivalent. Reconstructed images were co-registered to their respective CT image, resampled to 0.2 mm3 voxels, and cropped to a uniform size prior to analysis. The imaging bed was removed from the CT image and the injected activity was adjusted (normalized) to the average activity in the FOV in the final 6 frames of the dynamic PET acquisition. Estimating tissue uptake Regions of interest (ROIs) for the heart, lung, spleen, liver, gallbladder, small intestines, large intestines, and cecum were defined using VivoQuant software (Perceptive, Needham, MA). The heart was defined by fitting a fixed volume organ phantom in each CT image. The tracer uptake in the lungs and spleen was sampled by placing fixed volume spheres (2 and 3 spheres, respectively) in the appropriate locations as guided by CT. Based on a combination of both the PET and CT images, the liver, gallbladder, small intestines, large intestines, and cecum were segmented manually as completely as possible to capture all tracer signal as appropriate. Activity concentrations were reported as Standardized Uptake Values (SUV); for fully segmented organs, uptake was reported in percentage injected dose (%ID). Image generation Using VivoQuant, the data acquired was used to generate a variety of images. Using the frames of the dynamic PET acquisition, time-weighted MIP movies scaled 0–100% ID/g were generated. MIPs scaled 0–100% ID/g were generated using the final 15-min frame of the dynamic acquisition as well as the 30-min static acquisition at 4 h. Gamma counting analysis The activity of each collected blood sample was measured in units of counts per minute (CPM). Aliquots of the radiotracer were assayed (in triplicate) in the gamma counter to calculate the conversion factor for converting counts per minute to units of activity (CPM/kBq). Values were decay corrected to the time of injection and corrected for background radiation. Results Radiosynthesis of [18F]SFB-ICG The purified product was obtained in 0.25 to 1.0% non-decay-corrected yield with a total synthesis time of 90 min. For this proof-of-concept study, no efforts were made to optimize the yield. The identity of the 18F-labeled product was confirmed by comparison of the HPLC retention time to that of the 19F reference compound. The radiochemical purity of the final product was > 90% and the chemical purity was also > 90%. The average molar activity was 160 GBq/µmol (4300 mCi/µmol). Biodistribution [18F]SFB-ICG was predominantly localized to the liver within 5 min following injection (Fig. 2). The characteristics of animals and administered doses are listed in Table 1. Standardized Uptake Value (SUV) in the whole blood was 0.47 ± 0.04 (mean ± esd) at 1 h post-injection and 0.20 ± 0.02 at 4 h post-injection. Liver reached a peak mean SUV of 8.64 ± 0.76 at 15 min post-injection. At 4 h post-injection, the liver mean SUV was 4.72 ± 0.50. The percent injected dose (%ID) in various organs as a function of time is presented in Table 2. There was no appreciable uptake by the kidneys at any time point. Activity in blood samples at 1 and 4 h post injection is shown in Fig. 3. During the 4-h imaging period, no overt signs of distress, abnormal behavior, or respiratory compromise were observed in any of the animals. Because no body-weight follow-up, serum chemistry, hematology, or histopathology was performed and animals were not observed beyond 4 h, this does not constitute a formal toxicity evaluation and no conclusions can be drawn regarding sub-acute or delayed toxicity.Fig. 2[18F]SFB-ICG biodistribution. A Representative PET images of two mice at indicated time points during the 60-min dynamic phase (White arrow points to the presence of the tracer in the small intestine) and four mice at 4 h post-injection. B Percent injected dose (%ID) in various organs versus time. C Standardized uptake value (SUV) in the liver of four mice (the dark blue curve represents the mean values of the four other curves with error bars shown in B. Note: in one of the four animals, the apparent slight increase in liver SUV between 45 min and 4 h reflects repositioning of the animal between the dynamic and static acquisitions—a focal high-uptake region present at 45 min is sampled differently within the manually segmented liver ROI in the 4-h frame—rather than a true increase in hepatic activityTable 1Mouse characteristics and injected activityIDSexWeight (g)Injected activity (MBq/µCi)1Female19.912.1/3262Female19.811.7/3163Female21.56.63/1794Female20.36.40/173Table 2Biodistribution of [18F]SFB-ICG as percentage of injected dose (%ID)Time post-injection (min)Cecum (%ID)Gallbladder (%ID)Heart (%ID)Large intestine (%ID)Liver (%ID)Small intestine (%ID)000.51 ± 0.234.97 ± 0.45011.53 ± 1.680.39 ± 0.08503.96 ± 1.711.59 ± 0.44063.63 ± 5.250.90 ± 0.341506.20 ± 2.540.95 ± 0.27067.57 ± 2.691.59 ± 0.723008.53 ± 3.18 0.84 ± 0.25065.43 ± 2.373.16 ± 1.074509.51 ± 3.950.76 ± 0.23060.47 ± 1.656.77 ± 1.9824029.09 ± 5.251.22 ± 0.160.24 ± 0.026.09 ± 4.6235.99 ± 3.053.77 ± 0.84Fig. 3Percent injected dose per gram (%ID/g) of 18F-SFB-ICG in whole blood [18F]SFB-ICG biodistribution. A Representative PET images of two mice at indicated time points during the 60-min dynamic phase (White arrow points to the presence of the tracer in the small intestine) and four mice at 4 h post-injection. B Percent injected dose (%ID) in various organs versus time. C Standardized uptake value (SUV) in the liver of four mice (the dark blue curve represents the mean values of the four other curves with error bars shown in B. Note: in one of the four animals, the apparent slight increase in liver SUV between 45 min and 4 h reflects repositioning of the animal between the dynamic and static acquisitions—a focal high-uptake region present at 45 min is sampled differently within the manually segmented liver ROI in the 4-h frame—rather than a true increase in hepatic activity Table 1Mouse characteristics and injected activityIDSexWeight (g)Injected activity (MBq/µCi)1Female19.912.1/3262Female19.811.7/3163Female21.56.63/1794Female20.36.40/173 Mouse characteristics and injected activity Table 2Biodistribution of [18F]SFB-ICG as percentage of injected dose (%ID)Time post-injection (min)Cecum (%ID)Gallbladder (%ID)Heart (%ID)Large intestine (%ID)Liver (%ID)Small intestine (%ID)000.51 ± 0.234.97 ± 0.45011.53 ± 1.680.39 ± 0.08503.96 ± 1.711.59 ± 0.44063.63 ± 5.250.90 ± 0.341506.20 ± 2.540.95 ± 0.27067.57 ± 2.691.59 ± 0.723008.53 ± 3.18 0.84 ± 0.25065.43 ± 2.373.16 ± 1.074509.51 ± 3.950.76 ± 0.23060.47 ± 1.656.77 ± 1.9824029.09 ± 5.251.22 ± 0.160.24 ± 0.026.09 ± 4.6235.99 ± 3.053.77 ± 0.84 Biodistribution of [18F]SFB-ICG as percentage of injected dose (%ID) Percent injected dose per gram (%ID/g) of 18F-SFB-ICG in whole blood Discussion [18F]SFB-ICG represents a promising lead compound for the development of PET radiotracers for noninvasive liver function assessment. Our preclinical findings demonstrate that this tracer exhibits rapid hepatic uptake, sustained liver retention, and hepatobiliary excretion with negligible renal clearance. These pharmacokinetic features qualitatively resemble those of native ICG, which has been used for decades as a clinical standard for evaluating hepatic perfusion and clearance capacity (Reinhart et al. 2016). The rationale for using ICG analogs in liver imaging is well supported by the established physiology of ICG itself: After intravenous injection, ICG binds almost completely to plasma proteins and is selectively extracted by hepatocytes via glutathione S-transferases, followed by biliary excretion without undergoing metabolism (Engel et al. 2008). In healthy individuals, ICG is not excreted in the urine and does not cross the blood brain barrier or the placental barrier (Sakka 2018). Native ICG has a serum half-life of 3–4 min with hepatocyte uptake via solute carrier proteins serving as the rate limiting step (Landsman et al. 1976; Paumgartner and Reichen 1975). This mechanism is the basis for the ICG plasma disappearance rate (ICG-PDR) assay, which remains a clinically accepted dynamic test for liver function assessment (Paumgartner 1975). ICG has been shown to have a longer serum half-life in individuals with liver disease (Leevy et al. 1967). In fact, ICG elimination kinetics were demonstrated to be independent predictors of postoperative mortality following liver resection (Hemming et al. 1992; Nonami et al. 1999). In addition, several studies have demonstrated the use of ICG as a predictor of mortality and organ dysfunction following liver transplantation (Olmedilla et al. 2009; Schneider et al. 2011; Hori et al. 2006; Levesque et al. 2009). While traditional ICG clearance measurements are useful, they are limited by operator variability, reliance on serial blood sampling or transcutaneous sensors, and lack of spatial and anatomic information (Sakka 2018). In contrast, the use of PET imaging with [18F]SFB-ICG provides dynamic, quantitative, and spatially resolved insights into liver function in vivo. These advantages align with emerging clinical needs as highlighted by Duan et al., who emphasized the critical role of noninvasive imaging in evaluating hepatic dysfunction, particularly in early or subclinical stages where conventional serum biomarkers and grading systems lack sensitivity (Duan et al. 2022). Compared to hepatobiliary MRI using agents like gadoxetate, which also reflect hepatocyte transport and biliary excretion, [18F]SFB-ICG PET may offer superior temporal resolution and quantitative kinetic modeling, without requiring specialized liver-dedicated MR protocols (Duan et al. 2022). One of our key findings was the fact that [18F]SFB-ICG demonstrates similar pharmacokinetics to ICG despite the addition of the FSB moiety. This was shown by rapid (under 5 min) liver uptake of [18F]SFB-ICG, reaching a maximum concentration of 68% of the injected dose at 15 min following injection. Subsequent rapid excretion into the biliary system was demonstrated by the presence of the tracer in the gallbladder by 5 min post-injection. The transit of the tracer through the intestinal tract is demonstrated by activity first appearing in the small intestine, followed by cecum and the large intestine. [1⁸F]SFB-ICG retains the qualitative hepatobiliary profile of native ICG—selective hepatic uptake, biliary excretion, and negligible renal clearance—but the kinetics are quantitatively slower. Native ICG in humans has a plasma half-life of 3–4 min and a plasma disappearance rate of approximately 18–25%/min, reflecting first-pass hepatic extraction of greater than 70% per circulatory passage (Desmettre et al. 2000) which is comparable to mouse pharmacokinetics (Zhang et al. 2018). By contrast, [18F]SFB-ICG shows substantially slower kinetics in our hands: the whole-blood activity declines from approximately 2.3%ID/g at 1 h to approximately 1.0%ID/g at 4 h, hepatic retention remains high at 4 h (35.99 ± 3.05%ID; gallbladder 1.22 ± 0.16%ID), and there is concurrent activity accumulation in the small intestine, large intestine, and particularly the cecum (29.09 ± 5.25%ID at 4 h). Summation of the measured ROIs at 4 h accounts for approximately 76% of the injected activity; the remainder is attributable to fecal excretion during the 1–4 h inter-imaging interval (animals were not housed in metabolic cages) and to unsampled tissues (carcass, urinary bladder contents, skin). This pattern is consistent with intact hepatobiliary clearance proceeding at a slower rate than that of native ICG and continuing well beyond the 60-min dynamic acquisition. The most likely explanation for the slowed kinetics is the addition of the fluorobenzoyl moiety, which increases molecular weight and lipophilicity and may reduce affinity for the OATP-family uptake transporters and/or for MRP2-mediated canalicular efflux that mediate native ICG handling. The slower clearance is, importantly, well matched to the 110-min physical half-life of 18F and lends itself naturally to static late-time-point PET acquisition in addition to dynamic uptake-rate measurement. It is also instructive to compare [18F]SFB-ICG to the rapidly developing class of ASGPR-targeted PET tracers, in particular 68Ga-Dolacga, which has progressed furthest toward clinical use (Hua and Yu 2024). 68Ga-Dolacga PET data is dependent on receptor-density imaging to predict functional liver reserve. Therefore, ASGPR-mediated and ICG-handling pathways report on biologically distinct aspects of hepatocyte function: receptor expression on the one hand, and OATP/MRP2-mediated transport and biliary excretion on the other. In several disease states—most notably cholestasis and drug-induced transporter dysfunction—these pathways may be discordantly affected, and an ICG-based PET probe could therefore provide diagnostic information that is complementary to 68Ga-Dolacga. From a translational standpoint, 18F (t1/2 = 110 min) confers logistic advantages over 68Ga (t1/2 = 68 min) for satellite distribution and for late-time-point imaging of biliary transit. Realizing the diagnostic potential of [18F]SFB-ICG will require establishment of dedicated PET acquisition and analysis protocols, in analogy to those developed over the past three decades for 99mTc-GSA SPECT and now extended to 68Ga-Dolacga PET. Based on the kinetics observed in the present study, a clinical protocol would likely combine a short dynamic acquisition (0–30 min) over the liver to derive hepatic uptake-rate parameters analogous to ICG-K and to the 99mTc-GSA HH15 and LHL15 indices, together with a late static acquisition (e.g. 60–120 min) to quantify biliary transit and to derive whole-liver and segmental functional volumes. Population-input-function approaches or image-derived input functions from cardiac chambers may obviate the need for arterial sampling. Establishment of normal-range diagnostic parameters and their correlation with established benchmarks—ICG-R15, MELD, Child–Pugh score, and gadoxetate-enhanced liver-function MRI—will be a necessary step prior to clinical adoption. The non-decay-corrected radiochemical yield of 0.25–1.0% obtained in this proof-of-concept study is modest and will require systematic optimization prior to any clinical translation. The two-step SFB conjugation strategy is well established but inherently lossy: yield is limited at both the 18F-fluorination/activation step that produces [18F]SFB and at the subsequent acylation of the ICG-amine precursor. We anticipate specific strategies will improve overall yield in future work. Based on the observed pharmacokinetics of [18F]SFB-ICG in this study and the current uses of ICG in the clinical setting, we hypothesize that [18F]SFB-ICG can be a valuable addition to the current imaging modalities used for medical and surgical management of liver conditions. Specifically, we envision that [18F]SFB-ICG may have utility in assessment of the following: liver function, perfusion, liver mass, bile leak, or biliary obstruction. In combination with CT imaging, it can provide a new approach for functional and anatomical imaging of the liver. Although our study provides a synthesis methodology for [18F]SFB-ICG and foundational pharmacokinetic data, further validation in disease models—such as fibrosis, cholestasis, liver tumor, biliary leak, or drug-induced liver injury—is essential to demonstrate sensitivity to pathological alterations. In conclusion, this work establishes that [18F]SFB-ICG is a promising lead compound for further development of PET tracers for hepatobiliary imaging. By combining the biochemical specificity of ICG with the quantitative power of PET, this approach addresses an important gap in dynamic liver function and biliary imaging. Important next steps include radiochemical-yield optimization, formal stability, toxicity, and dosimetry studies, and evaluation in animal models of hepatic disease (fibrosis, cholestasis, focal hepatic lesions, biliary obstruction, drug-induced liver injury) prior to consideration of clinical translation. Conclusions This proof-of-concept study establishes [18F]SFB-ICG as a lead compound for 18F-PET imaging of hepatobiliary function. The tracer preserves the qualitative hepatic-uptake and biliary-excretion profile of native ICG, exhibits negligible renal clearance, and was well tolerated by all animals during the 4-h observation period of this study. Several developmental steps remain—optimization of radiochemical yield, characterization of plasma protein binding and in-vivo stability, formal toxicity and dosimetry studies, and most importantly evaluation in animal models of hepatic disease (fibrosis, cholestasis, focal hepatic lesions, biliary obstruction, drug-induced liver injury)—before the tracer’s diagnostic utility relative to existing modalities, including 68Ga-Dolacga PET, 99mTc-HIDA SPECT, and gadoxetate-enhanced MRI, can be established. The work nevertheless provides a foundation for further development of a PET tracer that, uniquely among hepatic functional imaging agents under investigation, directly recapitulates the biology of the clinically validated ICG retention test.