A Novel Injection Protocol Using Voluven®-Assisted Indocyanine Green with Improved Near-Infrared Fluorescence Guidance in Breast Cancer Sentinel Lymph Node Mapping—A Translational Study

All reagents were purchased from Sigma‒Aldrich, Alfa Asear, TCI, and Thermo Fisher and used as received unless indicated elsewhere. Highly pure water (18.2 MΩ·cm) was used in all of the experiments. ICG for injection (25 mg/vial) was purchased from Daiichi Sankyo Propharma Co. Ltd. (Japan). Voluven® (1000 ml/bag) was obtained from Fresenius Kabi Deutschland GmbH (Germany). Flexbumin human serum albumin USP (25% solution, 50 ml/bag) was purchased from Bazalta US Inc. (IL, USA). For absorption and fluorescence measurements, three stock solutions were prepared: (1) an ICG solution with a concentration of 2.5 mg/ml using normal saline as the solvent, (2) an ICG solution with a concentration of 2.5 mg/ml using Voluven® as the solvent, and (3) an ICG solution with a concentration of 2.5 mg/ml using 10% HSA as the solvent. The ICG solutions were diluted in the three different solvents to obtain ICG solutions with concentrations of 0.25, 0.125, 0.0625, 0.05, 0.0313, 0.025, 0.0167, and 0.0125 mg/ml. This allowed measurement of the absorbance and fluorescence intensity (Fig. 1).

The clinical trial was approved by the Institutional Review Board (IRB) of National Taiwan University Hospital Hsin-Chu Branch in accordance with the ethical standards of the Helsinki Declaration of 1975. It was designed as a dose-adjustment trial and was registered at clinicaltrial.gov prior to the recruitment of the participants (NCT05365204). The off-label use of ICG and Voluven® for breast cancer SLN mapping was evaluated and approved by the IRB. All participants were Taiwanese women with early breast cancer eligible for SLNB who provided written informed consent before the surgery. The exclusion criteria were known allergy to conventional tracers, known allergy to ICG, or any change of the disease resulting in the patient no longer being indicated for SLNB. The trial included a feasibility test (phase I) of using ICG:Voluven® for breast cancer SLNB, and a dose-adjustment trial for best imaging quality (phase II). The default ICG:Voluven® concentration of 0.25 mg/ml was selected based on the results from laboratory experiments and animal models. We stepwisely tested each new concentration with a group of three patients. After each group of three patients was tested, we performed image analysis. The new concentration group was stopped if the outcome (see the Section “Endpoints” below) was significantly worse than the previous concentration group. This translational research is the result of the phase I study.

During the surgical procedure, a repetitive injection-observation protocol was implemented (see Fig. 2). A 5-ml syringe was filled with the prepared ICG solution and connected to a winged infusion set. The needle of the winged infusion set delivered a subareolar injection (see Fig. 3). With the Stryker SPY-PHI (Stryker Corporation, Kalamazoo, MI) portable handheld NIR imaging system, 0.5 ml of the ICG solution was infused slowly. The filling of the solution into the areola and drainage to the subcutaneous lymphatic system was observed. If the whole areola was filled and the lymphatic duct traveling toward the axilla was observed, the winged infusion set was fixed in place using sterile 3M tape. If the lymphatic duct traveling toward the axilla was not observed within 2 min, the needle tip position was adjusted to another quadrant by the surgeon. The injection and observation were repeated. Light palpation of the breast was performed if draining seemed stationary. Injection example videos are provided in the supplementary files (see Vid. S1 for ICG:Voluven® and Vid. S2 for ICG:water).

Given that ICG is not yet an FDA-approved dye for breast cancer SLN mapping, an FDA-approved primary tracer was used, forming a dual-tracer mapping technique. Since the study goal was to evaluate the performance of the Voluven®-assisted ICG solution in SLN mapping of breast cancer, we carefully avoided synchronous injection of multiple tracers at the same location to prevent interaction between the drugs affecting the outcome evaluation. Peritumoral blue dye (Patent Blue V®) was selected as the primary (conventional) tracer for SLN mapping.

We usually waited for at least 5 min after blue dye injection and then initiated the SLNB procedure. All blue nodes were removed. ICG-fluorescent SLNs were retrieved sequentially, guided by the fluorescent lymphatic ducts. It is a known fact that ICG fluorescence guided SLN biopsy tends to retrieve more nodes than the conventional methods, and there are considerations of over-extensive dissection to the axilla.^7,16 Therefore, during the procedure, the first three fluorescent nodes receiving drainage from the breast were retrieved. If additional sequential nodes were observed during NIR fluorescence imaging, the first three nodes were sent for frozen section analysis. Further dissection was not carried out if the frozen section was reported to be benign.

All lymph node images were taken using the SPY-PHI NIR fluorescence mode (grayscale SPY mode) with a medical HDMI video recorder. The SBR values of the lymph nodes were measured by ImageJ bundled with Java 1.8.0_172 software. R version 4.2.2 (R Foundation, Vienna, Austria) was used for plotting.

The primary endpoint was the number of retrieved lymph nodes and SBR under the SPY-PHI imaging system. Other endpoints included the areola-to-axilla traveling time (AAT) of ICG fluorescent lymphatics (the time from the development of subcutaneous lymphatic fluorescence to the fluorescence reaches the axillary fascia), the status of blue dye and ICG fluorescence in retrieved SLNs, and any adverse effects of this injection procedure.

We measured the absorption and fluorescence spectra of ICG at concentrations of 0.25 mg/ml and 0.0125 mg/ml in various environments, including saline, Voluven®, and 10% HSA (Fig. 1a–c). These experimental findings offered insights into the ACQ effect of ICG and its emission intensities, which are concentration-dependent in different environments. Additionally, we investigated the dilution ratio factor of the three stock solutions (0.25 mg/ml), as depicted in Fig. 1d. Table 1 provides a summary of the fluorescence intensities for each data point depicted in Fig. 1d.

Fluorescence imaging of ICG:Voluven® and ICG:saline in Eppendorf tubes was conducted using a 1300-nm longpass filter and an exposure time of 500 ms (Fig. S1a). Clear visualization of whole-body blood vessels was achieved with NIR imaging with ICG:Voluven® as well as ICG:water (Fig. S1b, see Supplementary Information, SI) within 1 min of injection. In the in vivo imaging in mice (Fig. S1b and c), ICG was intravenously injected into live mice, bearing 4T1 tumors, through the tail vein. ICG:Voluven® and ICG:water were administered to three mice each. Within 1 min of the injection, ICG dispersed throughout the entire body, enabling clear visualization of the vascular structures in the live mice. As early as 1 min postinjection, ICG uptake in the liver, kidney, and vessels was detected. However, ICG:Voluven® provided greater vessel conspicuity and enabled prolonged visualization of vasculature for up to 5 min, while with ICG:water, blood vessels were not visualized beyond 2 min. The signals from background soft tissues and the liver gradually decreased due to quick elimination, accompanied by increasing signals in the tumor. Furthermore, the SBR was consistently higher for ICG:Voluven® than for ICG:water at all time points (Fig. S1c). At 1 min postinjection, the SBR values were 1.58 for ICG:Voluven® (Fig. S1d) compared with 1.43 for ICG:water (Fig. S1e). At 5 min postinjection, the SBR of ICG:water dropped to less than 1, while that of ICG:Voluven® remained higher at 1.35. After 24 h, the mice were sacrificed, and isolated tissues, including those of the pancreas, liver, spleen, kidney, and tumor, were collected (Fig. S1f and g). ICG fluorescence was predominantly concentrated in the kidneys, and stronger fluorescence was observed in tumors from mice treated with ICG:Voluven® compared with ICG:water.

In total, nine patients undergoing SLNB for breast cancer were enrolled from April 2022 to August 2022. Six patients underwent a repetitive injection-observation protocol using ICG:Voluven® (0.25 mg/ml). Three patients were administered ICG:water (0.25 mg/ml) for comparison. The patient characteristics and information about tumor typing and tumor staging are shown in Table 2. All patients were female, and their ages ranged from 34 to 80 years. Their BMIs ranged from 15.1 to 34.9. The invasive tumors were T1 and T2 lesions, ranging in size from 8 to 32 mm. One patient had one SLN pathologically positive for metastasis.

All patients had lymphatic vessels successfully identified under NIR fluorescence. No adverse reactions were observed intraoperatively or reported by patients postoperatively. The type and volume of injected solution, AAT, measured SBRs, and pathological status of the retrieved SLNs are all listed in Table 3. The SBRs of the retrieved nodes were measured (Fig. S2 as an example) and are plotted in Fig. 4. The volume of the injected solution ranged from 1.5 to 2.5 ml. The AAT ranged from 163 to 196 s for ICG:water and 21 to 133 s for ICG:Voluven®. Except for one patient (V2) with metastasis detected in the first SLN, three SLNs were retrieved from all other patients. In the case of patient V1, the needle tip was inadvertently inserted too deeply into the breast parenchyma, leading to leakage of the ICG solution from the nipple (Fig. S3). The mapping of the SLNs was not affected after adjustment of the needle tip position. One patient (W1) had no blue nodes identified. As per the trial design, the three ICG:water participants presented obviously worse results in the endpoints (lymph node SBRs and AATs) compared with ICG:Voluven®. Thus, no more participants were subsequently invited into the ICG:water group.

We observed a significant adverse ACQ effect of ICG at concentrations exceeding 0.25 mg/ml, as evidenced by very low fluorescence intensities across all three different solvents (Fig. 1 and Table 1). Even under such concentrated conditions, the fluorescence of ICG in 10% HSA exhibited 8–8.5 times higher intensity compared with saline and Voluven®. This phenomenon can be attributed to the noncovalent binding between the dye and proteins, facilitated by hydrophobic van der Waals interactions as well as ion pairing between the sulfonate groups of ICG and cationic amino acid residues such as histidine, lysine, and arginine.^18,19 Similar behavior has been observed for cyanine derivatives with sulfonate moieties in solutions containing serum.²⁰ The mechanism underlying this phenomenon involves the hindrance of dye aggregation by protein binding, thus maintaining a rigid dye conformation that minimizes torsional effects and, consequently, reduces nonradiative decay. When the concentration of ICG was significantly diluted to 0.0125 mg/ml, the emission intensities of ICG showed a substantial increase. Specifically, the fluorescence intensities increased by approximately 5 and 5.2 orders of magnitude for saline and Voluven®, respectively. In contrast, a smaller increase in ICG fluorescence, less than 5 orders of magnitude, was observed for 10% HSA. The observed increase in fluorescence relative to the dilution ratio (Fig. 1d) suggests that hydrophobic HES plays a similar role to HSA in relation to ICG. Notably, concentrated ICG in water exhibited significant H-aggregates, which was evident by the strong absorption shoulder at ~700 nm (Fig. 1a).²¹ On the other hand, when concentrated ICG was mixed with Voluven® and 10% HSA, a similar absorption profile was observed, suggesting the absence of aggregation (Fig. 1b and c). This finding again suggests that the HES in Voluven® effectively prevents aggregation of ICG and exhibits behavior similar to that of HSA.

From the brightness measurements of ICG:Voluven® and ICG:water in Eppendorf tubes, both at a concentration of 0.25 mg/ml (Fig. S1a), the emission brightness of ICG:Voluven® was ~1.5 times higher than that of ICG:water, which is consistent with the spectroscopic results (Table 1). Interestingly, we found that the SBR and the spatial resolution of blood (Fig. S1d and e) were notably improved in mice treated with ICG:Voluven® compared with those injected with ICG:water. This finding verified our proposed concept that ICG has a tendency to adhere to the HES colloids in Voluven®, consequently reducing its diffusion into neighboring tissues and minimizing background noise levels. More strikingly, from the biodistribution of ICG in major organs 24 h after injection, it was observed that ICG:Voluven® exhibited specific accumulation within the 4T1 tumors. This targeting ability could be attributed to the enhanced permeability and retention effect, a phenomenon that facilitates the accumulation of macromolecules, lipids, and nanoparticles (e.g., HES colloids) in tumor tissues.^22,23 The performance of ICG with tumor targeting ability by Voluven® could promote the widespread adoption of ICG:Voluven® formulations in fluorescence-guided surgeries utilizing tumor targeting probes.

We clinically translated the hypothesis that ICG could be “protected” from the ACQ effect and unwanted binding with serum proteins in the lymphatics when dissolved in Voluven® solution. The observed shorter AAT and higher SBR of retrieved lymph nodes in this study show improved transportation of the ICG when assisted by the Voluven® solution. In 2011, a dose-image optimization trial by Mieog et al.²⁴ was published, concluding that the ACQ effect at the injection site and the dilution within lymphatic channels were major contributors to the signal strength of the SLN. Our study provides a protocol with lower transportation loss, which means a better possibility for sequential mapping of more draining lymph nodes.

After the ACOSOG Z0011 trial, AMAROS trial, and SINODAR-ONE trial, current guidelines propose omitting axillary lymph node dissection when meeting certain criteria and fewer than three SLNs are metastasized.^25‐28 In 2022, Yuan et al.²⁹ revealed a novel approach, stepwise limited axillary lymph node dissection, as a de-escalated form of traditional axillary lymph node dissection. This procedure enabled further mapping beyond the conventional SLNs and was reported to have a significantly lower lymphedema rate with no obvious compromise to the oncological outcomes. We assume that our repetitive injection-observation protocol with ICG:Voluven® could be an easier substitute for the stepwise blue dye injection procedure in this type of surgery. Future research could be conducted for further de-escalation of axillary surgeries in breast cancer patients.