A multi-modal embolic gel system for long-term fluorescence imaging and photothermal therapy

Abstract

Gel embolic agents are increasingly recognized for their versatility in minimally invasive vascular interventions. However, their application in real-time imaging, post-operative monitoring, and thermal treatment remains underexplored. In this study, we present a novel transcatheter injectable nanoclay-alginate (NCA) gel embolic agent integrated with indocyanine green (ICG) for dual fluorescence imaging and thermal ablation. The NCA/ICG embolic gel exhibits excellent shear-thinning properties, transcatheter injectability, and mechanical stability. Furthermore, the mechanism to enhance fluorescence for real-time imaging enhancement and extended post-operative monitoring was discussed. A 28-day fluorescence persistence shows the NCA/ICG gel’s long-lasting fluorescent signal, which was significantly stronger and longer compared to current clinically used ICG aqueous solution. Furthermore, the gel can effectively convert near-infrared (NIR) laser energy into heat for potential photothermal therapy. The biocompatibility and enhanced antibacterial properties further highlight the potential clinical benefits of this embolic agent as a multifunctional agent for vascular embolization.

1. Introduction

Minimally invasive procedures are widely used in clinical practice due to their superior outcomes compared to traditional open surgery, along with benefits such as reduced pain, shorter hospital stays, and fewer complications [1]. Among these, embolization, a minimally invasive vascular intervention, is used to stop bleeding or to block blood flow to abnormal/injured vasculature. Its advantages have led to its growing adoption as an interventional treatment [2,3] for conditions such as liver cancer, uterine fibroids, and aneurysms [4–8]. One of the important embolization treatment is the management of arteriovenous malformation (AVM), a condition where tangles of blood vessels create irregular connections between arteries and veins, bypassing the capillary system. Embolization helps reduce reflux into the feeding arteries and prevents premature embolization of the draining veins [9]. Thus, a nidus model can be a useful tool in developing embolic agents to ensure thorough occlusion of AVM.

A variety of embolic agents have been developed to address specific clinical and pre-clinical needs [7,10–12], including solids and liquids. Shear-thinning embolic gels have attracted notable attention due to their unique behavior, particularly their ability to flow like a liquid under shear force and revert to a solid-like state when the force is removed [13–16]. The shear-thinning behavior allows the gel for a versatile occlusion compared to fixed-size solid embolic agents. It also eliminates the toxicity risks associated with organic solvents used in liquid embolics [8,17].

We have previously developed laponite nanoclay/alginate (NCA) gel which showed promise as a shear-thinning embolic agent due to their excellent transcatheter injectability, mechanical properties, and biocompatibility [13]. Laponite nanoclay (NC) is a synthetic disc-shaped material with a diameter of 25 nm and a thickness of 1 nm [18]. Its unique charge distribution allows NC to form the “house of cards” structure in water, a microstructure that contributes to its shear-thinning property [18]. Alginate is a naturally occurring linear copolymer with an overall negative that exhibits shear-thinning properties [19]. Additionally, both NC and alginate are FDA approved materials known for their biocompatibility [20–22].

Imageability plays a crucial role in embolization procedures for ensuring precise occlusion to mitigate the risk of unintended embolization [23]. Indocyanine green (ICG), an FDA-approved fluorophore, has become a standard tool for intraoperative imaging due to its near-infrared (NIR) emission (~820 nm), deep tissue penetration, and well-established safety profile [24]. Its low toxicity and ability to provide real-time visualization make it highly valuable for surgical applications, including perfusion assessment and fluorescence-guided procedures.

While other FDA-approved fluorescent agents exist, they each have specific limitations [25]. Fluorescein sodium, primarily used in ophthalmology, has an emission peak (~520 nm) that overlaps with hemoglobin absorption, reducing fluorescence intensity and making it less effective for deep-tissue and vascular imaging. Methylene blue, although FDA-approved for lymphatic mapping, exhibits weak fluorescence due to its lower quantum yield, limiting its imaging efficiency. Other approved agents, such as 5-ALA (glioblastoma imaging), hexaminolevulinate (bladder cancer imaging), and pafolacianine (ovarian cancer imaging), provide tumor-specific targeting by selectively accumulating in cancerous tissues [25]. These fluorophores are restricted to specific indications and lack the broad intraoperative imaging versatility that ICG offers.

However, ICG also has limitations. First, ICG can be cleared rapidly by blood flow [26,27], limiting its imaging duration in vivo. Besides, the instability in aqueous solutions poses challenges such as decreased fluorescence intensity and lifetime [27–29]. Recent studies explored ICG embolic agents for real-time imaging and drug delivery [15,30–32], where only a limited amount of research focused on extending stability for long-term imaging. Clinical practice usually optimizes imaging results by controlling injection time and dosage, while it differs from case to case [24,33]. Studies have confirmed that aggregation is the main reason for fluorescent signal reduction [34,35] due to dynamic quenching [35,36]. Research has been done by enhancing ICG dispersion through supercritical fluids or dissolved in nonpolar solvent like ethanol, which carries a potential risk of cytotoxicity [35,37].

In addition to imageability, ICG can efficiently convert near-infrared (NIR) laser energy into heat with a high conversion efficiency of approximately 85 % [38]. This high efficiency makes ICG particularly effective for photothermal ablation, including those deeply seated blood vessels [38–40]. The multifunctional capabilities of ICG, which include improved fluorescent visibility and the ability to facilitate therapy, suggest its potential for effective embolization.

In this study, we report the development of a dual-functional NCA/ICG system for both embolization and visualization, with additional capability for potential photothermal therapy (PTT). We successfully improved the fluorescent signal intensity and endurance, as well as investigated the underlying mechanisms. The rheological properties of NCA/ICG were studied to understand their mechanical behaviors. Injection forces were tested to ensure an optimal experience for physicians. Biocompatibility was evaluated to support the safety of the agent. Laser heating was carried out for potential PTT. Overall, we seek to achieve accurate and stable occlusion, enhanced fluorescent signal for real-time visualization, improved ICG stability for extended post-operative assessments, and potential PTT with the newly designed NCA/ICG system.

2. Material and methods

2.1. Formulation of NC/alginate/ICG gel (NCA/ICG)

In this study, the NC/alginate/ICG gel (namely NCA/ICG) embolic agent was engineered using molecular-grade water (WFIMGW20L, Intermountain Life Sciences, UT, USA), Laponite-XLG nanoclay (NC) (BYK USA Inc., Gonzales, Texas, United States), medium viscosity alginic acid sodium salt derived from brown algae (Sigma-Aldrich, St. Louis, MO, USA), and indocyanine green (ICG) (Pfaltz & Bauer, Waterbury, CT, USA).

The specific composition of NCA/ICG embolic agents is listed in Table 1. ICG solutions were first made with varying concentrations. Subsequently, a 9 wt% NC/ICG gel was prepared by mixing NC and ice-cold ICG solution using a specialized mixer (DAC 330–100 SE, FlackTek SpeedMixer, Landrum, SC, USA). In parallel, a 2 wt% alginate/ICG gel was prepared by dissolving alginate in ICG solution at room temperature. Finally, various formulations were achieved by mixing the prepared 9 wt% NC/ICG and 2 wt% alginate/ICG.

Table 1.

Composition of NCA/ICG gels.

Sample name	NC (wt%)	Alginate (wt%)	Water (wt%)	ICG concentration in the final gel (mg/mL)
5NC0A/ICG0	5	0	95	0
5NC0A/ICG0.1	5	0	95	0.1
5NC0A/ICG0.3	5	0	95	0.3
5NC0A/ICG0.5	5	0	95	0.5
4.5NC0.5A/ICG0	4.5	0.5	95	0
4.5NC0.5A/ICG0.1	4.5	0.5	95	0.1
4.5NC0.5A/ICG0.3	4.5	0.5	95	0.3
4.5NC0.5A/ICG0.5	4.5	0.5	95	0.5

2.2. Rheology of NCA/ICG gels

The rheology properties of NCA/ICG gels were investigated using a rheometer (MCR 302e, Anton Paar, Graz, Austria) at ambient temperature (25 °C) to simulate clinical injection condition [41]. A 25 mm diameter sandblasted upper plate, and a sandblasted lower plate were used with a 1 mm gap in between for all tests. A solvent trap was used to maintain the chamber humidity to prevent gels from drying.

The shear-thinning behavior of the gels was examined through a shear rate sweep across shear strain, ranging from 10⁻³ s⁻¹ to 10³ s⁻¹. The gels’ strength and transition between gel and sol phases were assessed via large-amplitude oscillatory shear (LAOS) tests at a constant angular frequency of 10 rad s⁻¹. The yield strain was then determined by fitting two linear segments to the storage modulus data using the ordinary least squares method, denoted by a red cross at the intersection point in Fig. 1(a) [13]. The yield stress was then identified as the stress value corresponding to the yield strain (Fig. 1(b)). The loss factor (tan δ) was calculated as the ratio of the loss modulus (G″) to the storage modulus (G′) to assess the viscoelastic behavior of embolic agents. Both the shear rate sweeps and LAOS tests were conducted in triplicate.

Fig. 1.

An example of the linear fit of representative of LAOS to determine (a) yield point indicated by the red cross; (b) the yield stress is determined by the stress recorded corresponding to the shear strain through linear interpolation.

The time-dependent viscoelastic properties of the gels were investigated by frequency sweep at a fixed strain of 0.1 %, within the linear viscoelastic regime determined from LAOS, at an angular frequency ranged from 0.1 to 100 rad s⁻¹. Gel recoverability was assessed through thixotropic tests at an angular frequency of 10 rad s⁻¹. These gels were first broken down with strain oscillating at 100 % (high strain) for 2 min. Then, an alternating strain oscillation between 0.1 % (low strain) and 100 % (high strain) was applied for 2 min each, over a total duration of 16 min.

2.3. Microcatheter injectability

The injectability of NCA/ICG embolic agent through a 1 mL syringe (BD Luer-Lok^™ Syringe sterile, BD, USA) and a clinically used microcatheter (1.8F 150 cm, SuperCross^™ Microcatheter, Teleflex, USA) was measured by a universal testing machine (UniVert, CellScale, Waterloo, Ontario, Canada). The syringe was fixed and gradually pushed by the universal testing machine at a 1 mL min⁻¹ flow rate [42,43]. The injection force was determined as the mean value of the stable plateau within a certain range of piston displacement. Three tests were conducted for each sample.

2.4. Fluorescence imaging

Fluorescence signals of selected gels were evaluated by a microplate reader (Varioskan^™ LUX, Thermo Fisher Scientific Inc., Waltham, MA, USA) and an in-vivo imaging system (IVIS) (Xenogen, Perkin Elmer, Waltham, MA, USA). The embolic agents (5NC0A/ICG0.1, 4.5NC0.5A/ICG0.1) were loaded into a 96-well microplate (100 μL per well) and the fluorescence intensity was measured on days 0, 1, 3, 5, 7, 9, 14, 21, and 28. The samples were incubated at 37 °C to mimic physiological temperature. Microplate reader measurements were carried out at wavelengths of 780 nm excitation and 814 nm emission. For IVIS, an ICG filter (745 nm excitation and 840 nm emission) and a 1-s exposure time was used for imaging. Fluorescence intensity data obtained from IVIS were analyzed using the Living Imaging software (Perkin Elmer, Waltham, MA, USA).

2.5. Laser ablation

The laser ablation potential of the NCA/ICG gels, including 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1, was evaluated by heating the material with a near-infrared (NIR) laser. Equilibrium amounts (100 μL) of various ICG materials, along with their respective control groups (5NC0A and 4.5NC0.5A), were placed into a semi-micro cuvette (Fisher Scientific, Waltham, MA) with a transmittance around 0.9. A Ti:sapphire laser (Tsunami, Spectra Physics) was employed, generating a train of pulses at a repetition rate of 80 MHz. Each pulse had a temporal width of 250 fs; and the spectrum of pulses was centered at 808 nm. The laser’s average power intensity was kept at 2 W/cm², which corresponds to 1.8 W/cm² [44–46] focused on the gel with a beam diameter of 2 mm. Besides, a power of 5 W/cm² was performed to evaluate the gel’s photothermal performance under low power density [38]. The laser beam was directed onto the container for 5 min [44–46]. The temperature distribution along the cross-plane (laser heating) direction was measured using an IR camera (FLIR A655sc, Teledyne FLIR).

2.5.1. Tissue model for photothermal therapy (PTT)

A 1 wt% agar gel was prepared by dissolving agar powder (Fisher Scientific, Waltham, MA) in molecular-grade water under continuous stirring and heating until fully dissolved. A 200 μL of the solution was then pipetted into a semi-micro cuvette (Fisher Scientific, Waltham, MA) and left to solidify at room temperature overnight, forming a 1.3 mm uniform tissue-mimicking layer. After agar solidification, 100 μL of NCA/ICG gels or ICG aqueous solution were placed within the cuvettes for laser treatment, with an average laser power intensity kept at 2 W/cm².

2.6. Nidus model

The ability of 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1 embolic agents to penetrate confined spaces was studied using an in-vitro nidus model [47]. Specifically, glass balls with a diameter of 3 mm were loaded into a PVC tube with an inner diameter of 4.76 mm (3/16″) (Line 6516 T18, McMasterCarr, Atlanta, GA, USA). Embolic agents were loaded in a 1 mL syringe and gradually injected into the nidus model through a 1.8F 150 cm microcatheter at a flow rate of 1 mL min⁻¹ by the universal testing machine. Additionally, the fluorescent signals in the nidus model were captured using IVIS.

2.7. In vitro degradation

A degradation test was performed on 5NC0A, 4.5NC0.5A, 5NC0A/ICG0.1, and 4.5NC0.5A/ICG0.1. Approximately 200 μL of each gel was placed into a 1.5 mL Eppendorf tube, and its initial weight (W₀) was recorded. The gel samples were then incubated in 200 μL of PBS at pH 4, 7, and 9 at 37 °C (n = 4). At designated time points (1, 3, 5, 8, 24, 72, 120, 168, 240, and 336 h), PBS was removed, and the remaining gels were weighed to determine their final weight (W_R). The residual percentage of the gels was calculated using Eq. (1):

$$\mathit{\text{Residual}}\left(\%\right)=\frac{W_R}{W_0}\times 100$$ (1)

2.8. Shelf-life stability and injectability

Freshly prepared 5NC0A, 5NC0A/ICG0.1, 4.5NC0.5A and 4.5NC0.5A/ICG0.1 gels were stored at 4 °C to evaluate their stability over time. At 0, 7 and 15 days of storage, LAOS tests (n = 3) were performed at constant angular frequency of 10 rad s⁻¹. Additionally, injectability tests (n = 3) were conducted using a 1.8F 150 cm microcatheter (SuperCrossTM Microcatheter, Teleflex, USA), following the protocol described in Section 2.3.

2.9. Sterility and anti-bacterial test

The sterility and anti-bacterial properties of NCA/ICG0.1 gels were evaluated using Escherichia coli (E. coli, ATCC, 25922). Specifically, 5 mL of Luria-Bertani (LB) broth was added into a 15 mL centrifuge tube. Then, 1 mL 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1 were placed in the tubes respectively as the negative control. The same procedure was repeated for the treatment group, with a treatment of 1.6 × 10⁵ bacterial cells/mL for both 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1. The same bacteria concentration in LB alone was included as positive control. These samples were placed in an orbital shaker (C24 Incubator/Shaker, New Brunswick Scientific, Edison, NJ, USA) overnight at 37 °C and 100 rpm. The optical density of the supernatant was measured at the absorbance of 600 nm using a microplate reader (GENios, TECAN, Crailsheim, Germany). Three independent measurements were conducted per test.

2.10. Hemocompatibility

The hemocompatibility of 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1 was evaluated according to ISO-10993-4 [48]. In this study, 1 mL of each gel was added into a 15 mL centrifuge tube, followed by the addition of 9 mL of 1× Phosphate-buffered saline (PBS). These tubes were then pre-warmed at 37 °C for 30 min. Porcine whole blood (anticoagulated with sodium citrate, Lampire Biological Laboratories, Inc., Pipersville, PA, USA) was diluted with 1× PBS at a ratio of 4:5. Subsequently, 0.2 mL of diluted blood was added into each tube and incubated at 37 °C for 1 h. As negative and positive controls, 10 mL of PBS and 10 mL of deionized water were incubated with 0.2 mL of diluted blood, respectively. After the incubation, samples were centrifuged at 3000 rpm for 5 min and the supernatant was transferred to a 96-well plate. The absorbance (A) of the supernatant was measured using a microplate reader (GENios, TECAN, Crailsheim, Germany) at an absorbance wavelength of 550 nm. The hemolysis percentage was calculated according to Eq. (2).

$$\mathit{\text{Hemolysis}}\left(\%\right)=\frac{A_{\mathit{\text{sample}}}-A_{\mathit{\text{negative}}\mathit{\text{control}}}}{A_{\mathit{\text{positive}}\mathit{\text{control}}}-A_{\mathit{\text{negative}}\mathit{\text{control}}}}\times 100$$ (2)

2.11. Thrombogenicity

The thrombogenicity of selected gels was conducted to evaluate their interaction with blood. 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1 were loaded into a 96-well plate at a volume of 100 μL per well. The plate was then centrifuged at 1000 rpm to establish a uniform gel surface. The porcine blood (anticoagulated with sodium citrate, Lampire Biological Laboratories, Inc., Pipersville, PA, USA) was activated by gently mixing it with 10 % (v/v) 0.1 M CaCl₂. Then, 100 μL of the activated blood was added onto the top of gels. The gels were allowed to interact with the blood for designated time duration of 3, 4, 5.5, 7, and 8 min respectively. At selected time points, clotting was stopped by the addition of sodium citrate solution (0.109 M). Any uncoagulated liquid was immediately aspirated and washed repeatedly until the solution was clear, leaving only blood clots. Porcine blood alone and clinically used coils (2D Helical-35 micro-coils, Boston Scientific, MA, United States) were used as controls.

2.12. Cell viability

The biocompatibility of NCA/ICG0.1 was evaluated using a WST-1 assay (Cayman Chemical, Ann Arbor, MI, USA) [48] with L929 fibroblasts cells (ATCC, CCL-1, Lot #70008726) according to ISO-10993-5 [49]. The cell culture medium consisted of Dulbecco’s Modified Eagle Medium (DMEM, Gibco BRL, Grand Island, NY, USA), 10 % heat-inactivated fetal bovine serum (Cytiva, Marlborough, MA), and 1 % penicillin/streptomycin (Thermo Fisher Scientific, Inc., Waltham, MA).

To evaluate the biocompatibility of 5NC0A/ICG0.1 and 4.5NC0.5 A/ICG0.1, L-929 was treated with various concentrations of gel extracts (100 %, 50 %, 25 %, and 12.5 % v/v). L-929 mouse fibroblasts were seeded in 96-well plates at a density of 5000 cells per well and incubated for 24 h at 37 °C with 5 % CO₂ and 95 % humidity. For gel extraction preparation, 1 mL of each NCA/ICG0.1 gel was loaded into 10 mL of complete cell culture medium and incubated at 37 °C overnight. A series of extractions were obtained by carefully collecting the supernatant to achieve a dilution of 100 %, 50 %, 25 % and 12.5 % (v/v). A 100 μL of extraction was then added into each seeded well, following a 24-h incubation period. The viability of the treated cells was determined using the WST-1 assay according to the manufacturer’s instructions. DMSO (10 %) was used as positive control, while non-treated cells were used as the negative control. Six independent experiments with four replicates were conducted.

2.13. Statistical analysis

Statistical analysis was conducted by GraphPad Prism software (GraphPad Software, La Jolla, CA, USA). A t-test was performed to compare two groups. One-way analysis of variance (ANOVA) with Tukey’s comparisons method was performed for experiments containing more than two groups. Statistical significance in intergroup differences was assessed at the 95 % confident level (p < 0.05). Data is reported as average ± standard deviation (s.d.) unless otherwise noted.

3. Result

3.1. Rheology of NCA/ICG gels

Rheology studies were conducted to study the behaviors of engineered NCA/ICG gels. Hydrogels contained 5 wt% nanoclay (5NC0A), and those with 4.5 wt% nanoclay and 0.5 wt% alginate (4.5NC0.5A) were selected as base material according to our previous study [13]. A visual comparison of embolic agents with different base materials (5NC0A, 4.5NC0.5A) and varying ICG concentration (0, 0.1, 0.3, 0.5 mg/mL) are presented in Fig. 2(a). As the ICG concentration increased, the color of the gels gradually shifted from transparent to a deep green.

Fig. 2.

(a) Visual comparison of embolic agents. Rheological results of 5NC0A and 4.5NC0.5A gels with varying ICG concentrations, including (b) flow curves, (c) LAOS tests, (d) storage modulus, (e) yield strain, (f) yield stress strain, (g) loss factor, (h) thixotropy test and (i) angular frequency, Data are presented as mean ± standard deviation (n = 3 for (d–f)). ns, non-significant; *p < 0.1; ***p ≤ 0.001 ****p < 0.0001.

3.1.1. Shear rate sweeps

As shown in Fig. 2(b), the viscosity of all NCA/ICG gels decrease with increasing shear rate, confirming their shear-thinning behavior. The viscosity of 4.5NC0.5 gels have slightly higher viscosity than the 5NC0A gels with the same ICG content. Additionally, a linear viscoelastic region was identified at the shear rate from 10⁻² s⁻¹ to 10² s⁻¹ as indicated within the 2 dash lines.

3.1.2. Large amplitude oscillatory sweep (LAOS)

Large amplitude oscillatory sweep (LAOS) was conducted within the linear viscoelastic region of the gels to evaluate viscoelastic behaviors of NCA/ICG gels. Fig. 2(c) shows that within the shear strain ranges from 10⁻² to around 10⁰, both storage moduli (G′) and loss moduli (G″) remain constant, with the G′ surpassing G″. This indicates that the gels exhibit a solid-like behavior [15], characterized by a greater tendency for energy storage than dissipation. As shear strain exceeding 10⁰, G′ gradually decreases while the G″ exceeds G′, indicating a transition to liquid-like behavior.

LAOS results highlight differences in the mechanical response of two different gel formulations. In general, the 4.5NC0.5A group exhibited an overall higher G′ (Fig. 2(d)) and lower yield strain (Fig. 2(e)) compared to 5NC0A group. No significant differences (p > 0.05) were found in the storage modulus and yield strain across the various ICG concentrations, indicating that the addition of ICG does not significantly affect the storage modulus and yield strain of either gel formulation.

Fig. 2(f) represents the yield stress for both 4.5NC0.5A and 5NC0A gel group with varying ICG concentrations (0, 0.1, 0.3, and 0.5 mg/mL). In the 4.5NC0.5A gel group, statistically significant differences were observed between gels with different ICG concentrations. While for 5NC0A gels, only 5NC0A/ICG0 and 5NC0A/ICG0.3 have a significant difference in the yield stress (p < 0.05).

The dynamic viscoelastic behavior during sol-gel transition of NCA/ICG was analyzed using loss factor (tan δ) as the ratio of the energy lost to the energy stored (G″/G′). As shown in Fig. 2(g), tan δ values are below 0.1 when the shear strains are <10 %, indicating gels stabilizing with predominantly elastic behavior across a wide strain range. Subsequently, tan δ increases rapidly towards the sol-gel transition point (tan δ = 1), suggesting the transition from a solid-like state to a liquid-like state with internal energy dissipation [50,51].

3.1.3. Thixotropy testing

Thixotropy tests were conducted to evaluate the recoverability of the NCA/ICG embolic gels. As shown in Fig. 2(h), all gels exhibit their original G′ at the initial shear strain of 0.1 % (0–2 min). When the shear strain was increased to 100 % (2–4 min), a sharp drop in G′ was observed. As the shear strain reverted to 0.1 % (4–6 min), G′ rebounded and gradually recovered to a level near the initial G′. Through alternating high-strain (100 %) and low-strain (0.1 %) cycles (2–18 min), the gels exhibited a consist behavior between the break and recovery of the gel network. These results suggest potential recoverability of the gels under intermittent injection and delivery process during the intervention procedures.

3.1.4. Frequency sweep

The frequency sweep tests were conducted to evaluate the stability of NCA/ICG in response to angular frequencies. Fig. 2(i) showed consistently higher G′ compared to G″ for all NCA/ICG gels, suggesting the stable gel network across a wide range of angular frequencies. Furthermore, G′ is nearly independent of frequency, further indicating the stability of gel networks.

3.2. Microcatheter Injectability

In a typical transcatheter injection test (Fig. 3(a)), the injection force rapidly increased and followed by a plateau. The 5NC0A groups, with an injection force ranging from 22.3 ± 0.26 N to 25.1 ± 0.10 N, showed an overall lower force compared to the 4.5NC0.5A groups, ranging from 46.23 ± 1.27 N to 46.56 ± 1.01 N. Statistical analysis shows the addition of ICG has no significant influence on the transcatheter injection force (p > 0.05) in 5NC0A and 4.5NC0.5A groups.

Fig. 3.

(a) Injection force over time, and (b) the summary of injection force for 5NC0A and 4.5NC0.5A with varying ICG contents. Data are presented as mean ± standard deviation (n = 3 for (b)). ns, non-significant.

3.3. Fluorescence

Due to the instability of ICG in an aqueous solution, it is suggested to administer ICG as early as possible, ideally within 24 h before surgery [28,29], to ensure optimal effectiveness. Therefore, in this study, the fluorescence intensity of ICG dissolved in sterile water on day 1 (24 h) will serve as the benchmark value. The fluorescent intensity of embolic gels with varying ICG concentrations (0 to 0.5 mg/mL) in water, 5NC0A, and 4.5NC0.5 A is shown in Fig. 4(a) and (b).

Fig. 4.

Fluorescence analysis was conducted using signals detected by (a) microplate reader, (b) IVIS, and (c) images taken by IVIS. The analysis covered gel samples of 5NC0A/ICG, 4.5NC0.5A/ICG embolic gels. ICG aqueous solution is included representing current clinical method. Different ICG concentrations of 0, 0.1, 0.3, and 0.5 mg/mL were examined in this study. Data are presented as mean ± standard deviation (n = 3 for (a–b)).

3.3.1. Fluorescence analysis with microplate reader

The microplate reader results (Fig. 4(a)) showed that pristine gels and water (without ICG) exhibited minimal fluorescence signals as background control. For gel and water with ICG, the fluorescent signals gradually decreased over the first 3 days, followed by a sharp reduction until day 9, after which reading plateau. The NCA/ICG0.1 group yielded the highest signals overall, while increased ICG concentration (0.3 and 0.5 mg/mL) resulted in reduced fluorescence. Peak fluorescent intensity was detected on day 1, aligning with findings from previous studies [27–29].

It is important to note that NCA/ICG gels exhibited a significant enhancement in fluorescence signal compared to ICG aqueous solutions, which is the common practice in current clinical use. For instance, at an ICG concentration of 0.1 mg/mL on day 0, the signal for 5NC0A/ICG0.1 was the highest (354.53 ± 17.97), slightly surpassing that of 4.5NC0.5A/ICG0.1 (322.55 ± 8.90). These values were 3.65 and 3.32 times higher compared to those of 0.1 mg/mL ICG aqueous solution (97.12 ± 1.93), respectively.

Moreover, better fluorescence signal persistence was observed in NCA/ICG embolic gels compared to ICG aqueous solution over the 28-day monitoring. Generally, fluorescence signals decreased over time. The initial drop was sharp, with signals starting to plateau around day 9 for NCA/ICG embolic gels and day 3 for ICG aqueous solution. On day 1, the peak fluorescent intensities of ICG aqueous solutions were measured as 33.12 ± 3.11, 27.87 ± 3.20, and 11.15 ± 1.17 for ICG concentrations of 0.1, 0.3, and 0.5 mg/mL, respectively. These peak fluorescent intensities were set as the baseline in the study. Compared to these baseline measurements, stronger signals were detected for 4.5NC0.5A gels on days 5, 9, and 14, and for 5NC0A gels on days 9, 21, and 28. This indicates that the signal can persist for up to 28 days, which will be beneficial for post-monitoring.

3.3.2. Fluorescence analysis with IVIS

The quantitative analysis conducted using IVIS (Fig. 4(b)) validated the result in Section 3.3.1, confirming an enhanced and persistent signal effect for NCA/ICG gels compared to ICG solution. When comparing ICG solution and gels with identical ICG concentration on the same day, the fluorescent signal ranged from 4.8 to 55 folds for 5NC0A/ICG, and 4.37 to 48.9 folds for 4.5NC0.5A/ICG. Note that the signal saturation during imaging on day 0 and day 1 limited the recorded maximum values. Overall, a higher signal was also observed for 5NC0A/ICG gels compared to 4.5NC0.4A/ICG gels.

Fig. 4(c) confirmed that NCA gels and water without ICG had no detectable signal. For ICG aqueous solution at the ICG concentration of 0.1 and 0.3 mg/mL, signals became invisible by day 5, whereas the signal of 0.5 mg/mL ICG solution was undetectable by day 2. On the contrary, all NCA/ICG embolic gels were still detectable on day 28, showing persistent fluorescent signal.

Considering the minor influence of ICG addition in rheological properties and injectability, along with the strong and persistent signal intensity observed, an ICG concentration of 0.1 mg/mL was selected for further study, including both 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1.

3.3.3. Mechanism of fluorescence enhancement

The peak shift in absorption peaks in Fig. 5(a) evidenced the dominance of well dispersed ICG fluorophore due to the NCA system. The absorption peaks of ICG aqueous solution are at wavelengths of 700 nm and 780 nm (blue arrows), indicating the co-existence of H-aggregation and ICG monomer [35]. While the absorption peaks of NCA/ICG gels show a main absorption peak around 800 nm and a typical shoulder (orange arrows), indicating the dominance of ICG monomers [35]. Notably, the peaks do have a red-shift from 780 nm to 800 nm due to the rotation along the π-conjugated backbone of cyanine dye [35,52]. Furthermore, the absorption spectra shows that NCA/ICG0.1 exhibits an absorbance intensity three times stronger compared to 0.1 mg/mL ICG aqueous solution, indicating a higher energy absorption of NCA/ICG gels. Additionally, a figure of fluorescent spectra for 5NC0A, 4.5NC0.5A, water, 5NC0A/ICG0.1, 4.5NC0.5/ICG0.1, and ICG aqueous solution (0.1 mg/mL) is provided in Fig. S1, with a fluorescence peak around 830 nm for both NCA/ICG0.1 gels, and a much lower peak at 820 nm for ICG aqueous solution.

Fig. 5.

(a) Absorption spectra of the negative control, water, 5NC0A and 4.5NC0.5A, and treatment with 0.1 mg/mL ICG. (b) Embolic agents 5NC0A/ICG0.1 (top) and 4.5NC0.5A/ICG0.1 (bottom) were injected into the AVM mode, and observed by camera and IVIS. The embolic gels completely filled all confined spaces.

3.3.4. Nidus model

The IVIS tests demonstrate the applicability of NCA/ICG embolic agents in a clinical visualization setup. Fig. 5(b) illustrates the penetration ability of embolic agents, specifically 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1, in an arteriovenous malformation (AVM) model [47]. The artificial nidus was made of 3 mm glass balls in a PVC tube. Both gels completely filled the artificial nidus, accompanied by the detection of fluorescent signals.

3.3.5. Photothermal therapy (PTT)

In general, higher laser power and longer laser exposure time lead to more aggressive ablation. To evaluate the photothermal ablation potential of the NCA/ICG gels, NIR laser heating was conducted using the setup shown in Fig. 6(a). In this study, an effective laser power of 1.8 W/cm² was chosen based on previous research [38,44–46,53,54]. Usually, the treatment time ranges from a few minutes to hours [55]. In this study, the ablation time of 5 min was selected based on the point at which the temperature plateaued.

Fig. 6.

(a) Experimental setup for the NIR laser heating, where the width of cuvette surface is 12 mm, (b) temperature rise distribution across the surface of the cuvette after 5 min of heating, and (c) thermal images captured by the IR camera after 5 min heating. The length of the cuvette (12 mm) is marked.

The temperature profiles (Fig. 6(b)) and representative NIR images (Fig. 6(c)) showed differences in temperature distribution along the laser heating direction. In particular, the low standard deviation of temperature, highlighted by the shaded regions in Fig. 6(b), indicating consistent heating across multiple runs.

For gels without ICG (5NC0A/ICG0 and 4.5NC0.5A/ICG0), the temperature remained nearly constant with minimal rise, as shown in both the temperature profiles (Fig. 6(b)) and the uniform distribution in the corresponding NIR images (Fig. 6(c)). In contrast, both ICG aqueous solutions and NCA/ICG gels demonstrated a significant temperature increase under laser irradiation (Fig. 6(b)). The ICG solution exhibited a non-uniform temperature raise within 5 min under a 2 W/cm² laser power, reaching a peak increase of 11.8 ± 0.3 °C. The NCA/ICG gels demonstrated a more uniform heating pattern, with 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1 achieving temperature increases of 8.1 ± 0.3 °C and 7.4 ± 0.3 °C, respectively. This suggests that the NCA/ICG gels provide a more controlled and consistent heating profile.

Increasing the laser power to 5 W/cm² further raised the maximum temperature in both 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1, demonstrating a power-dependent heating effect (Fig. 6(b)). The higher laser intensity led to a more pronounced thermal response, with temperature increases from 8.1 °C to 10.6 °C for 5NC0A/ICG0.1, and from 7.5 °C to 8.6 °C for 4.5NC0.5A/ICG0.1. These results confirm the photothermal efficiency of NCA/ICG gels under varying irradiation conditions.

The NIR images (Fig. 6(c)) further illustrate the surface temperature distribution after laser exposure. A clear and intense hotspot was observed for ICG solution, indicating a more concentrated heat zone. In contrast, deeper penetration of the laser heating effect with a more uniform temperature distribution were observed for NCA/ICG gels. These low variations in heat distribution of NCA/ICG are beneficial for better PTT control.

The temperature distribution of NCA/ICG gels during laser heating depends on the presence of a tissue model layer (Fig. 7(a)). The tissue layer, modeled by agar, created a lower temperature zone on the tissue side. As shown in Fig. 7(b), samples with a tissue layer exhibited reduced temperatures on the tissue side, whereas samples without a tissue layer maintained a more uniform temperature distribution across the laser direction. Additionally, the inclusion of the tissue layer also resulted in lower maximum temperatures increase (ΔT) for both 5NC0A/ICG0.1 (7.7 ± 0.1 °C) and 4.5NC0.5A/ICG0.1 (6.7 ± 0.1 °C). The temperature increase (ΔT) was 0.4 °C and 0.7 °C lower than the samples without a tissue layer respectively. Moreover, the temperature rise was higher in 5NC0A/ICG0.1 compared to 4.5NC0.5A/ICG0.1, with a difference of 1.0 °C in the presence of tissue layer, and 0.7 °C without the tissue layer. Gels without ICG showed no photothermal response, whereas those containing ICG exhibited a temperature increase.

Fig. 7.

(a) Thermal images captured by the IR camera with and without a tissue mimicking layer under 2 W/cm², and (b) temperature distribution during laser heating with and without a 1.3 mm tissue model layer.

3.4. Shelf-life evaluation of the gels

3.4.1. Degradation

An in vitro degradation study was carried out to evaluate gels’ behavior over time. All gels exhibited an initial increase in weight, with residual masses percentages exceeding 100 % within the first 8 h, suggesting swelling (Fig. 8). This effect was more pronounced in alginate-containing formulations (4.5NC0.5A and 4.5NC0.5A/ICG0.1), especially under acidic conditions (pH 4). Following this initial swelling, the gels showed gradual mass loss over time, confirming their degradation behavior. After 14 days (336 h), all gels retained over 90 % of their mass across all pH conditions.

Fig. 8.

Degradation test showing residual percentages of (a) 5NC0A, 5NC0A/ICG0.1, and (b) 4.5NC0.5A, 4.5NC0.5A/ICG0.1 (d) at pH = 4, 7, and 9 at 1, 3, 5, 8, 24, 72, 120, 168, 240, and 336 h. Data are presented as mean ± standard deviation (n = 4).

5NC0A (Fig. 8(a)) demonstrated the highest stability, with the highest residual mass of 97.4 % at pH 9, and 91.8 % at pH 4. In contrast, 4.5NC0.5A (Fig. 8(b)) exhibited a more pronounced initial swelling effect, particularly at pH 4, where residual mass percentages exceeded 114.0 % at 8 h. It then degraded faster compared to 5NC0A, reaching 93.6 % remaining mass at pH 9 by 14 days.

Similar to 5NC0A and 4.5NC0.5A, ICG-containing gels (Fig. 8(a, b)) followed a pH-dependent degradation trend, with faster degradation in acidic conditions (pH 4) and greater stability in alkaline conditions (pH 9). By 14 days, 5NC0A/ICG0.1 retained 90.3 % of its mass at pH 4 and 98.8 % at pH 9. The 4.5NC0.5A/ICG0.1 hydrogel showed a similar pattern, with residual masses ranging from approximately 92.8 % to 98.9 % across different pH conditions.

3.4.2. Long term storage modulus

The gels exhibited progressive strengthening over time. Yield stress increased across all formulations, ranging from 3.5 % to 22.9 % on day 7 and from 20.0 % to 34.1 % on day 15 (Fig. 9(a)). The yield stress ranged from 57.1 ± 1.9 Pa to 79.4 ± 6.0 Pa on day 1 and from 76.1 ± 3.2 Pa to 95.3 ± 6.7 Pa on day 15. Similarly, the storage modulus increased by 22.6–36.4 % on day 7 and 41.9–49.2 % on day 15, indicating continued structural reinforcement during storage (Fig. 9(b)). The storage modulus ranged from 790.2 ± 6.4 Pa to 1329.8 ± 15.8 Pa on day 1 and from 1168.7 ± 34.4 Pa to 1960.3 ± 45.5 Pa on day 15, demonstrating enhanced rigidity and stability over storage time.

Fig. 9.

Shelf-life evaluation of gels over 15 days of storage, including (a) yield stress, (b) storage modulus, (c) injection force.

3.4.3. Long term injectable forces

The injection force of the gels through the microcatheter increased over time during storage. The maximum injection force among all gels was 46.40 ± 1.5 N on day 7 and 47.40 ± 0.4 N on day 15 (Fig. 9(c)). However, all injection forces remain below 60 N (dashed line), ensuring they are within the acceptable range for human capability and comfortable injections.

3.5. In-vitro characterization of NCA/ICG gels

3.5.1. Sterility and antibacterial properties

Sterility and antibacterial properties play a crucial role in reducing the risk of infections during medical procedures. The sterility and antibacterial properties of 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1 gels were evaluated using E. coli, with optical density at 600 nm (OD600) measured to determine bacterial growth (Fig. 10(a)). The gels showed OD600 values for 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1 are 0.001 ± 0.002 and 0.076 ± 0.001 respectively, indicating their inherent sterility.

Fig. 10.

In-vitro (a) sterility (b) antibacterial property, (c) cell viability, (d) hemolysis and (e) thrombogenicity of 5NC0A, 5NC0A/ICG0.1, 4.5NC0.5A, .5NC0.5A/ICG0.1 gels. Data was reported as mean ± standard error of the mean (s.e.m.) (n = 3 for (a, b, d), n = 6 for (c)). ns – not significant; ****p < 0.0001.

Antibacterial efficacy was tested for 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1 was evaluated. The results (Fig. 10(b)) showed a reduction in bacterial growth of 56 % and 44 %, respectively, suggesting their antibacterial properties.

3.5.2. Cell viability

The biocompatibility of 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1 was assessed using L-929 fibroblast cells following the ISO-10993-5 standard [49]. The results (Fig. 10(c)) showed consistent cell viability exceeding 70 % across sequential dilution of gel extractions. No statistical significance difference in cell viability was observed within each gel group (p ≥0.775). The analysis indicates excellent cell biocompatibility of 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1 [49].

3.5.3. Hemocompatibility

In this study, a hemolysis test was carried out following the standard of ISO-10993-4 standard [48]. As shown in Fig. 10(d), the hemolysis rates were analyzed to be 1.46 % ± 0.44 % for 5NC0A/ICG0.1 and 2.25 % ± 0.43 % for 4.5NC0.5A/ICG0.1. These values were below the permissible threshold of 5 % [48], indicating excellent hemocompatibility for both embolic agents. Furthermore, statistical analysis revealed no significant difference in hemolysis rates between the two embolic gels (p = 0.36).

3.5.4. Thrombogenicity

Thrombogenicity, the ability of a material to promote blood clotting [56], was further investigated. As shown in Fig. 10(e), all materials examined in the test showed a progressive formation of blood clots starting at 3 min of interaction. The clot formation exhibited and achieved complete coagulation in 7 min. The results indicate that the clotting capability of 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1 is comparable to that of clinically used coils.

4. Discussion

4.1. Rheology of NCA/ICG gels

In the study, 5NC0A and 4.5NC0.5A embolic gels were selected due to their optimized mechanical properties, minimized adverse reactions, and good physician experience from our previous study [13].

After introducing ICG, the shear-thinning behavior of NCA/ICG gels were confirmed through shear rate sweeps, LAOS, and frequency sweep tests. During injection, the gels gradually transited to a liquid-like state at higher strains (>10⁰) and fully became a liquid-like state at the sol-gel transition point (tan δ = 1), where the shear strain was approximately 10¹. The shear-thinning behavior ensured a smooth delivery through microcatheters.

Once the injection was completed and the force was removed, the gels reverted to a solid-like state. The gels were able to maintain solid-like behavior over a wide strain range from 10⁻² to around 10¹ (tan δ = 1), indicating their resistance to deformation. The yield stress of all gels were >13 times higher than the peak wall shear stress in human blood vessels (4.3 Pa) [57], demonstrating the ability of the gels withstand the dynamic mechanical forces under physiological conditions [58,59]. Additionally, the thixotropy tests confirmed the recoverability of the gels under repeated force application, ensuring consistent performance after multiple injections in clinical scenarios [9]. These rheological characteristics indicate the suitability of the gels for clinical embolization procedures.

Additionally, the ICG concentration did not significantly affect the storage modulus nor yield strain. While it did impact the yield stress, this value is much higher (13 folds) than the required force of 4.3 Pa [57]. The difference is negligible in the context of the overall yield stress. Thus, we can safely incorporate ICG without significantly compromising the gel’s mechanical properties and stability.

4.2. Microcatheter Injectability

Injectability is a practical factor that needs to be evaluated to ensure physicians’ comfort of transcatheter delivery. A microcatheter of 1.8 F was used due to its capacity to access fine vasculature, making it suitable for neurosurgical applications such as AVM. The average maximum finger force during injection is 79.8 N, where 95.4 N for males and 64.1 N for females. The injection force for both embolic agents is below 60 N, which is marked with dashed line in Fig. 3(b) [60–62], indicating good injectability.

4.3. Fluorescence

Current embolic agents used in clinical practice (Table S1), including beads and liquid embolic agents, do not carry inherent fluorescent imaging capability [63–77]. The clinical use of ICG is often restricted by its short signal duration. There have been limited studies on ICG-loaded embolic formulations; for instance, ICG has been mixed with Lipiodol for tumor localization and fluorescence-guided surgery [31]. This approach primarily involves liquid embolic agents designed to enhance surgical guidance during fluorescence-guided tumor resection.

In contrast, the NCA/ICG embolic agent overcomes these challenges by delivering strong and prolonged fluorescence signals, thereby enhancing real-time imaging without the need of pre-mixing. Furthermore, it features shear-thinning properties for versatile embolization, along with optimized rheology for controlled delivery and effective embolic performance, making it a promising option for embolization procedures [13].

4.3.1. Fluorescent signal enhancement

ICG is a fluorescent dye that is widely used for real-time anatomical visualization in clinical settings [24,78]. Despite its wide applicability, ICG is known for its low stability in aqueous solution [79–81]. Dynamic quenching, the decrease in fluorescence intensity due to the interaction between fluorescent molecular and a quencher, has been identified as the major contributor to the loss of ICG signals [35,36]. To preserve ICG signal, it is essential to minimize the self-quenching effect by ensuring a consistent and uniform dispersion of ICG, as well as preventing quenching from extensive blood-ICG interaction [24,33]. Additionally, its clearance due to blood flow poses a challenge for long-term imaging and signal stability [24,33,82].

In this study, ICG fluorophores were integrated into NCA embolic agents to prevent aggregation, a major concern acknowledged to fluorophore self-quenching and signal reduction [34,35]. Unlike free ICG fluorophores, which are randomly dispersed in aqueous solutions, NCA forms a stable “house of card” network [18], where electrostatic interactions [34] facilitated the adsorption of negatively charged ICG to the positively charged edges of NC particles [83].

Laponite has a negatively charged surface and positively charged edges [21,84]. Sodium alginate is an anionic polysaccharide, containing carboxylate (−COO⁻) groups that can interact with cationic molecules [20,22]. ICG contains sulfonate (−SO₃⁻) groups, which are negatively charged in aqueous solutions [35]. Electrostatic interactions are the dominant initial mechanism in NCA gel formation. The positively charged Laponite edges facilitate the adsorption of negatively charged ICG, while sodium alginate primarily interacts with cationic species, further reinforcing the gel network. Additional molecular interactions may further contribute to the stability of the system.

Absorption spectra confirmed that ICG remains well-dispersed in the NCA gels, primarily in the monomeric state. This state allows efficient π-electron transitions that lead to strong fluorescence emissions [34]. Besides, minimal intermolecular interactions in the NCA gel reduce self-quenching effects, preserving the optical properties of ICG (Fig. 11(c)).

Fig. 11.

Schematic representation of ICG fluorescence reduction and enhancement mechanisms. (a) ICG in water undergoes H-aggregation and self-quenching, leading to reduced fluorescence. (b) In the bloodstream, free ICG is rapidly cleared and experiences fluorescence quenching due to interactions with blood components. (c) ICG in NCA gel remains in a well-dispersed monomeric state, preventing H-aggregation and reducing self-quenching. (d) NCA gel retains ICG, preventing rapid clearance and minimizing fluorescence quenching by reducing blood interaction.

In contrast, H-aggregation was observed in the ICG aqueous solution, face-to-face π-π stacking interactions between ICG molecules led to exciton splitting and fluorescence quenching. This occurs because exciton coupling in H-aggregates results in an energy band structure where the lowest energy electronic transition is optically forbidden, preventing efficient radiative decay and leading to non-emissive energy loss. As a result, ICG fluorescence is significantly reduced in aqueous conditions due to strong intermolecular interactions (Fig. 11(a)).

No J-aggregation was detected at 890 nm in either the aqueous solution or the NCA/ICG gel [78]. J-aggregates, which typically form end-to-end (head-to-tail) molecular stacking, allow π-electron delocalization across multiple chromophores, facilitating a red-shifted absorption peak and enhanced fluorescence emission [34]. Since neither condition exhibited J-aggregation, the fluorescence enhancement in the NCA/ICG gel is primarily attributed to ICG remaining in its monomeric state rather than excitation coupling effects [34].

Additionally, the solid-like nature of the NCA gel limited the interaction of ICG with quenchers in the blood [82] (Fig. 11(d)), in contrast to the freely circulating ICG in aqueous solutions (Fig. 11(b)). This reduced mobility helps maintain the fluorescent signal over a longer period. The enhanced signal intensity and longevity in the NCA/ICG gels, compared to the clinically used ICG aqueous solution over the 28-day monitoring period, were confirmed by results from the microplate reader and IVIS.

Although NC particles dispersed in water are known to reduce light transmission primarily through absorption [85,86], the comparison of the absorbance intensity between NCA and water revealed only a slight enhancement in light absorption. This observation suggests that the enhanced fluorescent signal of the NCA/ICG gels is not attribute to the changes in the matrix components. Instead, by the effective dispersion of ICG is the primarily driven rather than by such as water or the NCA gel itself.

Overall, the NCA/ICG embolic agent consistently exhibits persistent and strong signals across varies ICG concentration, in contrast to the current practice of dissolving ICG in sterile water. These strong signals can improve real-time imaging and post- operative monitoring even after days of injection. The design ensured a strong and consistent fluorescent signal without introducing harmful solvents such as ethanol or DMSO [35,37].

4.3.2. Potential applications

4.3.2.1. AVM nidus models.

The results from IVIS tests highlighted the effective penetration capabilities of the NCA/ICG embolic agents in AVM models with tumors and vasculature, which is crucial for clinical applications. The ability of the NCA/ICG gels to fill confined spaces indicates their potential to navigate tortuous and intricate blood vessels effectively. The observed fluorescent signals confirm that these embolic agents reached the target areas with good visibility for real-time monitoring.

It is important to note that the model has limitations. It serves as an initial validation step and does not fully replicate the dynamic and complicated physiological conditions of the human AVM. To further evaluate the clinical applicability of the NCA/ICG system, future studies are clearly needed to use physiologically relevant in vitro models, such as microfluidic vascular networks, and ultimately, in vivo studies, such as porcine rete mirabile model [87], which closely resembles human cerebral vasculature.

4.3.2.2. Photothermal therapy.

ICG has been widely utilized in photothermal therapy for treating vascular anomalies [44,53], tumors [45], and diseases such as oral leukoplakia [54]. The significant temperature increase observed in ICG solutions underscores its capability to absorb laser energy effectively. However, this also raises concerns regarding the uniformity of heat distribution, as uneven heat can potentially cause damage to surrounding tissues.

By integrating ICG into NCA gels, NCA/ICG provides better thermal management compared to conventional ICG solutions. The minimal temperature rise in the gels without ICG reflects their lack of photothermal activity, highlighting the importance of ICG for effective heat generation. The low variations in heat distribution observed in the NCA/ICG gels enhance safety and targeting capabilities, making them a promising option for PTT.

Besides, the observed increase in maximum temperature at 5 W/cm² compared to 2 W/cm² highlights the power-dependent photothermal response of the NCA/ICG gels. While both 5NC0A/ICG0.1 and 4.5NC0.5A/ICG0.1 exhibited higher temperatures at 5 W/cm², the heating profiles remained more uniform compared to free ICG solutions. This indicates that the gel matrix effectively modulates heat diffusion, reducing localized overheating that is often observed with the currently clinical used ICG aqueous solution.

To better understand the thermal effects of laser energy on biological tissue, a 1 wt% agar tissue layer was prepared to mimic the hypodermis layer of blood vessels [88,89]. The results show that the presence of a tissue model layer affects the temperature distribution during laser heating of NCA/ICG gels. The lower temperatures on the tissue side suggest that the tissue model restricts thermal transmission, likely due to its distinct thermal properties. Scattering and absorption effects further reduce heat accumulation in the gel, leading to lower maximum temperatures compared to gels heated without a tissue layer. Additionally, the gradual decrease in temperature with the tissue model highlights the need to optimize laser parameters to ensure adequate heating in clinical applications.

Additionally, the calculation of photothermal conversion efficiency is provided in the Supplementary information. Notably, although the calculated photothermal conversion efficiency of the ICG aqueous solution is higher, this does not necessarily translate into greater therapeutic efficacy in vivo. In a biological environment, free ICG is rapidly cleared by the bloodstream [24,33,82], limiting its effective photothermal performance despite the high conversion efficiency measured in vitro. Moreover, in our tests, temperature measurements were taken linearly along the heating path, rather than throughout the entire sample volume. The higher photothermal conversion efficiency calculated for the ICG aqueous solution primarily results from fluid circulation, which facilitates uniform heating and interaction of a larger amount ICG throughout the entire solution. In contrast, the NCA/ICG gel, which behaves as a solid-like material upon deployment, confines ICG to localized regions, thereby achieving targeted, localized heating.

Overall, the NCA/ICG formulation enables localized heating along specific paths due to its solid-like properties once deployed, which prevents rapid clearance by the bloodstream. This targeted thermal management offers improved control over the photothermal effect compared to free ICG in aqueous solution.

4.4. Shelf-life evaluation of the gels

The shelf-life studies of 5NC0A, 5NC0A/ICG0.1, 4.5NC0.5A, and 4.5NC0.5A/ICG0.1 gels showed an increase in yield stress and storage module, indicating progressive strengthening over time, which is consistently higher than the peak wall shear stress in human blood vessels (4.3 Pa) [57]. The injection force also increased over time, with a maximum recorded force of 47.40 N on day 15 for 4.5NC0.5A/ICG0.1. However, this value remains well below the 60 N threshold for comfortable injection, ensuring continued feasibility for clinical use. Overall, the NCA/ICG gel maintained its structural integrity and injectability after 15 days of storage, demonstrating stability.

Long-term stability and controlled degradation are essential for embolic agents. Degradation studies demonstrated that NCA/ICG gels maintain over 90 % residual mass after 14 days, indicating gradual and predictable degradation, which is ideal for vessel occlusion. Further animal studies are needed to investigate the in vivo degradation profiles of the gels. Furthermore, storage stability is a critical factor for clinical translation, and the NCA/ICG gels maintained structural integrity over prolonged periods, supporting their practicality for embolization.

4.5. In-vitro characterization of NCA/ICG gels

The in-vitro characterization of NCA/ICG gels emphasizes their biocompatibility and safety, which are crucial for successful clinical applications. The antibacterial properties of these gels confirmed the lower infection risks during medical procedures. The observed antibacterial activity can be primarily attributed to the intrinsic antimicrobial properties of NC. Studies have shown that NC exhibits antibacterial effects due to its high surface charge density and positively charged edges, which can interact with bacterial cell membranes, leading to membrane destabilization and bacterial death [90]. Moreover, the cell viability test revealed that the NCA/ICG gels surpassed the acceptable threshold for biocompatibility [49], further supporting their suitability for clinical use.

Additionally, the results indicate that NCA/ICG gels can safely interact with blood, minimizing the risk of hemolytic reactions. Hemocompatibility tests revealed low hemolysis rates for NCA/ICG gels, which helps prevent complications and adverse effects such as red blood cell disorders and thrombosis [91–93]. Effective clot formation is crucial, as it can strengthen embolic agent/blood complex and reduce risk of distinct migration. The NCA/ICG gels exhibited clotting capabilities comparable to clinically used coils, ensuring effective occlusion while reducing the risk of recanalization and stroke [94].

Overall, the in-vitro characterization of the NCA/ICG gels presents a promising approach for improving the safety and efficacy of embolic agents, before moving forward with in-vivo studies.

4.6. Limitations

While our study provides comprehensive in vitro characterizations, we acknowledge that in vivo validation is essential to fully assess the physiological performance of the NCA/ICG embolic agent [95–97]. However, in vivo the evaluation represents a significant developmental step that extends beyond the scope of the current work, which lays the foundational groundwork for future investigations. Future research will build on these findings to further validate the in vivo performance and safety of the NCA/ICG embolic agents. Ultimately, this progress will bring the NCA/ICG embolic agent closer to clinical application.

5. Conclusion

We have successfully developed a transcatheter injectable NCA/ICG gel embolic agent, which demonstrated excellent rheological properties, injectability, and stability, ensuring precise and accurate embolization during procedures. This formulation effectively stabilizes ICG by physically absorbing it onto the edges of disc-shaped NC particles, significantly enhancing both the fluorescent signal intensity for real-time imaging and the longevity of the signal for post-operative monitoring over 28 days. Furthermore, the NCA/ICG0.1 exhibited a more stable heating profile compared to clinical used ICG aqueous solution, suggesting potential for photothermal therapy. These embolic agents showed biocompatibility, confirming their safety. Additionally, NCA/ICG embolic agents possessed significant antibacterial properties, reducing bacterial growth by 44 % for 4.5NC0.5A/ICG0.1 and 56 % for 5NC0A/ICG0.1, broadening their potential application for use in patients with bacterial infections or sepsis. Overall, the NCA/ICG embolic agent is a promising material for a wide range of embolization procedures, offering additional benefits of imaging, photothermal therapy, and antibacterial effects.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.bioadv.2025.214298.

Data availability

Data will be made available on request.