The Dual-Targeted Peptide Conjugated Probe for Depicting Residual Nasopharyngeal Carcinoma and Guiding Surgery

Detecting residual nasopharyngeal carcinoma (rNPC) can be difficult because of the coexistence of occult tumours and post-chemoradiation changes, which poses a challenge for both radiologists and surgeons using current imaging methods. Currently, molecular imaging that precisely targets and visualises particular biomarkers in tumours may exceed the specificity and sensitivity of traditional imaging techniques, providing the potential to distinguish tumours from non-neoplastic lesions. Here, we synthesised a HER2/SR-BI-targeted tracer to efficiently position NPC and guide surgery in living mice. This bispecific tracer contained the following two parts: IRDye 800 CW, as an imaging reagent for both optical and optoacoustic imaging, and a fusion peptide (FY-35), as the targeting reagent. Both in vitro and in vivo tests demonstrated that the tracer had higher accumulation and longer retention (up to 48 h) in tumours than a single-targeted probe, and realised sensitive detection of tumours with a minimum size of 3.9 mm. By visualising the vascular network via a customised handheld optoacoustic scan, our intraoperative fluorescence molecular imaging system provides accurate guidance for intraoperative tumour resection. Integrating the advantages of both optical and optoacoustic scanning in an intraoperative image-guided system, this method holds promise for depicting rNPC and guiding salvage surgery.


Introduction
Nasopharyngeal carcinoma (NPC) is an epithelial malignancy characterised by a distinct geographical distribution [1]. Up to 95% of patients diagnosed with NPCs are treated with radiotherapy [2]. Patients with early-stage NPC have encouraging survival rates, with the 5 year local control rates reaching 90%, while approximately 10% of patients with residual NPC (rNPC) demonstrate poor survival, and their 5 year local control rates are only 41% [3]. Recently, salvage surgery has been reported to provide effective tumour control, which is beneficial for achieving long-term survival in patients with local persistent NPC or rNPC [4]. However, intraoperative findings under endoscopic sinus surgery are inconclusive for delineating residual tumour margins, due to the presence of inflammatory tissue, fibrous hyperplasia, and oedema after radiotherapy [5]. Moreover, rNPC is not completely accessible preoperatively using conventional imaging techniques, with insufficient specificity [6]. Therefore, the accurate identification of rNPC has been a great challenge for both early detection and salvage surgery.

Synthesis and Characterisation of Optical Tracers
The dual-targeted peptide FY-35 (FAEKFKEAVKDYFAKFWD, termed F; LTVSPW YLTVSPWY, termed Y) and a single-targeted peptide FL-35 (FAEKFKEAVKDYFAKFWD termed F; SYTYSTPVWVPLWL, termed L) were obtained by solid-phase peptide synthe sis in Apeptide Co., Ltd. (Shanghai, China). The synthesis cycle consisted of the following (I) Fmoc cleavage: 20% piperidine in N, N-dimethylformamide (DMF), (II) DMF washings (III) coupling, and (IV) DMF washings. The resin was treated with 2 mL dichloromethane (DCM) for 5 min three times and 2 mL DMF for 5 min three times before it was used. The resin was washed thoroughly before it was incubated with HBTU-preactivated amino acid, or commercially-available fluorophores (FITC/IRDye800 CW) activated ester. Cleav age from the solid support was performed with trifluoroacetic acid (TFA)/water/triiso propylsilane (95/2.5/2.5, vol/vol/vol) over 3 h at room temperature. All products were pu rified by high-performance liquid chromatography (HPLC) using a Gemini-NX 10μ C18 4.6 × 250 mm column (Phenomenex, Torrance, CA, USA) on a Shimadzu LC-2010A HPLC System (Kyoto, Japan). Mass spectroscopy was used to characterise the probes. The FITC labelled probes were used for in vitro cell-binding assays.

Cell Culture
Human NPC cells 5-8F were transfected with firefly luciferase for in vivo imaging Transfected cells (5-8F-fLuc) and non-transfected cells (5-8F) were cultured in high-glu cose Dulbecco's modified Eagle's medium (DMEM), containing 10% foetal bovine serum 1% penicillin, and streptomycin in a humidified atmosphere of 5% CO2 at 37 °C. The cells were harvested at 80% confluence and maintained in the exponential growth phase. Cel density was calculated using a haemocytometer and performed before any experiments.

Laser Confocal Microscopy
Confocal imaging was performed to explore the targeting efficiencies of FY-35 and FL-35. The 5-8F cells were cultured on confocal plates (500 μL of DMEM containing 5 × 10 2 cells) for 24 h at 37 °C and then co-incubated with FITC-FY-35 or FITC-FL-35 (1.5 μM for 1 h. The cells were then washed three times with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde (PFA) for 8 min. Then, 4, 6-diamidino-2-phenylindole and rhodamine B were used to stain the cellular nuclei and cytoskeleton, respectively Immunofluorescence staining was performed using SR-BI and HER2 antibodies. The cells were fixed with 4% PFA for 8 min, permeabilised with 0.1% Triton™ X-100 for 15 min and blocked with 2% bovine serum albumin (BSA) for 1 h at room temperature. The cells were then co-incubated with the SR-BI antibody (ab217318; 1:250 dilution in 0.1% BSA) or HER2 antibody (ab134182; 1:500 dilution in 0.1% BSA) overnight at 4 °C. After the stand ard washing process, the cells were co-incubated with rabbit anti-goat IgG H&L

Synthesis and Characterisation of Optical Tracers
The dual-targeted peptide FY-35 (FAEKFKEAVKDYFAKFWD, termed F; LTVSP-WYLTVSPWY, termed Y) and a single-targeted peptide FL-35 (FAEKFKEAVKDYFAK-FWD, termed F; SYTYSTPVWVPLWL, termed L) were obtained by solid-phase peptide synthesis in Apeptide Co., Ltd. (Shanghai, China). The synthesis cycle consisted of the following: (I) Fmoc cleavage: 20% piperidine in N, N-dimethylformamide (DMF), (II) DMF washings, (III) coupling, and (IV) DMF washings. The resin was treated with 2 mL dichloromethane (DCM) for 5 min three times and 2 mL DMF for 5 min three times before it was used. The resin was washed thoroughly before it was incubated with HBTUpreactivated amino acid, or commercially-available fluorophores (FITC/IRDye800 CW) activated ester. Cleavage from the solid support was performed with trifluoroacetic acid (TFA)/water/triisopropylsilane (95/2.5/2.5, vol/vol/vol) over 3 h at room temperature. All products were purified by high-performance liquid chromatography (HPLC) using a Gemini-NX 10µ C18 4.6 × 250 mm column (Phenomenex, Torrance, CA, USA) on a Shimadzu LC-2010A HPLC System (Kyoto, Japan). Mass spectroscopy was used to characterise the probes. The FITC-labelled probes were used for in vitro cell-binding assays.

Cell Culture
Human NPC cells 5-8F were transfected with firefly luciferase for in vivo imaging. Transfected cells (5-8F-fLuc) and non-transfected cells (5-8F) were cultured in high-glucose Dulbecco's modified Eagle's medium (DMEM), containing 10% foetal bovine serum, 1% penicillin, and streptomycin in a humidified atmosphere of 5% CO 2 at 37 • C. The cells were harvested at 80% confluence and maintained in the exponential growth phase. Cell density was calculated using a haemocytometer and performed before any experiments.

Laser Confocal Microscopy
Confocal imaging was performed to explore the targeting efficiencies of FY-35 and FL-35. The 5-8F cells were cultured on confocal plates (500 µL of DMEM containing 5 × 10 2 cells) for 24 h at 37 • C and then co-incubated with FITC-FY-35 or FITC-FL-35 (1.5 µM) for 1 h. The cells were then washed three times with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde (PFA) for 8 min. Then, 4,6-diamidino-2-phenylindole and rhodamine B were used to stain the cellular nuclei and cytoskeleton, respectively. Immunofluorescence staining was performed using SR-BI and HER2 antibodies. The cells were fixed with 4% PFA for 8 min, permeabilised with 0.1% Triton™ X-100 for 15 min, and blocked with 2% bovine serum albumin (BSA) for 1 h at room temperature. The cells were then co-incubated with the SR-BI antibody (ab217318; 1:250 dilution in 0.1% BSA) or HER2 antibody (ab134182; 1:500 dilution in 0.1% BSA) overnight at 4 • C. After the standard washing process, the cells were co-incubated with rabbit anti-goat IgG H&L (ab150141) at a dilution of 1:1000 for 45 min at room temperature. All the above-mentioned staining processes were conducted away from light. Confocal images were acquired using a confocal laser scanning microscope (LSM-710, Carl Zeiss, Oberkochen, Germany). Imaging was performed using ZEN 2.3 lite (Zeiss, Oberkochen, Germany). Quantification of mean fluorescence intensity (MFI) was conducted using Image J 2.X (LOCI, University of Wisconsin, Madison, WI, USA).

Flow Cytometry Analysis
The 5-8F cells were cultured in 6-well plates (1.2 × 10 3 cells/mL) for 24 h at 37 • C. When the cells reached the desired density, cells in the six plates were co-incubated with FITC labelled FY-35 or FL-35 at six concentrations (0, 1.5, 3, 6, 12, and 24 µM) for 1 h. Before flow cytometry analysis, the cells were washed three times with PBS, dissociated from the wells with trypsin-ethylenediaminetetraacetic acid (EDTA), and then resuspended in PBS. Cell counting was conducted using a flow cytometer (BD Accuri™ C6, BD Biosciences, Franklin Lakes, NJ, USA) with an excitation wavelength of 488 nm for FITC (525/30 emission filter). The data were analysed using FlowJo software (FlowJo 10.8, Ashland, OR, USA).

Animal Models
Animal experiments were performed in accordance with the approval of Animal Welfare and Ethics Committee of Jinan University (approval number: IACUC-20211210-03). Four-week-old BALB/c nude mice were purchased from the Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). Subcutaneous tumour models received a subcutaneous injection of 5-8F (5.0 × 10 6 in 200 µL PBS) in the lower right flank. Experiments were started 10-14 days after tumour implantation, when the tumour diameter was 6-8 mm. Tumour volumes were measured using a Vernier calliper and calculated using the formula for a hemiellipsoid (volume = 0.5236 × length × width 2 ), as this form best approximated the tumour shape [24]. For orthotopic tumour models, mice were anaesthetised using isoflurane with a precision vaporiser at 1% and injected with 25 µL of 1.1 × 10 5 5-8F-fLuc cells into the nasopharynx using a 28-gauge needle, as previously reported [25]. Mice were fed a specific pathogen-free unit, and animal surgery was performed in a sterile hood. Orthotopic tumours were visualised and tracked by bioluminescence imaging (BLI) using an IVIS ® Spectrum In Vivo Imaging System (PerkinElmer, Waltham, MA, USA). Specifically, mice received an intraperitoneal injection of 3 mg/mL luciferin bioluminescent substrate (Perkin Elmer, Waltham, MA, USA) 4 min before imaging.

Biodistribution and Tumour Selectivity of IR800-FY-35
Mice with subcutaneous 5-8F tumours were randomly divided into three groups (n = 3 per group). The three groups received intravenous (i.v.) injection with 12 µM FY-35 (dual-target dye), 12 µM FL-35 (single-target dye), and equimolar quantities of free dye, respectively. Fluorescence images were acquired pre-, 2 h, 6 h, 10 h, and 24 h post-injection. The mice in each group were euthanised. The heart, lung, liver, pancreas, spleen, kidney, intestine, and tumour tissues were collected for ex vivo imaging. The MFI was analysed to determine organ biodistribution and tumour selectivity in the three groups. The TBR was calculated by drawing regions of interest within the tumour region and ipsilateral auricle of each mouse.

Detecting Residual Tumours Using IR800-FY-35-Based Fluorescence Imaging
To identify the minimal size of detectable tumours, subcutaneous tumours of different sizes (1-7 mm; n = 8) were injected with 12 µM FY-35, and imaged at 24 h post-injection to measure MFI in the tumour region. A linear regression model was used to demonstrate a linear relationship between tumour volume and MFI. To confirm the feasibility of IR800-FY-35 for detecting residual tumours, mice with orthotopic tumours (n = 5) were intravenously injected with 12 µM IR800-FY-35. Fluorescent images at various time points (2,4,6,8,12,24, and 48 h) were acquired with a 745-nm excitation wavelength and 800-nm filter using the IVIS ® Spectrum In Vivo Imaging System (PerkinElmer, Waltham, MA, USA). After 48 h post-injection, the mice were sacrificed. The orthotopic tumours and main organs were imaged ex vivo. To simulate salvage surgery in patients with NPC, we performed tumour resection on the sacrificed mice. Repeated imaging was performed to identify residual signals. The resected tumours were embedded in optimal cutting temperature (OCT) and stored at −80 • C. Serial sections were cut at a thickness of 10 µm for fluorescence imaging using an inverted microscope (Leica, Wetzlar, Germany), and 4 µm for haematoxylin-eosin (H&E) staining.

Imaging-Guided Surgery
Mice bearing subcutaneous NPC xenografts were IV injected with 12 µM IR800-FY-35, and surgery was performed at 24 h post-injection. The mice were anaesthetised, and the tumours were first subjected to optoacoustic imaging using a customised hand-held optoacoustic scanner to visualise tumour vessels. The tumours were then visualised and excised using the FMI intraoperative guiding system (DPM-III-01, Zhuhai Dipu Medical Technology Co., Ltd., Guangdong, China) with the advantage of convenient operation. The FMI intraoperative system is a dual-mode imager that combines the following two systems: NIR fluorescent imaging to detect targeted cancer cells, and visible-light reflectance imaging to observe the contours of the tissue itself. This system improves the image quality via feature point algorithms and allows rapid and precise imaging fusion for two cameras [26]. Surgically excised tissues were embedded in OCT, cut into serial sections, and subjected to H&E staining and confocal imaging.

Statistical Analysis
Statistical analysis was conducted using GraphPad Prism (9.0.0, GraphPad Software, San Diego, CA, USA). Data are reported as mean ± standard deviation. Comparative analysis was performed using an independent two-sample t-test or a chi-square test. Simple linear regression was used to estimate the relationship between two quantitative variables. Here, p < 0.05 was considered significant.

Synthesis of Optical Tracers and Identification of Their Targeting Efficiency Ex Vivo
Here, FY-35 was a fusion peptide targeting both HER2 and SR-BI for our imaging studies, while FL-35 was designed as a single-target control, only binding with SR-BI. The chemical structure of IRDye 800 CW-labelled FY-35 and FL-35 were shown in Figure 2. The FITC-labelled FY-35 and FL-35 were synthesised for ex vivo tests ( Figure S1, Supplementary Materials). The mass spectra were presented in Figures S2 and S3, and the purity of these products was about 98% (Figures S4 and S5). The targeting efficiency was first identified by confocal imaging and flow cytometry analyses in vitro. Based on effective receptor-mediated endocytosis, probe uptake within the two groups was qualitatively and quantitatively analysed (Figure 3a identified by confocal imaging and flow cytometry analyses in vitro. Based on effective receptor-mediated endocytosis, probe uptake within the two groups was qualitatively and quantitatively analysed (Figure 3a,b). As indicated in Figure 3a, bright green fluorescence was observed in the cytomembrane and cytoplasm of each group, indicating that the two tracers were easily internalised by cells. The cellular uptake of FITC-FY-35 by 5-8F cells was significantly higher than that of FITC-FL-35 (p = 0.0001). Immunofluorescence staining demonstrated stable co-expression of SR-BI and HER2 in 5-8F cells, providing the basis for effective targeting of the two tracers. The flow cytometry results further validated the findings of confocal imaging (Figure 3c). With increasing co-incubated concentrations of FITC-FY-35 and FITC-FL-35, the fluorescent peaks shifted to the right, indicating that the number of probe-binding cells was positively correlated with the concentration. Although the two tracers demonstrated selective binding with 5-8F cells, the ratio of marked cells in the FITC-FY-35 group (100%) was higher than that in the FITC-FL-35 group (84.7%) at the first concentration (1.5 μM), demonstrating the additional targeting capacity of FY-35.

Biodistribution and Tumour Selectivity of IR800-FY-35
To further explore in vivo targeting efficiency, a biodistribution test was performed using a subcutaneous NPC mouse model. At 2 weeks following tumour inoculation, the tumour-bearing mice were randomised into three groups (n = 5 per group) with equal average tumour burden (Figure 4a-c). The mice were IV administered IR800-FY-35, IR800-FL-35, or free dye. Similarly, both targeted tracers revealed apparent tumour selectivity at 6 h and 10 h post injection, indicating that SR-BI and HER2 might be suitable imaging markers for NPC. Significantly, the higher signal intensity within the tumour region (3.575 ± 2.953) in the IR800-FY-35 group was demonstrated in comparison with the IR800-FL-35 group (1.835 ± 2.178) or free dye group (1.65 ± 1.923) (Figure 4d,e). The maximum TBR was 13.83, which was 2.3-fold higher than that of IR800-FL-35 (maximum TBR = 5.99). Except for the superior optical contrast, the longer retention of IR800-FY-35 in 5-8F tumours was observed at 24 h after injection, demonstrating the potential for applying this tracer in imaging-guided surgery, since preoperative preparation in clinical practice generally requires more time. Ex vivo fluorescence imaging of the organs and tumours resected from the two groups of mice demonstrated that both tracers mainly accumulated in the kidney, liver, and tumours ( Figure S6). Quantitative analysis of MFI revealed a significant difference in tumour uptake between the two groups (p = 0.0079; Figure 3f). Altogether, IR800-FY-35 demonstrated a higher binding affinity and longer retention in NPC than IR800-FL-35.

Detecting Minimal Residual NPC Using IR800-FY-35
To confirm the feasibility of IR800-FY-35 in identifying residual tumours, we first tested the minimal size of tumours that could be detected in subcutaneous tumours. Mice bearing 5-8F tumours (n = 8) of different sizes (1-7 mm) were injected with equivalent IR800-FY-35 and imaged at several time points. The MFI in the tumour region at the 24 h time point was recorded in each mouse, and then a linear regression model was performed to exhibit a linear relationship between the tumour volumes and MFI. We observed a linear relationship between these two variables (Y = 7.178 × X-41.31) ( Figure 5). The MFI increased remarkably when the tumour volume reached 32 mm 3 , indicating that the minimum size of detectable tumours was 3.9 mm (Figure 6a).
To further evaluate the feasibility of IR800-FY-35 in detecting residual tumours, an orthotopic mouse model (n = 5) was constructed by injecting 5-8F-fLuc cells into the nasopharynx and experimented when the bioluminescent signal reached 10 8 . Considering that the parotid gland presented apparent probe uptake (Figure 4a-c), the delayed imaging time points might have avoided troublesome signal interference. Thus, mice with orthotopic tumours were imaged at 2, 5, 8, 12, 24, and 48 h post-injection. Similarly, superior optical contrast was observed in the tumour region, which was consistent with that in the subcutaneous tumour, but surprisingly persistent until 48 h post-injection (Figure 6b). The mice were then euthanised and exposed to the tumours and the main organs (Figure 6c). To simulate salvage surgery, we performed tumour resection ex vivo. Following the first white-light resection and removal of the macroscopic tumour, the residual signal was adjacent to the tumour bed, as shown in Figure 6d, indicating a suspicious tumour that was not easily detectable under white light. The presence of cancer cells was confirmed by H&E staining. Cancer cells exhibited strong NIR signals in fluorescence microscopy imaging, whereas the muscle had weak signals on the same slice (Figure 6e). These results verified the feasibility of IR800-FY-35 in detecting minimal residual NPC.

Detecting Minimal Residual NPC Using IR800-FY-35
To confirm the feasibility of IR800-FY-35 in identifying residual tumours, we first tested the minimal size of tumours that could be detected in subcutaneous tumours. Mice bearing 5-8F tumours (n = 8) of different sizes (1-7 mm) were injected with equivalent IR800-FY-35 and imaged at several time points. The MFI in the tumour region at the 24 h time point was recorded in each mouse, and then a linear regression model was performed to exhibit a linear relationship between the tumour volumes and MFI. We observed a linear relationship between these two variables (Y = 7.178 × X-41.31) ( Figure 5) The MFI increased remarkably when the tumour volume reached 32 mm 3 , indicating that the minimum size of detectable tumours was 3.9 mm (Figure 6a).    To further evaluate the feasibility of IR800-FY-35 in detecting residual tumours, an orthotopic mouse model (n = 5) was constructed by injecting 5-8F-fLuc cells into the nasopharynx and experimented when the bioluminescent signal reached 10 8 . Considering that the parotid gland presented apparent probe uptake (Figure 4a-c), the delayed imaging time points might have avoided troublesome signal interference. Thus, mice with orthotopic tumours were imaged at 2, 5, 8, 12, 24, and 48 h post-injection. Similarly, superior optical contrast was observed in the tumour region, which was consistent with that in the subcutaneous tumour, but surprisingly persistent until 48 h post-injection (Figure 6b). The mice were then euthanised and exposed to the tumours and the main organs (Figure 6c). To simulate salvage surgery, we performed tumour resection ex vivo. Following the first white-light resection and removal of the macroscopic tumour, the residual signal was adjacent to the tumour bed, as shown in Figure 6d, indicating a suspicious tumour that was not easily detectable under white light. The presence of cancer cells was confirmed by

Photoacoustic and Fluorescence Imaging-Guided Resection
The common complications of salvage surgery are intraoperative haemorrhage and inadvertent damage to the carotid vessel; intraoperative tumour vascular and adjacent artery imaging could effectively reduce surgical complications. To avoid severe bleeding in orthotopic mouse models due to tumour infiltration into adjacent muscle tissues, the subcutaneous tumour model (n = 5) was IV injected with IR800-FY-35, and image-guided resection was conducted using both photoacoustic and fluorescence imaging. The 3D volumetric optoacoustic imaging was performed to clearly visualise the tumour vessels at 24 h after administration (Figure 7a). Subsequently, a surgeon attempted to resect all tumours completely under white light, while carefully ligating the tumour vessels to reduce bleeding during the procedure. The resected tumour and postoperative tumour bed are presented in Figure 7b. Confocal imaging of tumour slices revealed accurate microscopic co-localisation of tumour cells and the IRDy800 signal (Figure 7c,d). Based on these imaging data, the dual-targeted tracer could enhance the ability of surgeons to accurately conduct tumour resection by integrating fluorescence with optoacoustic imaging to guide surgery. presented in Figure 7b. Confocal imaging of tumour slices revealed accurate microscopic co-localisation of tumour cells and the IRDy800 signal (Figure 7c,d). Based on these imaging data, the dual-targeted tracer could enhance the ability of surgeons to accurately conduct tumour resection by integrating fluorescence with optoacoustic imaging to guide surgery.

Discussion
Here, we conjugated IRDy800CW to produce a fusion peptide (FY-35) that targeted HER2/SR-BI, and confirmed its superior targeting efficiency and longer retention in NPCs, with the contrast being improved 2.3-fold at 24 h post-injection and the retention time extended up to 48 h, when compared to a single-targeted tracer (IR800-FL-35). These results reveal the potential for IR800-FY-35 to be utilised for preoperative detection and to guide intraoperative resection of NPCs. Although it is not technically feasible to perform salvage surgery in live mice with orthotopic NPC, resection surgery on the executed mice with orthotopic tumours simulated the difficulty of residual tumour visualisation under white light and identified the feasibility of IR800-FY-35 in detecting minimal residual tu-

Discussion
Here, we conjugated IRDy800CW to produce a fusion peptide (FY-35) that targeted HER2/SR-BI, and confirmed its superior targeting efficiency and longer retention in NPCs, with the contrast being improved 2.3-fold at 24 h post-injection and the retention time extended up to 48 h, when compared to a single-targeted tracer (IR800-FL-35). These results reveal the potential for IR800-FY-35 to be utilised for preoperative detection and to guide intraoperative resection of NPCs. Although it is not technically feasible to perform salvage surgery in live mice with orthotopic NPC, resection surgery on the executed mice with orthotopic tumours simulated the difficulty of residual tumour visualisation under white light and identified the feasibility of IR800-FY-35 in detecting minimal residual tumours. Moreover, with the good performance of vascular network mapping by 3D volumetric optoacoustic imaging, surgical excision of subcutaneous NPC tumours was successfully achieved with the guidance of our intraoperative FMI system.
Despite advancements in NPC therapy, approximately 10% of patients develop residual or recurrent disease; the genomic landscape of rNPC is complex, with enriched mutations leading to enhanced tumour heterogeneity [27]. The dual-targeting strategy with two binding receptors copes with heterogeneity and allows for improved uptake in the region of interest for specific imaging or targeted therapy [9]. Here, the HER2/SR-BI targeting tracer in our study indicated that the addition of another binding monomer had advantageous effects on cellular internalisation and tumour uptake (Figures 3 and 4). Both in vitro and in vivo tests of FY-35 revealed significantly improved binding affinity with cancer cells, which was consistent with previous results [15].
Except for superior targeting efficiency, longer retention in the tumour region can provide a one to two day window that offers flexibility in scheduling surgical procedures. There are many platforms for tumour-targeted fluorescent probes [28]. Besides a peptide probe in our study, small molecules and antibodies are available in tumour-targeting platforms for NPC. Antibodies, the most widely used platform, have effective accumulation and longer retention in tumours, but suffer from pharmacokinetic limitations, including slow tissue clearance (normally 2-5 days or longer) and non-specific organ uptake [7]. Small molecules or peptide probes show rapid diffusion into the circulation blood stream, with the advantage of low cost and minor side effects and, thus, have promise in clinical settings [29]. However, the short retention time in tumours is a disadvantage of peptide probes when utilised in fluorescenceguided surgery [30]. Here, the dual-targeted probe in our study surprisingly makes up for the disadvantage, with the longest retention time reaching to 48 h. There may be two explanations for the durable tumour accumulation. First, doubling the binding monomer can double the molecular weight of the targeted tracer, leading to a long circulatory half-life and slow clearance. Second, the HDL receptor SR-BI is critical for cholesterol haemostasis, and the ApoA-I mimetic peptide (FAEKFKEAVKDYFAKFWD) could function as HDL and possibly increase the efficiency of receptor-mediated endocytosis, contributing to its remarkable accumulation in tumours. Previously, several studies underscored the use of reconstituted HDLs as target-specific nanoparticles for drug carriers to tumours, and highlighted that these peptides can serve as safe targets for a number of cancers [31,32]. In addition to the good performance of the reported peptide, IRDye 800CW was selected because of its safety and preliminary success in several clinical studies [19,33]. These studies have reported its low autofluorescence and good quantum yield in receptor-targeted fluorescence-guided surgery and, most importantly, its compatibility with both the fluorescence-guiding system and optoacoustic scanning system. These findings encouraged us to examine the capacity of the dual-targeted tracer to detect residual tumours and delineate tumour margins in optical and optoacoustic-guided surgery.
The imaging strategy used in our study has the potential for translation into clinical practice. First, both fluorescent and optoacoustic imaging systems were approved by the US Food and Drug Administration. Second, a fibre-delivered optoacoustic guide was previously reported, which mimicked the traditional guide wire and allowed intraoperative placement into the tumour mass to visualise the tumour with sub-millimetre variance [34]. Considering the compatibility of the flexible fibre optoacoustic guide with any conventional endoscope and the advancement of hand-held optoacoustic scanning systems [23,35], it can be incorporated into endoscopic nasopharyngectomy. Optoacoustic imaging possesses the advantages of both optical and ultrasonography that offers a superior vertical resolution at a depth of 2-4 cm and excellent vascular network mapping [36]. Considering that fluorescence-guided endoscopy has been widely reported [37,38], the complementation of optoacoustic imaging would compensate for the limitations of optical-guided surgery [39,40]. Altogether, a possible clinically related image-guiding strategy for endoscopic surgery in rNPC should be as follows. Optoacoustic imaging can be used for localisation with tissue microvasculature and tumour infiltration depth at the microscopic scale. Then, NIR imaging can be used to depict tumour boundaries on a macroscopic scale.
There are several limitations in our study. Firstly, we only selected the peptide FL-35 that had the same molecular weight as FY-35, but only targeted SR-BI as a single-target control. To evaluate both, two separate peptides as control would be more persuasive. Secondly, while we successfully performed tumour resection under fluorescent and optoacoustic imaging guidance, there was a lack of analysis concerning the volume of intraoperative bleeding, which would be worthy of testing in our future study.

Conclusions
Our data demonstrated the power of the dual-targeted tracer for optoacoustic and fluorescence-guided surgery. Utilisation of this method enables the detection of residual tumours and tumour delineation in NPC in vivo, providing a basis for image-guided surgery. Combining the advantages of optical imaging (i.e., wide field-of-view) with optoacoustic scanning (i.e., high-resolution and vascular network mapping), this method holds promise in guiding salvage surgery for rNPC, which might reduce reoperation rates and shorten surgery time.