The multifunctional fiber probe was realized by immobilization of Gr/AuNS hybrid nanomaterials on the fiber taper (Fig. 2a). At first, graphene was activated for the introduction of reactive functional groups 33. To endow activated graphene with enhanced photothermal conversion efficiency and address the fluorescent quenching mediated by graphene, Gr/AuNS suspensions were prepared by mixing the synthesized AuNS with the activated graphene. To integrate Gr/AuNS onto the optical fiber, the procedures of hydroxylation, silanization and immobilization are followed sequentially as reported previously 15,18. Activated graphene with enriched functional groups and AuNS can both strongly interact with silane through hydrogen bonding and electrostatic adsorption, and thus generate stable Gr/AuNS composite on the fiber 34. In the following, the NTR fluorescent molecules, which can emit green fluorescence via blue light excitation, was also tethered to the fiber by the covalent bonds. The tailored AuNS (S10) with the plasmonic peak of 550 nm can be orchestrated with the fluorescent emission, heralding the enhancement of fluorescent signal.
Figure 2b showed the preparation procedure of the iFOT and the microscopic photograph of the fiber tip. The cone structure of the fiber tip cannot only facilitate the tissue intervention, but also enhance the interaction between the delivered light and NTR-activated fluorophore via evanescent field. The configuration of Gr/AuNS decorated fiber (Gr/AuNS fiber) sensor was confirmed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM) (Fig. 2 and Fig. S1). As shown in Fig. 2c, Gr/AuNS were dispersedly attached on the surface of the 150 µm fiber tip, providing the basis for enhanced sensing and photothermal conversion. Two-dimensional graphene sheets with surface diameters ranging from 2 to 9 µm were observed to provide favorable anchor sites for AuNS, which were randomly distributed on and around the graphene sheets. This result was further confirmed by TEM images (Fig. 2d). As shown in Fig. 2e, AuNS with an average aspect ratio of 70 nm could be observed on graphene. Compared to gold nanospheres, the engineered AuNS possessed the emission-like branches that can drastically enhance the local electromagnetic (EM) field to boost fluorescent signal. The selected area electron diffraction (SAED) pattern displays two sets of diffraction rings that can be indexed to the graphic structure (blue rings) and the gold structure (red rings). It corresponds to the two kinds of lattice fringes shown in high-resolution TEM (Fig. 2f). The lattice spacing of about 0.34 and 0.24 nm corresponds to the (002) planes of the hexagonal graphite structure of graphene and the (111) plane of cubic Au, respectively. Figure 2g demonstrated the elemental mapping images also confirmed the distribution of graphene and AuNS directly.
Figure 2h showed the XRD patterns of the pure graphene and the Gr/AuNS with different interfaces. The diffraction peaks of Gr/AuNS in the XRD pattern matched well with those of graphene (JCPDS No. 41-1487) and Au (JCPDS No. 04-0784), which suggested the successful combination of graphene and AuNS. Raman spectra of the Gr/AuNS exhibited D-band at 1315 cm− 1 and G-band at 1580 cm− 1 (Fig. 2i), respectively, which is the typical characteristic of graphene. Due to the sensitization mechanism of AuNS, the Raman spectral signal of Gr/AuNS is remarkably higher than that of graphene. Moreover, the UV–vis spectra in Fig. 2j showed that the Gr/AuNS had a broad-spectrum absorption and relatively higher absorption than pure graphene in the near infrared light window, indicating that the hybridized AuNS could boost both the fluorescence intensity and photothermal conversion to the pure graphene.
To achieve fluorescence-based detection and photothermal actuation simultaneously with the same fiber probe, we designed the fiber probe by customizing the diameter of the fiber end and modulating the decorated nanomaterials as shown in Fig. 3a.The light-mater interaction with regard to the fluorescent molecules and photosensitizers on the fiber surface can be directly influenced by the thickness of fiber probes, which provide different intensity of evanescent fields. For another, the structural morphology also determines the mechanical strength of fiber probe. Therefore, it is important to balance the evanescent field utilization and rigidity of fiber probe. Furthermore, the interference between fluorescent agents and photosensitizers on a fiber probe needs a deeper understanding to strengthen both effects.
The evanescent field utilization, which can be represented by the photothermal effect and fluorescent intensity, related to the diameter of fiber probe is described in Fig. 3b.As the diameter of fiber-end decreases, the photothermal conversion efficiency is enhanced, but the efficiency curve presents a flat region as the diameter is less than 150 µm due to the evanescent field is already taken fully advantage of below that diameter, enabling a presence of ~ 186 oC in the air under the incidence of 250 mW pump light. By contrast, the fluorescent sensing curve has a distinct style. The fiber probe has the lowest Limit-of-Detection (LOD) of 2.6 ng mL− 1 in the calibration test when the fiber end diameter is designed at 150µm (Fig. S2 and Fig. S3). This parabolic-like curve implies there should be a compromise between the intensity of evanescent field and the amount of fluorophore loading regarding the thickness of fiber. Therefore, the fiber with an end diameter of 150 µm was selected as the optimal probe to realize the integrated theranostics.
More importantly, the optimized structure of fiber probe can also ensure its mechanical strength, which is suitable for realizing in vivo test. Two types of fiber end with different diameters (D = 150 µm/TL = 5 mm and D = 40 µm/TL = 10 mm; D: diameter, TL: transition length) were fabricated to improve the understanding on the relationship between the mechanical strength and the structure of fiber. In Fig. 3c, after a two centimeter of lateral translation, the thicker fiber probe (which is preferred to fabricate iFOT) maintained its straightness, while the thinner microfiber exhibited bending. Subsequently, to evaluate the feasibility of in vivo operations of the fiber probe, we further tested the rigidity of the two fiber probes through puncturing them into the soft rubber. As shown in Fig. 3d,the optimized fiber probe (150 µm) can easily penetrate through the white rubber, but the microfiber was broken during the piercing process, and the original fiber (D = 400 µm) also failed to stab into the rubber (Fig. S4). The results reveal that the optimized needle-like head structure not only allows a high pressure exerting on the contact surface to realize paracentesis, but also yields the pronounced counter-bending ability-a key feature for the clinical operation, especially the interventional surgery, which relies on the catheter and puncture needle to reach the lesion. A highly bendable head structure will undermine the compatibility of the probe to the interventional surgery due to the lack of controllability in the rigid needle and silky ductility in the soft catheter. The rigidity of the optimized probe head derives from the larger area inertia moment and sectional resistance moment provided by the gradually diameter changed (cone) structure (Fig. S5 and Fig. S6).
For tumor PTT, the temperature elevation of the probe is prioritized. Graphene is a perfect photothermal absorber and its quantity of decoration is crucial to determine the photothermal efficiency. As the graphene quantity of decoration increases, represented by the density of the graphene suspension in the decoration, the iFOT can reach higher temperatures under the same laser power stimulation. As is shown in Fig. 3e, the fiber probe using graphene suspension of 5 mgmL− 1reached 155°C, showing a significant temperature elevation compared with the suspension of lower density. However, further increasing the density of graphene, for example, to 10 mg/mL did not bring about a greater enhancement of photoheating outcome due to the limited capacity of surface area of the fiber end, but would severely quench the fluorescent intensity for TME sensing (Fig. 3f).
AuNS was then introduced to mitigate the fluorescent quenching resulted from the presence of graphene and potentiate the fluorescence detection efficiency. As shown in Fig. 3g, the bare silica fiber, Gr only fiber, and Gr/AuNS fiber were employed to make comparison regarding the fluorescent sensing performance in the PBS-NTR-fluorescent molecule mixture solution. Due to the fluorescent quenching effect of graphene, the fluorescence signal intensity collected by the Gr fiber sensor was weaker than the bare silica fiber. Taking advantage of the electromagnetic enhancement effect of the AuNS, the Gr/AuNS fiber could not just overcome the graphene induced deterioration of fluorescent signal but even enable a 1.5 times higher fluorescent intensity compared with the bare silica fiber. The fluorescence intensity of the fiber probe samples was in agreement with the photographs under of fluorescence microscopy. The effect of the morphology of gold nanostructures on the intensity of the fluorescence signal was analyzed and customized to obtain the high fluorescence signal (Fig. S7). Like the photothermal characterization using graphene, the fluorescence signal can as well enhance by increasing concentration of AuNS colloidal solution and the saturation concentration regarding the fluorescent intensity enhancement was 0.1 nM, as is shown in Fig. 3h. The reason can also be attributed to the limitation of space for anchoring AuNS.
To characterize the fluorescence sensing properties of iFOT probe, we conducted in vitro calibration by inserting the fiber-tip probes into PBS-NTR-fluorescent molecule mixtures with different concentrations of NTR (Fig. 3i). The normalized intensity of the emission peak enhanced with the increase of NTR concentration (Fig. S8). For the wide NTR concentration range in Fig. S9, the measured points of the peak intensity ratio can be well depicted by logistic fitting (R2 = 0.989). And at the range of lower concentrations ranging from 0 to 30 ng mL− 1, an approximately linear correlation was deduced, presenting the LOD of 2.6 ng mL− 1 and limit of quantification (LOQ) of 8.5 ng mL− 1. The long-term stability and high-temperature resistance of iFOT were also verified, as is shown in Fig. S10.
Surprisingly, AuNS not only enhanced fluorescence, but also further upgraded the photo-thermal efficiency of the fiber probe. Figure 5b showed the temperature rising of different fiber probes in the air under laser pumping. It can be seen that both Gr and Gr/AuNS fiber can achieve high local temperatures (> 100°C) and the presence of AuNS provides an additional elevation of local temperature due to the gold plasmonic effect 35 (Fig. S11). Combined with the above description, we achieved the optimization strategy of the iFOT probe, which is important for the implementation of fluorescence detection and photothermal therapy in vivo.
Since the local endogenous hypoxia is a significant characteristic for the TME, the hypoxia-associated biomarker, NTR, was selected as the tumor-associated marker 36. The distinction between normal tissue and solid tumors could be evaluated by the presence of the NTR. NTR fluorescent molecules were immobilized on the fiber tip using the chemical bonding method 15,18. The sensing principle of the iFOT probe for tumor diagnosis is shown in Fig. 4a. The presence of NTR results in the reduction reaction of fluorescent molecules, which would then emit strong green fluorescence at 550 nm, as they were excited by 450 nm blue light. To characterize the fluorescence sensing properties of iFOT probe, we injected the PBS-NTR-Fluorescent molecule mixtures into pork tissue and used iFOT to detect NTR in situ. The result shows that a significantly positive response can be obtained in contrast with the test on the blank sample (Fig. S12).
Further, we conducted in vivo experiments using tumor-bearing mice to demonstrate the feasibility of iFOT for solid tumor detection. As shown in Fig. 4d, the tumor tissues in mice bearing orthotopic breast cancer and the corresponding sites of normal mice were punctured and detected through the iFOT probe, separately. For tumor detection, the fluorescent signal responds rapidly and offers a positive peak in 10 s. With the accumulation of NTR fluorescent molecules and NTR reactions, the signal intensity enhanced 1.7 times within 1 min and the wavelength was redshifted up to 5.27 nm on the spectrum. Conversely, no obvious fluorescence signal was detected in normal mouse tissues tested by focused fluorescence for 1 min in the same way. Compared to the normal tissue, the tumor trial showed positive results within the same detection time. Notably, the different fluorescent signals were also observed at different sites of the same tumor (Fig. 4e and Fig. S13). The intensity of fluorescence signal was proportional to the tumor density. The results indicated that the fluorescence signal was highest in the central region of tumor while was lowest at the tumor edge, which was significantly variable compared with normal tissue. Moreover, iFOT can evaluate NTR levels in vivo after treatment (Fig. S14). Furthermore, the brief structure and rapid response of iFOT were visually displayed in video S4. Taken together, these results showed the iFOT can be used for rapid solid tumor detection in parallel with an effective tool for tumor margin detection. It is promising as an effective tool for intraoperative precise tumor localization.
Next, we investigated the solid tumor PTT capability of the same iFOT probe. As shown in Fig. 5a, taking advantage of the excellent photothermal conversion efficiency of graphene, the graphene decorated fiber can release a large amount of heat in the surrounding area under 980 nm pump laser excitation, which is essential for tumor PTT.
Then, the photothermal imager was used to characterize the thermal radiation range of Gr/AuNS fiber tip to verify the effectiveness of fiber PTT and the action interval of PTT in vitro pig liver tissues (Fig. 5b). The range of fiber thermal radiation was proportional to the power of the pump laser. After 5 min laser radiation of 200 mW, the photothermal capacity reached the peak, and the diameter and average temperature of the heated region were 9.9 mm and 42.6°C, respectively. Compared to the previous reported PTT fiber18, Gr/AuNS fiber performs a higher efficiency of photothermal conversion. (Fig. S15). In addition, iFOT probe also has good biosafety thanks to the low-power laser and non-residual nanoparticles (Fig. S16).
To verify the PTT efficacy for tumor treatment, we conducted the iFOT therapy experiment in vivo. For characterizing the PTT effect in vivo, the photothermal imager and an additional optical fiber Bragg grating sensor were used to monitor the temperature of the lesion area. As shown in Fig. 5c, the tumor temperature was quantified in the pump power range of 0-250 mW, and the tumor temperature increased as the pump power was increased. When the power went up to 200 mW, the tumor temperature reached above 43°C, which was a proper temperature to necrotize the tumor cells while maintaining the normal tissues within minimal damage. Then, we had recorded an in vivo PTT cycle in real-time (Fig. 5d). As the 200-mW pump laser was transferred to the fiber-tip probe, and the tumor lesion area showed a rapid and clear temperature response as the pump was "off-on". After 250 s of treatment start, the tumor temperature remained steady at 43.5–45°C, and when the pump was turned off tumor temperature decreased to normal within 200 s. Moreover, during the 15 min-treatment as illustrated in video S2, the photothermal imager data and the fiber optic temperature sensor measurements showed unanimous fluctuations to confirm the reliability of the multifunctional fiber-based photothermal treatment strategy.
Then, the therapeutic efficacy of the iFOT probe was further evaluated by the in vivo group study of tumor bearing mice. The female BALB/c nude mice bearing MDA-MB-231 tumors (20 g of average body weight and 70 mm3 of average tumor volume) were randomly assigned to four groups, including control, Gr/AuNS fiber without laser radiation(F + G-L), bare fiber with laser radiation (F-G + L), and Gr/AuNS fiber with laser radiation (F + G + L). The “F + G-L” and “F-G + L” groups were set to rule out the independent variable influence of the graphene and laser.We conducted only one operation of treatment lasting for 15minfor each mouse. The tumor volumes and body weights of the mice were observed continuously every 3 days (Fig. 6a). Figure 6b and Fig. 6c showed the evolution of tumor volume after treatment in each group, respectively. The mice in the “F + G + L” group displayed significant tumor suppression due to the PTT effect. The average tumor volume was approximately only 7.62% of the blank group at the 30 days after treatment. Quantitative analysis reveals the final efficacy of PTT at day 30 (Fig. S17). However, intriguingly, the tumor volume of partial controlled groups grew faster probably due to the stress reaction of the tumor after the exogenous stimuli. It implies a considerable antitumor effect of PTT using Gr/AuNS fiber probe, as is further demonstrated in continued individual recordings of mice (Fig. 6d). The mice of “F + G + L” group exhibited visible ablative marks on the flank immediately after treatment and formed eschars after 2 days. After 18 days, the eschars fell off and no signs of tumor recurrence were observed in the mice of the treatment group, with a high complete cure rate of 83.33%.By contrast, the mice in the controlled groups did not show any eschar and tumors were not suppressed after treatment. Therefore, fiber puncture, Gr/AuNS materials and laser cannot eradicate solid tumors independently.
To verify the therapeutic effects of the PTT strategy, further histological reports were developed as shown in Fig. 6e. Hematoxylin and eosin (H&E) staining showed that the most severe tumor cell damages were observed in the group of “F + G + L”, while control groups had much lower degree of damage. Moreover, in order to further reveal the mechanism of PTT, the Ki67+ tumor cells and Cleaved-Caspase 3+ tumor cells in each group were examined by immunohistochemical staining. Compared to other groups, the number of proliferative tumor cells was significantly decreased while the apoptotic tumor cells were drastically increased in the group of “F + G + L”. These results indicate that fiber PTT can effectively kill tumor cells and suppress tumor proliferation to achieve the purpose of anti-cancer treatment.
To cope with the deep-seated lesion of the tumor, endoscope and intervention operation are feasible pathways for driving the fiber-optic theranostic probe into practice. However, the guidance of optical fiber to the appropriate position of tumor remains challenging. Fortunately, regarding the anti-interference nature and pliability of optical fiber, the medical imaging technologies can greatly facilitate the navigation of the optical fiber in this new strategy. For example, the US-guided optical needles probe 37, the PAI-guided percutaneous needles 38, and the optically controlled MR-compatible active needle 39 are excellent paradigms that orchestrate the optical fiber and image navigation technologies. Therefore, in order to assess the viability of merging the proposed iFOT with the medical imaging techniques.
Several routinely used imaging methods were tested, including endoscopy, US, PAI, and MRI in Fig. 7a. As shown in Fig. 7b, at first, the optical fiber could be flexibly manipulated using the endoscopic trocar that was integrated by a medical laparoscopic training equipment. The thermochromic material was deployed inside the abdominal cavity phantom to reveal the photothermal effect of Gr/AuNS fiber. Furthermore, as is shown in video S3, the color variation according to moving track of fiber probe heater (following the handwriting as "J-N-U") on the thermochromic plate was confirmed by the temperature value indicated by the distributed FBG sensors. The visibility and controllability revealed the ability of the iFOT probe to perform as an endoscopic diagnostic and therapeutic tool in minimally invasive surgery. In the following, the iFOT probe, which was intervened into the tumor of mouse in vivo, was displayed under US imaging, as is shown in Fig. 7c. The tumor morphology and location of the intervened fiber could be clearly observed in the field of view. A video of fiber that slowly withdrew from the tumor was recorded to demonstrate the real-time feedback of the position of the iFOT regarding the tumor under the US imaging (Video S4). In addition, the iFOT could as well gained the assistance under PA imaging(Fig. S18).
The medical care compatibility of iFOT was also demonstrated by employing the theranostic probe in the minimal-invasive interventional surgery for deep lesions. As is shown in a phantom test (video S5), iFOT can be easily conducted into the microcatheter to navigate the natural channel inside the body, such as blood vessel of artery and vein, for targeting tumor model. After passing the curved soft tube mimicking blood vessel through the microcatheter, the probe reached the tumor phantom, and the optical diagnosis and photothermal treatment could be subsequently performed. Thanks to the flexibility and robustness, iFOT can be crossed through microcatheters and vascular models with different curvatures from 0 to 20 m− 1, as is shown in Fig. 7d. The results reveal that the iFOT can unleash its potential in interventional treatment for tumor, which would have been out-of-reach.
Finally, we demonstrated further possibilities for in-situ diagnosis and treatment of the iFOT under MRI taking advantage of the electric isolation and anti-electromagnetic interference characteristics of optical fiber. Similar to the above imaging techniques, MRI could also visualize the tumor and punctured fiber tip in vivo (Fig. 7e). Meanwhile, the MRI supports 360-degree scanned images and continuous monitoring with a stable field of view. As shown in Fig. 7fand video S6, the entire process of fiber movement in the tumor was recorded by MRI. The high-resolution images provided by MRI enable accurate targeting and positioning of the fiber, ensuring that iFOT can effectively focus on the lesion. Furthermore, the optical fiber pathway enables the possibility of remote sensing and controlling, which can bridge the magnetic resonance receiving coil and optoelectronic instruments, the latter of which are vulnerable under the extremely strong magnetic field (Fig. S19). Real-time monitoring of tumor variation in mice with MRI while using the PTT of fiber probe in the tumor (Fig. 7g). The continuous scan with a total scan time of 10 min (assessed every 100 s) was conducted. To perform a quantitative analysis of tumor T2-weighted signal intensity (SI), we delineated region-of-interests (ROIs) at the center of tumor. The average SI of these ROIs was computed and subsequently normalized using the signal intensity of the adjacent normal tissues. As the PTT continued, the tumor in the vicinity of the fiber tip turned darker in the image and the SI decreased from 2.69 to 2.17 in Fig. 7g, which indicated the decrease of hydrogen ions that was caused by heat mediated tumor dehydration. With the support of the image navigation terminal, we were able to guide the fiber probe to the deeper tumor lesions in vivo. Image-assisted fiber probes can accurately detect the density of tumor cells and even identify tumor margins. The dual tumor diagnosis strategy further enhances the success rate of PTT. Hence, regarding the high compatibility of the iFOT with the medical imaging techniques, the combination of iFOT probes and imaging tools can be an important advance in ensuring the efficiency and biosafety of the action of the optical fiber probe in vivo.