Precisely Controlled Surface-enhanced Raman Spectroscopic Nanoprobes for In Situ Bioimaging with High Signal-to-Noise Ratio

WANG Li-xiang，SHI Yang，CHEN Feng*，CAO Yue*

（1．Department of Forensic Medicine，Nanjing Medical University，Nanjing 211166，China；2．Department of Pharmacy，Nanjing Medical University，Nanjing 211166，China）

Abstract：Visualization of biological behavior is essential to basic and clinical medical research. Surface-enhanced Raman spectroscopy（SERS） as a rising powerful optical technology attracted a lot of interest due to noninvasive，highly sensitive，rapid and multiplex characteristics. Despite in vitro molecular sensing is highly sensitive，it is not possible to real-time visualize the spatiotemporal dynamics of small molecule functions in vivo. The existing SERS imaging speed and accuracy isn’t enough for the demand of in vivo imaging. This review focuses on the update breakthrough of the development of SERS imaging through more sensitive，controllable and accurate nanoprobes，such as gap-enhanced core-shell nanoparticles，semiconductor composite nanoparticles，Raman silent-region probes. SERS imaging is increasingly being studied for bacteria，tumor surveillance and photothermal therapy. When preparing SERS probes，it is necessary to consider the clearance of background signals，the penetration depth of in vivo detection，the targeting of targets，the non-destructive nature of normal tissues，and the in vivo clearance of probes. Further，we illustrate the application of SERS in vivo imaging ranging from cell imaging to tissue imaging. Finally，we provide prospects on the possible obstacles of SERS bioimaging in future development. Encouraging application results show that SERS imaging has great potential for clinical disease prevention，diagnosis and treatment.

Key words：bioimaging；surface-enhanced Raman spectroscopy（SERS）；theranostics

The physiological and pathological behavior of biological molecules，such as enzyme activity and oxidation and reduction metabolism，is determined by the microenvironment in vivo［1］. Disorders in these variables may serve as clinical indications of diseases like cancer and coronary heart disease. Therefore，precise，quick，and sensitive detection and real-time monitoring of these features are crucial for the early identification of clinical disorders and the evaluation of treatment effectiveness.

Surface-enhanced Raman spectroscopy（SERS） has gained increasing attention since its discovery. When the incident light excitation wavelength is coupled with the substrate resonance，a local electromagnetic field can be generated on the metal surface，which is known as local surface plasmon resonance（LSPR）. The hot spot generated can greatly enhance the Raman signal of the analyte. SERS，as an emerging optical imaging technology， presents the following advantages than other imaging techniques： ①unique fingerprint information； ②ultrahigh sensitivity for non-invasive detection； ③multiplexing capabilities； ④antiphotobleaching and anti-photodegradation［1］. Nowadays， SERS bioimaging involves various kinds of applications.

Firstly，it’s available for biochemical characterization of cells. The multiplex synchronous bioimaging of biomarkers，such as intracellular dysregulated proteins and abnormally expressed microRNAs in exosomes quantification and spatial positioning，has attracted much attention in clinical diagnostic applications such as early clinical screening， rapid efficacy evaluation， real-time monitoring， and postoperative evaluation，especially distinguish different kinds of cells like normal and malignant cells at the molecular level by aptamer or peptide targeted SERS probes. These biomarkers include classical matrix metalloproteinases（MMPs），epidermal growth factor receptor（EGFR），P-glycoprotein，and insulin-like growth factor 1（IGF1）［2-3］.

Secondly，it’s available for spatial-temporal profiles of tissues. In regard to in vivo imaging of lesions and pathology tissue imaging，it’s essential to for individualized precision treatment to identify the heterogeneity and differential expression of biomarkers［4］. In vivo tissue imaging could be applied into the following aspects：①the location of sentinel lymph node（SLN），which is beneficial for SLN biopsy and assessing cancer staging［5］，② delineating the margin of tumor tissues and identifying the existence of postoperative residual microtumors，which is crucial for intraoperative removal［6］. Recently，SERS probes not only play a key role in in vitro and in vivo imaging，but also act as NIR photothermal anti-tumor therapeutic agent in vivo［7］.

Thirdly，it’s a grand challenge for new strategy update to combine in situ track antibiotic resistance dynamic evolution and highly heterogeneous of bacteria with antimicrobial efficiency. As all known，antibiotic tolerance and resistance are the major cause of infection tough to treat，superinfection induction even to the point that there is no cure. SERS nanoprobes can only supply chemical information based on averaged physiological responses of hundreds of thousands of bacteria due to its much smaller size. D2O isotope［8］，4-MPBA label，surface cell imprinted substrate（SCIS） thus were applied to target bacteria for single cell SERS detection［9-10］. Furthermore， multi-function nanoprobes exhibited robust antibacterial effectiveness by Ag+releasing of AAS-NPs［9］，M13 bacteriophage@Au@AgNR composite［11］，photothermal inactivation by MOS2or graphene based nanomaterials［12-13］.

Although SERS has been successfully employed into bioimaging，the recent developments in SERS bioimaging application especially in situ in vivo imaging have been greatly limited because of poor target，reproducibility deficiency，background interference，limited tissue penetration depth，and simultaneously multiplex imaging demands. Size-induced nonspecific adsorption and incubation time-induced endocytosis，producing false positive results，would interfere detection accuracy. The heterogeneity of samples and different batches of materials reduce the reproducibility of results［14］. The existing SERS imaging speed lags far behind clinical demands. It usually takes hours to acquire in vivo images. Meanwhile，The SERS signal obtained in biological sample may be accompanied by a strong background，which causes the signature Raman peaks of the signal to be flooded by the background，affecting the image quality［15］.

In this review，we will focus on how to break limitations that hinder development of Raman bioimaging technology. First，we’ll summarize the recent progress of SERS platform with more sensitive，controllable and accurate nanoprobes. Further，the application of SERS in vivo imaging ranging from cell imaging to tissue imaging are going to be introduced. Finally，we’ll provide prospects on the possible obstacles of SERS bioimaging in future development.

1 In situ SERS bioimaging platform

In general，to obtain high-speed，high-resolution and reliable images，there are several issues about SERS probes that should be considered carefully：sensitivity，specificity，stability，uniformity. To overcome current limitations，one of the remarkable methods is optimizing design of SERS nanoprobes. Of course，as Raman equipment has advanced，so too has the applications for Raman imaging，which overcame the optical diffraction limit to achieve super-resolution spatial resolution. High sensitivity imaging is achieved by the spatial offset of the excitation point and the signal collect［16-17］. This is different from the traditional laser confocal Raman system，which achieved high sensitivity imaging by sacrificing time resolution［18-19］，to the stimulated Raman system， which achieved high sensitivity imaging with background-free through highfrequency phase sensitive detection，but the high peak power will damage the sample［20］.

A classic SERS probe consists of a plasma nanoparticle core，Raman reporters and targeted ligands with or without a protective coating shell. The sensitivity of noble metal SERS probes is influenced by the electromagnetic field produced by nanogap or sharp tip and the charge transfer between particles and molecules. The stability and uniformity can be obtained by fabricating uniform-sized plasmonic core and ratiometric labels. We will focus on novel and effective nanoprobes which hold promising future in the next paragraphs.

1.1 Anisotropic noble metal nanoparticles

According to higher enhancement effector，the morphology of non-spherical and porous plasmonic core have come into notice. Because SERS intensity usually rely on the creation of sharp tips and rough surfaces through morphologically controlled synthesis，like nanostars，nanoflowers，triangle nanoplates，nanocubes.What’s more，this kind of enhancement tactic does not depend on the aggregation of particles like dimers，and this feature improves the reproducibility in SERS imaging［21］. Numerous sharp tips generate strong electromagnetic field arise from the excitation of localized surface plasmon resonances（LSPRs），which amplify the vibrational signal of Raman molecules by 9-10 orders of magnitude［22-24］.

1.2 Gap-enhanced core-shell nanoparticles

Such nanoparticles like nano stars，nanoflowers，suffer from poor controllability in creating uniform hot spots and generating stable SERS signals. Therefore，the precise synthesis of plasmonic nanostructures with powerful EM fields，as well as precise control of the number and position of Raman reporters，are key to obtaining strong，controllable，stable，and quantifiable SERS signals. At the point，a new type of core-shell SERS nanoprobes，gap-enhanced Raman tags（GERTs），metal core-shell tags with inner nanogaps，have been developed for SERS imaging［25-27］.

It has been shown that the location of RMs in ultrasmall gaps rather than particle surface tips resulted in a higher SERS signal，because signal fluctuations induced by desorption or random aggregation of particles can also be avoided by hiding the reporter molecule in the inner gap［27-30］. By this theory，much spacer，such as thiolated DNA［31］，1，4-BDT，4-MBT［32］and silica［33］have been employed to synthesize nanogap successfully.

In addition to nanogap，the shell structure also has cooperative effect with plasmonic core and intrinsic characteristics that result in a much higher SERS performance. First，protective coating shell，for instance，covalent organic framework（COF） with stable chemical property guarantees the dispersion stability of nanoparticles in solution［34］. Second，hybrid bimetallic shell，the SERS activity of Ag-Au core-shell mainly depends on Ag shell thickness［35-36］. Third，porous and mesoporous shell，likem-silica，can be used for effective drugs encapsulation. At the same time，the pore structure enrich analyte selectively which can not only enhance Raman signal but also prevent interference from other molecules［34，37］. Forth，internal standard（IS），whose Raman characteristic peaks can be used to correct signal fluctuation［34］.

1.3 Semiconductor composite nanoparticles

Semiconductor nanoparticles，as opposed to noble metal material，have already been shown to process higher targeted specificity，stability，reproducibility，anti-interference ability and near infrared adsorption，which are more feasible for accurate quantification detection in complex system and photothermal therapy.However，SERS activity of semiconductor still lag far behind noble metal plasmonic particles. Promoting photoinduced charge-transfer（PICT） resonance owing to flexible controllability of semiconductor nanostructure is the dominating strategy that enhance SERS performance of semiconductor，which attributed to molecular polarizability and Raman scattering cross section increasement［38］.

It was found that the introduction of surface defect can dominantly create PICT activity. The study suggests that face-dependent nanostructure enable efficient charge transfer due to the lowest electronic work function through combining Kelvin probe force microscopy（KPFM） technology and first-principle DFT calculations［39］.Thus， improved SERS performance can be obtained from the generation of abundant defective surface structure，which are introduced into ultrathin 2D-dimensional semiconductor nanosheets，especially metal oxide materials. Zhao’s team demonstrated that the SERS detection of methylene blue（MB） was found with sensitivity close to the nanomolar level based on ultrathin WO3substrate［40］.

Some research found that amorphous semiconductor has a significant effect on enhancing substratemolecule system interaction，which was ascribe to the weak constraint of the surface electrons，resulting in efficient interfacial charge transfer. Wu et al.［38］designed crystal-amorphous core-shell-structured TiO2NPs that integrating the advantages of both crystallic and amorphous semiconductor materials.

1.4 Ratiometric nanoparticles

Changes of single Raman signal-based SERS imaging platform tend to be subject to individual heterogeneity differentials，extrinsic factors from complex biosystem，a new detection method thus is urgently needed. Ratiometric SERS is crucial for heterogenous correction and quantification analysis normalization.This detection framework generally consists of an analyte-responsive unit and a self-referenced Raman reporter unit. In the presence of analyte，signal change in the SERS-molecule system may be caused by the following aspects. First，specific group could be recognized by biomarkers and then chemical bond break apart，while the internal standard module signal remains unchangeable. Liu et al.［41］reported a synthetic module comprised of phenyl-alkyne responsive unit and triisopropylsilyl（TIPS）-terminated alkenyl IS， whose detection is triggered by Glu recognition of GUSB. Second，aptamers modified on the surface of nanoparticles by π-π chemical bond lead to dissociation due to the higher affinity with analytes. Song et al. presented a rhodamine X modified V-shaped double-stranded DNA serve as responsive unit and 4-NTP labeled fixed substrate as IS［42］.Third，some Raman reporter possess intrinsic chemical group switch. Yu et al.［43］modified MPBE onto the surface of probes，the boronic pinacol ester groups of the probes would turn into phenol groups upon reaction with analyte，inducing the significant typical peak changes of SERS probes.

1.5 Raman silent-region SERS probes

Additionally，eliminating background is another effective approach to improve the sensitivity of SERS tags. Furthermore，tissue penetration depth should not be neglected when involved in bioimaging，especially for in vivo bioimaging. The methods to solve this problem as well as background elimination. Generally，background interference includes the following sources：①ambient light；②the intrinsic optical background of metal nanoparticles due to photoluminescence（PL） or electronic inelastic scattering；③autofluorescence interference of biological tissues；④Raman spectral crosstalk in bioimaging［15］.

Then there are corresponding measurements. First，Raman spectral crosstalk and ambient light can be evitable when performed in the Raman-silent region. Finally，near-infrared（NIR） laser excitation is used to effectively reduce tissue autofluorescence and deepen tissue penetration.

Traditional Raman reporters（like crystal violet，rhodamine） fall far behind Raman-silent region Raman reporters in in vivo Raman intensity enhancement. Raman signals of the former locate in the fingerprint region，which overlapped with that of endogenous biomolecules，so that it will affect the detection accuracy. Raman Reporters located in the Raman-silent region（1 800-2 800 cm-1），are perceived as biorthogonal with no background interference. Therefore，many study paid particular attention to alkenyl，azide，nitrile［41，44-45］，whose Raman peaks in the biological Raman-silent region. This feature means that Raman detection and imaging in this region can eliminate background interference，thus providing high signal-to-background ratios for SERS sensing in complex biological systems.

1.6 NIR plasmonic nanoparticles

Currently，the development of plasmonic nanostructures suitable for SERS-imaging both in the NIR Ⅰand NIR Ⅱ window triggered a hot topic. For ideal in vivo imaging，the LPSR of NPs should be shifted towards NIR window（generally 700-950 nm）， where tissue absorption and autofluorescence are minimized.Wavelength in the visible spectrum（<650 nm） have very limited penetration into the tissue due to tissue scattering and background autofluorescence. The NIR window for in situ high-resolution imaging，with deeper tissue penetration，is further broken down into two regions，classified as the first（650-900 nm） and second（1 000-1 700 nm） NIR windows［33，46］. Appropriate nanoparticles and Raman-active molecules which can generate resonance in these regions and possess high sensitivity necessarily involve inordinate amounts of time and investigation.

Through changes in their size and shape，the LSPR of the AuNPs can be tuned to be in resonance with the excitation laser wavelength. However，macro-nanoparticles would inhibit tumor passively accumulation of SERS probes due to EPR effect［47］. In vivo SERS imaging requires appropriate size and large near-field enhancement of nanoparticles［48］. When the electronic excitation energy of reporter molecules（RMs） is close to the incident laser，Raman signal enhancement is several orders of magnitude compared to that on SERS，which is called surface enhancement resonance Raman spectroscopy（SERRS）. This kind of RMs belong to NIRabsorbing dyes，which is typically based on Cy7，such as the heptamethine cyanine derivative（Cy7-SH），IR-780［27，49-50］，IR-780 whose LPSR peaks at 780 nm match well with excitation laser wavelength（785 nm）thus produce additional NIR-SERS［51］. However，the number of NIR-absorbing Raman dyes are limited thus still needed to be developed.

Operating in the NIR Ⅱ window would be of favorable benefits that further reductions in absorbance，scattering，and the autofluorescence background. What deserves to be mentioned is aforementioned plasmonic nanostructures can transfer light energy into photothermal to kill cancer cells，so called plasmonic photothermal therapy（PTT），and targeted releasement of drugs triggered under NIR laser［37，52］. Although SERS imaging in the second NIR window has still been a budding application so far，it does a promising technology for in vivo imaging especially.

2 In situ SERS bioimaging applications

SERS detection technology was initially widely applied into in vitro biological fluid samples，but in vitro detection was not accessible for in situ spatiotemporal information and real-time dynamic monitoring of molecules. Recently，in situ imaging has been attracted increasingly attention. Research of in situ imaging would be of great value in biological and medicine fields.

2.1 Single cell imaging

In order to reveal the spatiotemporal dynamics of biomolecules，in situ high-resolution SERS imaging technology has been widely used in cell imaging. At present，SERS cell imaging is mainly applied into differentiate normal cells from cancer cells，into differentiate cell subtypes and single cell level analysis by biomarkers expression assessment，which is convenient for clinical diagnosis［53-54］. However，classify different grade of malignant cells in detail is a purpose which have a distance with the reality. Biomarkers for specific identification of tumor cells continue to be developed，and in situ detection by endocytosis SERS probes modified with labels has become an emerging topic. There are some factors to be considered in the design of in vivo imaging nanoprobes include endocytosis efficiency， targeting specificity， biocompatibility， and bioenvironmental stability［3，55-56］.

The detection of cell extracts hasn’t satisfied the needs of current in situ research. The dual-mode strategy utilized two or three-channel fluorescence/SERS signal to distinguish cell lines in situ，which broke through the limitation of the study of the interaction between biomarkers’ location and cell physiological characteristics.The distribution of biomarkers was indicated by fluorescent dyes，and targeted precise quantification are labeled by aptamers［57］. Miao and Song’s team，respectively，developed a highly sensitive strategy for catalytic hairpin assembly（CHA） based Signal amplification nano-polymers，which contributed to the development of single-cell in situ imaging in complex environments［58-59］.

2.2 Tissue imaging

Traditional pathological tissue sections can only obtain information from morphological perspective under the microscope，and information at the level of proteins，DNA and even microRNAs rely on technologies such as immunofluorescence staining. But there still remain some problems such as non-specific detection and cumbersome steps. Human tissue ex vivo and mice tissue in vivo imaging challenge the existing SERS technology due to biocompatiblity，penetration depth and target specificity.

Halina’s group demonstrated that the reduced cytochrome c levels upregulating triggers the development of aggressive cancer through monitoring redox state of mitochondrial cytochromes，which promote clinically diagnose cancers and understand underlying biomechanism of cancer development for future improved therapeutic foundation. They set ex vivo human brain tumors and breast cancer to evaluate redox state. More importantly，Raman imaging can be used for invasive grading，which supports precise clinical diagnosis［60］.They found that cytochrome c，cardiolipin and palmitic acid were released from the epithelial cells into the lumen in the cancerous duct in contrast to the lumen in the normal duct. This important mechanism would be a starting point of studying cytochrome c-based anti-cancer development［61］.

Sentinel lymph node（SLN） imaging is also a powerful tool to study lymphatic metastasis of malignant tumors to determine tumor staging and prognosis assessment. Sentinel lymph nodes（SLNs） are theoretically the first site of metastasis of cancer. SERS nanotags as SLN tracer for accurate locating，margin visualization ensures enough retention time intraoperative imaging of the SLN［62］. Current development was mainly the combination between ratiometric dual-SERS tags strategy or GERT and data processing method. Some study demonstrated that the diagnosis of meta-static SLN with a high accuracy of 87.5%［63-64］. To break through background interference，Ye et al.［15］designed a high signal-to-noise ratio probe that could be performed for SLN imaging under ambient light，under the off-resonance mode.

In addition to cancer，sudden cardiac death（SCD） is another disease with a very high a high morbidity and mortality that seriously affects public health. Some study confirmed that multiplexed SERS imaging of vascular inflammation related to atherosis may facilitate early diagnosis of fragile plaques and stratification of patients to receive anti-inflammatory therapy. Maffia et al.［65］accomplished in situ imaging of human coronary artery pathological tissue ex vivo to superficial vascular tissue in vivo，through a single Raman fiber probeconfigured custom-made spectroscopic system. Furthermore，monitoring the efficacy of current treatments for heart disease is also an urgent clinical need. Hu et al. achieved high-resolution and real time monitoring of the in vivo behavior of the therapeutic stem cell，including migration，distribution and metabolism，which improved the clinical transfer of stem cell therapy［66］.

2.3 Intraoperative imaging

Current clinical imaging technologies of tumor surgery mostly rely on CT/MRI，but they are limited to preoperative simple imaging. Additionally，aggressive，and metastatic tumors are too poorly demarcated and indistinctly distributed to de detected. The existence of residual tumor tissue causes recurrence and metastasis，which is almost impossible to be visualized and resected. To remove this microscopic lesion，current clinical strategies rely on intraoperative removal of large margins of normal tissue， or postoperative adjuvant chemotherapy and radiation therapy，which have potential side effects. Thus，it is essential to establish precise SERS imaging technology for surgery guidance.

There are various practical application values in the development of SERS imaging techniques for cancer detection.①early cancer diagnosis，quantitative monitoring of cancer development，and real-time assessment of treatment efficacy which need high reproducible imaging［41］. ②guide surgery through differentiating diseased tissue from normal tissues and reduce re-excision in patiens with residual tumor found at surgical margins［67］.③SERS imaging-guided precision photothermal treatment［7］.

For in vivo tumor therapeutic strategy，tissue detection depth at only millimeter scale so far. Regardless，the present level is enough to identify microtumors following surgical resection of primary tumors. Yu’s team reported a “three-in-one” theranostic nanoprobe for intraoperative SERS-guided complete elimination of residual microtumors. They successfully identified multiple residual tumor foci and several satellite residual microtumors on the tumor-normal tissue boundary while real-time clearing them. Intraoperation implemented in the SERS-positive position with 808 nm laser irradiation（1 W/cm2） by local plasmonic heating. Subsequent examination results indicated complete inhibition of tumor growth without recurrence［68］.

However，clearance of nanomaterials from the imaged tissues，severely limited SERS imaging in vivo applications. Qiu et al.［69］explored a novel degradable material mainly consisted of hollow CuS. The laserinduced photothermal effect causes CuS NP to decompose from the shell structure into individual crystals，thus implementing self-clearance. As a SERS-active substrate，hollow CuS NPs also enable substantial Raman enhancement and photothermal therapy in vivo. This new CuS SERS probe with photodegradable properties opens avenues for Raman imaging to countless biomedical applications.

2.4 Bacterial imaging

Bacteria imaging is another hot issue in the field of biomedical and public health. However， the application of SERS in the area of bacteria imaging has fallen far behind of cell imaging due to sensitivity limitation. Generally，SERS-based bacteria imaging has been divided into two approaches：label-free［70-71］and label-based［72-73］. three-dimensional imaging of bacteria is used to investigate chemical composition，spatial distribution，cellular responses，and genetic mechanisms. And in this regard，the label-free method has an excellent advantage over than the label-based one cause tags may interact with molecules within bacteria，affecting their biochemical characteristics. To acquire physical and chemical information of biomolecules，machine learning［74］has been combined with spectral technology to process complex data from subtle signal changes in the spectrum. Liu et. al［71］successfully employed in situ biosynthesized Ag NPs combined with PCA and MCR model to explore and deduce the bio-reduction mechanism within a single bacteria cell.

The purpose of bacterial imaging is to predict and screen drug-resistant bacteria and dynamically track their evolutionary traces，so as to accurately kill bacteria. Antibiotic abuse leads to the emergence of multidrugresistant（MDR） bacteria，which threaten health severely. Therefore，it’s essential for real-time monitoring of bacterial infection and detection of treatment efficacy. He’s group conjugated a positively charged hollow SERS probe，AuAg nanoshells@DTTC，preferentially bind to bacterial membranes with more negatively charged comparing with normal cells，to specifically target residual bacteria to assess the anti-bacterial treatment efficacy with high sensitivity. They proposed that the electrostatic interaction increasing affinity to bacteria may be one of the most important interactions which contribute to sensitivity increasing. What’s more，nanoparticlemediated photothermal therapy has a good anti-bacterial efficacy［72］.

However，the photothermal therapy damage normal and infectious tissues indistinguishably，even if they aggregated the bacterial specific targeting groups. It’s impossible to totally prevent the non-specific binding.Therefore，it’s necessary to develop a new strategy to meet the standards for both SERS imaging and the photothermal therapy. Wang et al.［70］employed bioorthogonal reaction into the process of gold nanoparticles aggregation. The key point to ensure no damage to healthy tissues is the specific binding affinity of Van-TCO towards gram-positive bacteria because GNPs could only aggregate in the present of Tz and Van-TCO.Therefore，the biorthogonal reaction steps are easier，more sensitive than traditional label-based methods for SERS imaging and PTT. Xu’s team［73］presented another rational design that photoactive generation of ROS result in controllable silver etching. Based on this， they achieved kill MDR bacteria precisely with dramatically reduced acute and systemic toxicities.

3 Conclusion

In this review，we summarize the design and development of highly sensitive and high signal-to-noise ratio SERS labeling for biological imaging，then introduce how to break the technical barriers of SERS in biological imaging，and finally the latest progress of SERS-labeled nanoprobes in live-cell imaging，pathological tissue imaging，in vivo imaging，and bacterial imaging detection. SERS labeling offers significant advantages for developing detection and bioimaging platforms，including high sensitivity，non-destructive and photostatic，multiplexing，low external interference，and ease of integration of multiple functions. It is ideal to develop multimodal probes for biomedical applications by integrating multiple functions such as diagnostics，therapeutics，into a single SERS probe. In vivo imaging still be underdeveloped，most applications are currently in the initial stage. Further efforts should remain focused on finding high-performance plasmon or nonplasmon nanostructures，including tuning their optical properties and optimizing material composition，as well as SERS-labeled surface ligands. Although highly sensitive targeted tissue imaging is expected，there are still problems with insufficient penetration depth，probe carryover，and imaging time. Especially，the clearance of nanoparticles injected in vivo are still a concern for unknown chronic toxicity，though pre-biosafety test demonstrated that nanoparticles have no significant toxicity cancer-targeted agents can minimize the nanoparticles accumulation in the normal organs. Effective strategies for rapid in situ precision SERS imaging should be further explored. We believe that interdisciplinary collaboration with scientists and clinicians from different specialties will undoubtedly transform the new thinking of the development of the SERS platform for biomedical applications，leading to new applications in the clinic.