Raman scattering (RS) is a spectroscopic analytical technique based on molecular vibrational characteristics. The underlying principle is that when incident light interacts with molecules, the energy of the photons undergoes a minute change, resulting in a frequency shift in the scattered light. This frequency shift is correlated with the molecular vibrational modes. Consequently, Raman spectroscopy can provide fingerprint-like information about molecular structures and is widely used in fields such as chemistry, biology, and materials science. However, the scattering cross-section of conventional Raman scattering is extremely small, on the order of 10^-6~10^-8of the incident light, leading to an extremely weak signal. In addition, background fluorescence interference and low detection sensitivity have limited the application of Raman technology in trace analysis. In 1974, Fleischmann et al. observed an anomalously enhanced signal from pyridine molecules on a roughened silver electrode surface. In 1977, Van Duyne and Jeanmaire et al. revealed that the underlying mechanism stems from the enhancement of local electromagnetic fields by noble metal nanostructures, and formally introduced the concept of surface-enhanced Raman scattering (SERS)^[17-18]. With the advancement of laser technology and nanomaterials and nanotechnology, the synergistic mechanisms of electromagnetic field enhancement and chemical enhancement have been progressively elucidated, and the base materials such as gold and silver have been optimized, laying the theoretical foundation for single-molecule detection.

Since the 21st century, breakthroughs in nanotechnology have driven the rapid development of SERS. By precisely tuning nanostructures (such as core-shell structures and nanogap arrays) to generate “hotspots,” sensitivity has been significantly enhanced, and applications have expanded into biomedicine, environmental monitoring, and public safety^[19-20].Current research focuses on non-precious metal substrates, smart responsive materials, and optimization of single-molecule detection, while also exploring multimodal integration with microfluidics and machine learning. In the future, SERS is expected to achieve broader practical applications in areas such as precision medicine and real-time environmental monitoring, and its theoretical refinement and technological innovation will continue to drive advances in analytical science.

The introduction of nanomaterials is pivotal to the development of SERS technology. Nanomaterials, particularly noble-metal nanostructures (such as silver and gold) and two-dimensional nanomaterials (such as graphene), have demonstrated tremendous potential in the diagnosis of various diseases^[21-23]. As shown in Figure 1, SERS-based disease detection methods have attracted considerable attention. Under the influence of nanomaterials, Raman signals can be enhanced by several orders of magnitude (typically by a factor of 10³to 10⁸)^[24-26]. When molecules adsorb onto the surface of specific structures (such as noble-metal nanostructures), the Raman signal is significantly enhanced through two mechanisms: electromagnetic enhancement and chemical enhancement^[27-28]. Electromagnetic enhancement typically occurs on the surfaces of noble-metal nanostructures and arises from the excitation of localized surface plasmon resonance (LSPR) (Figure 1A). When incident light interacts with metal nanoparticles, the free electrons on their surfaces undergo collective oscillations, generating a strong electromagnetic field near the nanostructure. This localized enhancement of the electromagnetic field can significantly increase the intensity of the Raman signal. The LSPR phenomenon can be excited by electromagnetic waves ranging from the ultraviolet to the visible spectrum and constitutes a key mechanism underlying SERS enhancement in noble-metal nanostructures. However, this enhancement mechanism is generally not applicable to most semiconductor nanomaterials, as they typically lack the free-electron density required to generate LSPR^[29-30]. In semiconductor nanomaterials, chemical enhancement is the primary source of SERS signal enhancement (Figure 1B). This mechanism relies on electronic coupling between the molecule and the substrate, which is influenced by the material’s energy-level structure and electron-state density, including the bandgap energy and the positions of the highest occupied molecular orbital and the lowest unoccupied molecular orbital. When a molecule adsorbs onto a semiconductor surface, charge-transfer processes may occur, thereby enhancing the Raman scattering signal. Although the effect of chemical enhancement is usually several orders of magnitude weaker than that of electromagnetic enhancement, it offers greater selectivity and is suitable for certain molecular systems that cannot be enhanced via LSPR.

SERS can achieve single-molecule detection largely because molecules adsorbed onto nanostructured metals, such as the surfaces of gold or silver nanoparticles, experience an enhancement in inelastic light scattering (10⁶~10⁸)^[31-32]. Compared with other high-sensitivity techniques, SERS detection does not require complex sample preparation procedures, thereby reducing pre-processing steps. Overall, SERS is rapidly emerging as a key analytical technique with successful applications in bioanalysis, disease detection, and diagnosis^[33]. In addition, biocompatible SERS-active nanoparticles or substrates can be coated with appropriate surfactants or protective shells to enhance their stability in biological environments. These nanoparticles or substrates can be used for direct detection or for indirect detection via specific receptors, aptamers, etc., thereby enabling efficient identification of bioanalytes. Whether for direct or indirect detection, SERS technology demonstrates high precision and exceptional sensitivity, providing a reliable means of detection for the early diagnosis and treatment evaluation of breast cancer metastasis.

SERS technology encompasses two methods: direct detection and indirect detection. Direct detection involves identifying target molecules by analyzing their intrinsic Raman fingerprint signals, without the need for any external signal labels (Figure 1B) ^[34-35]. This method is highly attractive, particularly for its outstanding performance in high-sensitivity detection at the single-molecule level, as it can directly obtain unique information about the target molecule while avoiding potential interference or errors that may be introduced by conventional methods^[36].

However, its practical application is limited by two key issues: First, most biomolecules themselves have small Raman scattering cross-sections, resulting in weak signals. Second, in biological systems, the surfaces of nanoprobes are prone to adsorption of serum proteins, forming a “protein corona.” This structure not only shields active sites but also significantly attenuates SERS signal intensity, reduces selectivity, and compromises in vivo circulation stability^[37-39]. In contrast, indirect SERS detection can, to some extent, overcome the limitations of direct detection. This approach typically involves modifying the surfaces of noble metal nanoparticles with various organic molecules as Raman reporter molecules, coating the outer layer with a protective shell, and attaching specific targeting ligands, thereby enabling indirect detection of proteins, DNA, and other biomarkers (Fig. 1C)^[40-41]. In indirect SERS detection, the targeting moiety is first attached to the nanomaterial, and then the SERS tag is bound to the molecular target, with the target being identified through the Raman signal of the reporter molecule^[42-43].

Compared with spontaneous Raman scattering, the significant enhancement of SERS signals at the nanoscale enables the detection of biomolecules at extremely low concentrations, even at the single-molecule level. By modifying the surface of SERS probes with targeting units—such as monoclonal antibodies, small-molecule ligands, or specific peptide chains—it is possible to achieve specific recognition of breast cancer–related molecules, including Human epidermal growth factor receptor 2 (HER2), Epidermal growth factor receptor (EGFR), and mucin 1 (MUC1)^[44].Targeting units bind with high affinity to receptors on the surface of tumor cells, effectively enhancing signal selectivity and thereby reducing background interference. This strategy not only improves the reliability of detection but also facilitates the differentiation of tumor cell subpopulations, providing data support for the analysis of tumor heterogeneity^[45].

SERS technology can be integrated with other clinical imaging modalities and molecular imaging techniques to enable multimodal imaging. At the molecular level, SERS provides high-resolution, fingerprint-like information, while other imaging techniques offer macroscopic structural visualization or anatomical localization capabilities. Multimodal fusion not only enhances imaging accuracy and spatial resolution but also enables multidimensional visualization of breast cancer lesions, such as detecting the precise location of tumors, the clarity of their boundaries, and potential metastatic foci. When used in conjunction with fluorescence imaging, MRI, and other techniques, SERS can simultaneously achieve both microscopic detection and macroscopic observation, providing a more reliable basis for precise diagnosis^[46]..

In real-time monitoring of breast cancer metastasis, SERS boasts high detection sensitivity and rapid response capabilities, enabling direct detection of trace CTCs and Exos without the need for complex sample pre-processing. Its spectral characteristics—narrow peak width and low background interference—allow it to precisely distinguish between structurally similar biomolecules, thereby overcoming the spectral overlap issues inherent in traditional fluorescence detection. Furthermore, compared with conventional methods such as hematoxylin-eosin staining (HE staining), immunohistochemistry (IHC), and polymerase chain reaction (PCR), SERS enables immediate spectral acquisition, significantly enhancing detection efficiency and making it particularly suitable for intraoperative rapid assessment and dynamic disease monitoring^[47]..

Breast cancer is a malignant tumor characterized by high molecular heterogeneity and a propensity for metastasis. Tumor cells from different patients, or even from different sites within the same patient, often exhibit significant differences in molecular expression^[48]. For example, the expression levels of markers such as HER2, the estrogen receptor (ER), and the progesterone receptor (PR) may be downregulated or absent due to heterogeneity among cell subpopulations or changes in the tumor microenvironment^[49]. This situation can easily lead to false-negative results in single-marker detection, thereby compromising the accuracy of disease diagnosis and treatment decisions^[50]. SERS probes, through the design of multi-target recognition platforms, can simultaneously carry multiple Raman reporter molecules and target different molecular markers^[51]. This multiplexed detection strategy can significantly enhance the ability to capture breast cancer–related biological signals, improve detection sensitivity and specificity, and effectively reduce the false-negative rate. For instance, simultaneous detection of HER2, EGFR, and MUC1 can cover a broader range of breast cancer subtypes, thereby compensating for lesion information that might be missed by a single marker^[52]. From a clinical perspective, identifying multiple molecular markers not only helps precisely locate small or occult metastatic lesions but also enables the assessment of tumor invasiveness and drug resistance, providing an important basis for developing personalized treatment plans. In particular, during preoperative subtyping, intraoperative navigation, and postoperative efficacy monitoring, multi-target detection can provide clinicians with more comprehensive and dynamic disease information, helping physicians achieve earlier and more accurate interventions^[53]. Moreover, the multi-marker strategy is also suitable for monitoring treatment-related biological changes, such as alterations in markers associated with immune cell infiltration in the tumor microenvironment and the epithelial-mesenchymal transition (EMT) process, thereby providing strong technical support for research on drug resistance mechanisms and the identification of new therapeutic targets^[54]. Therefore, SERS-based multi-target detection platforms not only represent an effective approach to addressing the molecular complexity of breast cancer but also offer a viable solution for advancing precision medicine.

To further enhance the efficiency and accuracy of breast cancer diagnosis, SERS platforms are progressively integrating with a variety of advanced technologies, driving their development toward intelligent and integrated systems. This cross-technology integration not only enhances the multidimensional capabilities of detection but also endows SERS with greater application flexibility, meeting the diverse needs of complex clinical scenarios. First, the introduction of microfluidic chip technology enables automation and high-throughput operation of sample processing^[55]. The miniaturized structure allows precise control of liquid flow, enabling sample preprocessing, mixing, and reaction to be completed entirely within microchannels. This significantly reduces human error and sample loss while markedly improving detection efficiency. Such integrated systems are particularly well suited for applications requiring rapid response, such as liquid biopsies and intraoperative rapid testing. Second, artificial intelligence (AI) algorithms, especially deep learning and pattern recognition techniques, are playing an increasingly important role in the signal analysis of SERS spectra^[56]. By training models to recognize Raman feature peaks associated with different biomarkers, AI can effectively improve the accuracy of signal identification, with particular advantages in samples featuring weak signals or complex backgrounds. The combination of immunoassay technology with SERS forms an immuno-SERS platform that enables highly sensitive detection of breast cancer–related proteins through antibody–antigen–specific recognition^[57]. Compared with traditional immunostaining, SERS provides faster and more quantitative results. At the genetic level, SERS probes can specifically target cancer–related mutation sites, such as PIK3CA (Phosphoinositide-3-kinase catalytic subunit alpha) and Breast cancer 1 (BRCA1)^[58]. By functionalizing nucleic acid probes and modifying them onto the SERS substrate, specificity is enhanced, and mutation information can be obtained directly without PCR amplification, providing a technological foundation for early molecular diagnosis^[59]. Finally, the application of smart nanoprobes expands SERS’s capabilities in metabolic detection. By loading pH–responsive units, enzymes, or targeting molecules onto the probe surface, these nanostructures can enable real–time monitoring of metabolites such as lactate, glucose, and glutathione, thereby supporting the assessment of tumor metabolic status and the dynamic tracking of treatment efficacy^[60]. Overall, the integration of SERS platforms with these cutting-edge technologies not only broadens their application scenarios across multiple stages of breast cancer, including screening, intraoperative navigation, treatment efficacy evaluation, and personalized therapy, but also enhances operational flexibility and system scalability. Clinically, physicians can flexibly select detection modes based on specific clinical needs, enabling the integration of multidimensional information from the molecular level to the systemic level and providing stronger support for precision medicine.

In summary, existing scoring systems have limited predictive capabilities for bleeding events, and their results are inconsistent^[25,30,33].

Despite significant improvements in SERS substrate performance, large-scale applications still face challenges such as complex fabrication, difficulty in controlling uniformity, and poor signal consistency. Although noble metal nanomaterials are highly sensitive, they are prone to oxidation or aggregation, requiring additional modifications to function effectively in complex environments. To address these issues, researchers have proposed several solutions: (1) biomimetic modification by coating the substrate surface with a silica or polymer layer to enhance stability and biocompatibility; (2) loading metal nanostructures onto polymer or fiber membranes to fabricate flexible substrates that can adapt to curved surfaces or flexible devices for on-site detection; (3) using photothermal or pH-sensitive materials to dynamically regulate “hotspot” distribution, enabling on-demand enhancement. In addition, 3D porous substrates can significantly increase detection throughput by increasing molecular adsorption sites.

SERS probes enable highly sensitive detection of biomolecules by immobilizing Raman reporter molecules on the surface of noble metal nanoparticles. Commonly used Raman reporter molecules typically exhibit the following structural characteristics. First, they should possess a high Raman scattering cross-section (such as rhodamine and methylene blue) to significantly enhance signal intensity. Second, they contain specific functional groups (such as 4-mercaptobenzoic acid, mercapto pyridine, and mercapto pyrimidine) that allow them to adsorb firmly onto the surface of noble metal nanoparticles. In addition, they feature a conjugated electron system (such as tetramethylrhodamine and crystal violet), which can increase the change in polarizability and enhance the Raman signal. Finally, these molecules exhibit good stability and are resistant to degradation (such as azo dyes and triphenylmethane dyes), thereby ensuring signal reliability. In SERS probe applications, these Raman reporter molecules are immobilized on the surface of noble metal nanoparticles and, in combination with targeting ligands on the protective shell, enable highly sensitive detection of biomolecules (such as proteins and DNA).

To ensure that SERS probes can efficiently identify breast cancer cells, the design must involve functionalization modifications targeting specific molecular markers. By modifying antibodies, peptide chains, or other targeting molecules onto the surface of nanoparticles, the probes can precisely recognize specific antigens on the surface of breast cancer cells (such as HER2, ER, etc.), thereby achieving highly efficient targeted recognition of tumor cells. To enhance the signal intensity of the probes, the probe design must select appropriate particle size and shape (such as nanospheres, nanorods, nanostars, etc.) as well as suitable metal materials (such as gold or silver) and optimize the LSPR effect of the metal nanoparticles, thereby amplifying the signal^[61]. In addition, a protective layer or functionalization modification on the probe surface can prevent oxidation or aggregation of the metal nanoparticles in the in vivo environment, thereby maintaining their stability ^[62-64]. In terms of specific modification methods, the surface of SERS probes is typically functionalized using biocompatible materials such as polyethylene glycol (PEG) and albumin to reduce the toxicity and immunogenicity of the metal nanoparticles^[65]. Polymer materials or water-soluble substances like PEG are often used for surface modification to enhance the biocompatibility of the probes and minimize their adverse effects on cells or tissues^[66-67]. Targeting molecules can also be attached to metal nanoparticles via electrostatic adsorption or chemical covalent bonding. Through these modifications, the probes not only improve their biocompatibility but also enhance their specificity and stability in recognizing breast cancer cells. The sensitivity of SERS probes is crucial for their application in breast cancer detection. To enhance sensitivity, nanomaterials with strong surface-enhancement effects are typically selected, and it is ensured that their surfaces have sufficient functional groups to strengthen interactions with target molecules. By optimizing the size, morphology, and distribution of the metal nanoparticles, the intensity of the Raman signal can be further increased, thereby improving the detection sensitivity of the probes.

Liquid biopsy, as a non-invasive diagnostic technique, analyzes tumor-derived substances such as CTCs, EXOs, and microRNAs (miRNA), offering advantages in early diagnosis and dynamic monitoring^[68-71]. Liquid biopsy can overcome the heterogeneity of breast cancer tumor cells, comprehensively reflecting the biochemical information of tumor tissue, and provides reliable support for the early diagnosis, treatment monitoring, and surgical evaluation of breast cancer^[72-74].

Epithelial-mesenchymal transition (EMT) is a key mechanism that promotes breast cancer metastasis. It enhances the metastatic potential of CTCs by facilitating their detachment from the primary tumor, enhancing their migratory and invasive capabilities, increasing their circulation survival rate, and promoting distant colonization^[75-77]. Histopathological biopsy is currently the gold standard for EMT monitoring; however, SERS technology, with its non-invasive nature and high sensitivity, shows potential as an alternative^[78-82]. Monitoring strategies based on CTC phenotypic characteristics and the dynamic process of EMT have been developed. During EMT, tumor cells gradually transition from an epithelial phenotype to a mesenchymal phenotype. This transformation enables tumor cells to detach from the primary tumor, invade the vascular system to form CTCs, and subsequently disseminate via the bloodstream to distant organs (Figure 3A). To monitor CTC phenotypic heterogeneity in real time, researchers have developed a multiplexed SERS nanoprobe labeling technique (Figure 3B). This technique employs four gold nanoprobes, each targeting a specific marker: the epithelial marker EpCAM (1338 cm^-1), E-cadherin (1080 cm^-1), N-cadherin (1379 cm^-1), and the tumor stem cell marker ABCB5 (1000 cm^-1)^[83-85]. By analyzing SERS spectra (Figure 3C(i)), we can simultaneously detect the phenotypic characteristics of different CTC subpopulations. Combined with a multiplexed labeling strategy (Figure 3C(ii)Figure 3C(iii)

Exosomes are a class of nanoscale vesicles secreted by cells, widely present in various body fluids. As an important mediator of intercellular communication, they can carry a variety of bioactive molecules, including proteins, lipids, DNA, and multiple types of RNA (including miRNA and lncRNA)^[86-87]. Recent studies have shown that tumor-derived exosomes play a crucial role in the initiation, progression, metastasis, and treatment response of breast cancer. They not only participate in regulating the tumor microenvironment, promoting angiogenesis, and facilitating immune evasion, but can also alter the biological behavior of target cells through interactions with them. Given that the molecular composition of exosomes is closely linked to the tumor state, they are regarded as a highly promising non-invasive biomarker that can be used for the early diagnosis, prognosis assessment, and treatment monitoring of breast cancer^[88-89]. As shown in Figure 3D(i, ii), the schematic process depicts the incubation of exosomes derived from three SK human breast cancer cell lines (SK breast cancer 3, SKBR3) and Michigan Cancer Foundation-7 cells (MCF) with HER2 and MUC1 aptamer probes, respectively. By using high-concentration, equimolar aptamer probes, saturation binding to the target molecules is achieved. The single-exosome SERS spectra are then analyzed using the multivariate curve resolution-alternating least squares (MCR-ALS) method, successfully deconvoluting two characteristic end-member spectra: the HER2-specific spectrum (α coefficient, major peak at 1580 cm⁻¹) and the MUC1-specific spectrum (β coefficient, characteristic peak at 1120 cm⁻¹)^[90-91]. The α/β ratio in each exosome can quantitatively reflect the relative expression levels of HER2 and MUC1, providing a multi-parameter analysis at the single-vesicle level for molecular subtyping of breast cancer (Figure 3D(iii)). The statistical distribution of the average intensity ratio (α/β) between SKBR-EXOs and MCF-EXOs reveals significant differences in their surface marker expression (Figure 3E). This result is highly consistent with the HER2/MUC1 expression ratio determined by Western blotting (WB), thereby indirectly validating the reliability of the unmixing model. SERS spectral reconstruction confirms that weighted HER2/MUC1 signals can accurately resolve EXO surface markers (Figure 3F). This model demonstrates a comprehensive technological workflow for SERS-based single-exosome analysis, enabling integrated analysis from exosome capture to molecular phenotyping.

MicroRNAs are a class of endogenous non-coding RNAs that play a crucial role in regulating gene expression, cell proliferation, invasion, and migration. Their aberrant expression is closely associated with the development and metastasis of breast cancer^[92-93]. Existing studies have shown that in the serum of ER-/PR-negative breast cancer patients, microRNA-21 (miR-21), microRNA-106a (miR-106a), and microRNA-155 (miR-155) are significantly upregulated, while microRNA-126 (miR-126), microRNA-199a (miR-199a), and microRNA-335 (miR-335) are significantly downregulated, suggesting that miRNAs hold potential as biomarkers for molecular subtyping and treatment evaluation in breast cancer^[94-95]. Current mainstream miRNA detection technologies include Northern blotting, reverse transcription quantitative polymerase chain reaction (RT-qPCR), next-generation sequencing (NGS), and microarrays; however, each of these techniques has its own limitations, such as insufficient sensitivity, lengthy assay times, or difficulties in labeling^[96-97]. Therefore, there is an urgent need for a novel detection method that offers high sensitivity, convenient operation, and high-throughput analysis. SERS technology, with its exceptionally high signal enhancement capability (enhancement factor up to 10¹²–10¹⁴) and strong resistance to background interference, is regarded as an ideal tool for achieving precise miRNA detection^[98-99]. However, simultaneously detecting multiple miRNAs in complex clinical samples such as tissues, serum, or plasma remains challenging, primarily due to: (1) the subtle differences in Raman signals among different miRNA molecules, making them difficult to distinguish; and (2) the time-consuming nature of point-by-point scanning^[100-102]. To overcome these limitations, researchers have proposed combining SERS with microfluidic chips and digital holographic imaging technology to enable rapid, multiplexed, high-throughput detection. A recent study has developed a three-dimensional (3D) SERS holographic chip platform that can complete the qualitative and quantitative analysis of breast cancer–related miRNAs within 9 minutes^[103]. The system first uses Exo III enzyme for isothermal amplification of miRNAs, then directs the amplified products into a microfluidic device embedded with a silver nanoparticle array. Each miRNA triggers a change in the SERS signal by opening a probe-specific stem-loop structure, causing the fluorescent dye to move away from the metal substrate. Finally, by scanning the array area and integrating spatial–spectral information, a 3D SERS hologram is generated that can simultaneously identify nine breast cancer–related miRNAs, including lethal-7b microRNA (let-7b), microRNA-1 (miR-1), microRNA-10b (miR-10b), microRNA-125b (miR-125b), miR-126, microRNA-133a (miR-133a), microRNA-143 (miR-143), miR-155, and miR-21. The platform achieves a minimum detection limit of 1 aM and exhibits approximately 85% detection consistency with RT-qPCR in clinical samples, demonstrating excellent sensitivity and accuracy.

With the advancement of liquid biopsy and the concept of non-invasive benchmark medicine, the application of SERS in breast cancer screening has attracted increasing attention. Traditional breast cancer screening methods, such as mammography, are recognized as standard screening approaches; however, they have significant limitations in detecting early, small lesions, which can easily lead to missed diagnoses^[104]. In addition, mammographic screening involves radiation exposure, making frequent testing inadvisable, and it lacks the ability to detect molecular-level abnormalities. In contrast, SERS technology offers advantages such as high sensitivity, high molecular specificity, and multi-target detection. By capturing molecular fingerprint characteristics in serum or plasma, SERS provides new technical support for the early detection of breast cancer^[105]. Furthermore, SERS-based detection does not require complex sample pre-processing or tissue sampling; only a small blood sample is needed to complete the test, making it suitable for large-scale population screening. Existing studies have attempted to combine SERS spectra with multivariate statistical analysis methods (such as principal component analysis PCA, linear discriminant analysis LDA, and support vector machines SVM) to establish discrimination models that successfully differentiate blood samples from healthy individuals and breast cancer patients^[106]. This technological approach offers advantages such as low cost, high throughput, and rapid response, providing a viable new tool for future primary healthcare facilities to conduct non-invasive breast cancer screening.

In addition to screening for individual cancer types, SERS can also be used for differential analysis between breast cancer and other cancers. In a large-scale study involving 253 total serum samples, researchers collected samples from healthy volunteers as well as patients with breast cancer, lung cancer, colorectal cancer, oral cancer, and ovarian cancer, and systematically evaluated the classification accuracy based on SERS spectra. The results showed that SERS technology demonstrates good discriminative ability in distinguishing between different cancer types, suggesting its potential application value in cancer type–specific screening and differential diagnosis^[107]..

Preoperative detection of lymph node metastasis (LNM) in breast cancer has always been a major challenge in clinical diagnosis and treatment. To track potential metastatic foci, various imaging techniques have been widely used in clinical practice, including US, CT, MRI, SPECT, and PET-CT^[108].Although PET-CT is widely regarded as the “gold standard” for imaging detection of various malignant tumors, it still faces challenges in terms of insufficient sensitivity and specificity for detecting small metastatic foci in breast cancer. US is easy to perform, non-invasive, and provides real-time dynamic imaging, and is commonly used to assess the size, shape, and internal structure of lymph nodes. However, its ability to detect deep or small metastatic lymph nodes is limited, and it is highly dependent on the operator. US also has relatively low specificity: benign lesions such as inflammation and infection can cause lymph node enlargement, leading to an increased false-positive rate and compromising diagnostic accuracy. CT offers high spatial resolution for assessing anatomical structures and can visualize lymph node size and calcification; however, its ability to resolve small lesions is limited, and it relies primarily on morphological features for assessment, lacking functional information and making it difficult to accurately distinguish between benign and malignant lesions. MRI provides excellent soft-tissue contrast and can clearly depict anatomical details in the breast and axillary regions. However, MRI imaging is time-consuming and relatively expensive, and it has certain limitations in detecting small metastatic foci, especially those with low signal intensity differences. As a functional imaging technique, SPECT can reflect the metabolic and blood flow characteristics of tumor tissue and is used for locating lymphatic drainage pathways or sentinel lymph nodes. However, its spatial resolution is relatively low, making it difficult to clearly visualize small structures, and its images are prone to artifacts, which limits its suitability for evaluating fine anatomical structures. PET-CT combines the advantages of metabolic and anatomical imaging and is widely used in breast cancer staging and systemic metastasis evaluation. Although it exhibits good detection capability for large, highly metabolically active lesions, it is prone to missing low-metabolism or small metastatic foci, particularly those with diameters less than 5 mm. In addition, PET-CT is expensive, requires sophisticated equipment, and is not suitable for routine screening^[109-111].Traditional imaging techniques generally suffer from insufficient sensitivity, limited specificity, and inadequate spatial or metabolic resolution when it comes to detecting small metastatic foci in breast cancer. Morphology-based diagnostic approaches struggle to meet the clinical need for early detection of small metastases, which has driven the gradual development of highly sensitive molecular imaging technologies.

SERS technology demonstrates unique advantages in the visual detection of breast cancer metastatic lesions, particularly in terms of sensitivity, molecular specificity, and spatial resolution. SERS imaging relies on the specific binding of functionalized nanoprobes to tumor-associated molecules or tumor microenvironment characteristics (such as pH and enzyme activity), with Raman-active molecules labeled on the surface of metal nanostructures. The "hotspot" effect significantly amplifies the signal, enabling the visualization and tracking of trace biomarkers. In the detection of breast cancer metastases, SERS nanoprobes can enter metastatic lymph nodes or distant metastatic sites either through active targeting or passive permeation and accumulation. With the aid of a Raman spectrometer, metastatic and non-metastatic tissues can be effectively distinguished. Compared with traditional imaging modalities, SERS enables early and precise localization and monitoring of small metastatic lesions, providing strong technical support for the early diagnosis, intraoperative navigation, and efficacy assessment of breast cancer metastatic lesions. With its extremely high detection sensitivity and molecular specificity, SERS technology holds promise for offering a new approach to the early diagnosis and precise monitoring of breast cancer metastasis^[112-113]. Triple-negative breast cancer, as an aggressive subtype of breast cancer, is characterized by early metastasis and poor prognosis, and traditional detection techniques often lack the sensitivity and molecular specificity required for early detection of metastatic lesions. Moreover, due to the absence of effective target sites, recognition units such as antibodies, peptides, and aptamers struggle to specifically target tumor cells and track tumor metastasis in real time. To address these limitations, a dual-modal imaging strategy based on SERS and bioluminescence has been developed, combined with bioorthogonal metabolic sugar labeling technology. Tumor cells (4T1-Luc) labeled with luciferase are modified with a bicyclo[6.1.0]nonyne (BCN) group, and click chemistry is used to facilitate the precise binding of the probe to 4T1-Luc cells. This dual-modal imaging approach enables highly sensitive, real-time monitoring of small metastatic lesions, providing a non-invasive and accurate method for assessing in vivo tumor metastasis and treatment response (Figure 4)^[114-115].

In summary, early diagnosis of breast cancer is crucial for improving patient survival rates. Histopathological biopsy is the “gold standard” for confirming a breast cancer diagnosis, providing information on the tumor’s histological type, molecular characteristics, and degree of invasiveness, thereby guiding personalized treatment plans^[116].Currently, clinical practice primarily relies on fine-needle aspiration cytology (FNA), core needle biopsy (CNB), and surgical excision biopsy. Due to their invasive nature, histopathological biopsy methods may cause infection, bleeding, or tissue damage, and may result in insufficient sampling or failure to capture key lesion areas, leading to a relatively high false-negative rate^[117].Furthermore, sample processing and pathological analysis typically require several days to weeks, which may delay treatment decisions^[118]. These challenges have prompted researchers to explore more efficient alternative diagnostic methods.

SERS technology, with its ultra-high sensitivity, rapid detection capability, and molecular fingerprint specificity, shows great potential in breast tissue biopsies. SERS probes can bind to specific biomarkers in tissues via minimally invasive or non-invasive methods, enabling real-time detection and reducing patient discomfort^[119]. SERS signals are highly sensitive to molecular changes, allowing even low-concentration biomarkers to be efficiently detected, thereby enhancing diagnostic accuracy^[120]. As shown in Figure 5, compared with traditional pathological methods, SERS detection does not require complex sample preparation. By detecting the molecular fingerprint characteristics of specific biomarkers and their metabolites in breast cancer tissues, SERS enables highly sensitive detection of changes in the tumor microenvironment, thereby predicting the potential malignant transformation capacity of breast cancer tissues^[121]. Studies have shown that SERS combined with functionalized nanoprobes can enable real-time detection of EMT-related proteins (such as E-cadherin and N-cadherin), revealing early signals of increased tumor invasiveness^[122]. By detecting these tissue-type changes at an early stage, SERS technology can provide critical information for MBC and clinical interventions, helping to improve treatment precision and patient prognosis.