SERS-guided photodynamic therapy: pioneering strategies in advanced cancer diagnosis and treatment

Abstract

Abstract

Cancer remains a leading cause of mortality worldwide, and while advances in conventional therapies such as chemotherapy, radiotherapy, and targeted therapies have improved patient outcomes, these treatments often fail to meet the demand for precision and specificity. Many current therapies struggle with limitations such as non-specific targeting, drug resistance, and significant side effects, often leading to incomplete tumor eradication and damage to surrounding healthy tissues. The urgent need for more precise, minimally invasive, and efficient cancer treatment strategies has paved the way for novel therapeutic approaches. Cancer theranostics has evolved exceptionally by incorporating new diagnostic tools that ultimately improve the therapeutic outcome of the treatment. Surface Enhanced Raman Spectroscopy (SERS)-guided photodynamic therapy (PDT) is rapidly gaining attention as a highly sophisticated modality in cancer theranostics, offering a dual advantage of enhanced diagnostic precision and effective therapeutic action. SERS, an ultra-sensitive molecular imaging technique, provides real-time, high-resolution detection of cancer biomarkers, enabling precise tumor localization and characterization. SERS-based theranostic probes can show potential results in both in-vivo and in-vitro studies. Herein this review critically discusses the key roles of SERS during the PDT treatment and also focuses on providing brief information on the fundamentals of both SERS and PDT. In combination with PDT, which selectively destroys cancer cells through photosensitizer activation under light exposure, SERS-guided PDT ensures a targeted therapeutic approach that minimizes damage to surrounding healthy tissues and reduces side effects.

Graphical abstract

Introduction

Cancer is a multifactorial and multifaceted chronic disease that is considered a global issue. According to the World Health Organization (WHO), cancer is one of the leading causes of death and it is estimated that by 2040 the annual cases could surge up to 28.9 million [1]. Treating cancer at an early stage is the key to reducing mortality and morbidity rates and minimizing treatment expenses, also the transition of cancer cells to metastatic tumors can be avoided with early diagnosis and screening [2, 3]. Currently, there is an arsenal of therapeutic options for the doctor to choose from depending on the stage and type of cancer (Fig. 1). However, conventional treatment options like surgery, chemotherapy, and radiation therapy are widely used though they have many drawbacks. For example, surgical resection is considered an effective method for early-stage cancer treatment, but it gets expensive and difficult when dealing with complex cases [4]. Similarly, chemotherapy and radiation therapy have been used to stop or slow the proliferation of tumor cells with the help of anti-cancer drugs and X-rays, however, most chemotherapies are not selective towards cancer cells and can affect the nearby healthy cells. Chemotherapy also causes major drawbacks like drug resistance and severe side effects (hair loss, nausea, etc.) [5]. On the other hand, radiation therapy suffers from high cost, patient immobilization and complex procedures [6].

Fig. 1.

Overview of various treatment modalities available for cancer theranostics. The figure depicts the comprehensive and multifaceted approach to cancer diagnosis and treatment, highlighting advancements and the importance of personalized medicine in improving patient outcomes

Accounting for the limitations of conventional therapies, huge efforts have been implemented to develop state-of-the-art strategies which overcome complex treatments and severe side effects. Recently, nanomaterials have been used to diagnose and treat cancer due to their tunability and ability to penetrate the tumor by enhanced permeation and retention effect (EPR) [5]. Also, nanomaterials have intrinsic properties which are utilized in advanced treatment methods like drug delivery, hyperthermia and PDT [7]. Additionally, functionalizing the surface of the nanomaterials with target-specific antibodies can improve treatment efficiency by improving its specificity and precise dosimetry [8]. Compared to antibody–drug conjugates, known for their high specificity and improved excretion pathways, antibody–drug nanoparticles offer additional perks like improved drug release, stability and higher therapeutic efficacy. Moreover, nanomaterials like plasmonic nanoparticles have intrinsic properties which aid in phototherapies and functionalizing it with antibody–drug conjugates can act as a multifunctional probe with greater treatment efficacy [9–11]

In phototherapies PDT has gained worldwide recognition due to its non-invasive and cost-effective procedure, and lesser post-treatment care time [12]. Additionally, it is a clinically approved phototherapeutic method that requires essential components like photosensitizers (PS), oxygen and light. The PS can generate reactive oxygen species (ROS) upon light irradiation which can disrupt the cellular equilibrium and trigger the cell death mechanism of the neoplastic cells, commercially available PS like photofrin, levulan and foscan are still used for PDT treatment [13].

PDT can be combined with other techniques for greater efficacy, therapeutic modalities like immunotherapy, chemotherapy, radiation therapy and surgery have improved results when combined with PDT [14]. While limitations are attributed to the PS used, light sources and tumor location, more research has delved into developing multifunctional theranostic probes capable of performing multiplex or multimodal imaging with targeted PDT [15, 16].

In the courtroom of cancer theranostics, imaging and monitoring techniques are the primary evidence or visual testimony for rendering a verdict, thus the results from these techniques must be reliable and precise. Early diagnosis and specific targeting hinges on the recognition of over-expressive biomolecules which serve as a biomarker for disease stratification [17]. In the case of PDT, the imaging and monitoring techniques can aid in dosimetry, tumor cell monitoring and predicting the cell death mechanism [18]. Traditional techniques like fluorescence are being eclipsed due to their low specificity and lack of producing three-dimensional tumor images [19].

Currently, non-invasive or minimally invasive imaging methods are encouraged for cancer diagnosis and early detection. SERS has proved to be an excellent imaging, sensing and monitoring technique in the biomedical field due to its unique molecular fingerprinting, high sensitivity and multiplexing capabilities [20]. Additionally, SERS offers high spatial resolution, contrast agent free tumor monitoring and single molecule detection. Although SERS shows high potential compared to other alternative imaging methods it still struggles with complications in outdoor sample analysis [21], hence developing proper experimental conditions and protocols can solve the issue.

In recent years, SERS has been used as a monitoring and imaging tool for phototherapies and there are numerous studies on SERS-guided photothermal therapy (PTT) and PDT, but as far as we know, there have been very few literature reviews specifically addressing SERS-guided PDT. So in this review, we start by providing an overview of PDT and SERS, proceed to assess the possibilities of using SERS during PDT, and finally provide future perspectives to improve the current limitations.

Photodynamic therapy for cancer treatment

Light as a tool for treating diseases dates back over 4000 years when ancient Indian, Egyptian and Chinese civilizations used it to treat various diseases like skin cancer, psoriasis and other skin diseases [22]. The beneficial effects of phototherapies were later shown by Niels Finsen for inhibiting growth and stopping the spread of smallpox pustules using red-light illumination. In another work, he also utilized ultraviolet light to treat cutaneous tuberculosis. These contributions led to his recognition and subsequent award of the Nobel Prize in 1903 [23]. These events led to researchers exploring different mechanisms for treating diseases with the usage of light. Phototherapy branched into different types and each had its challenges and advantages. In 1960, two scientists, R.L. Lipson and S. Schwartz at Mayo Clinic, reported the treatment of tumor using hematoporphyrin derivative and light, this approach of using light and PS was called PDT [24]. However, the concept of PDT was shown by Herman Von Tappeiner in 1909, and he called the phenomenon “photodynamic action” [25, 26]. In 1976, PDT showed a good response in human trials for treating bladder cancer and later, with constant research and development, it showed promising results with other tumor types, such as oral cancer, breast cancer and gynecological cancers [27]. Unlike other treatment methods, PDT induces cell death by three inter-combined mechanisms and is highly effective in preventing tumor recurrence (explained in upcoming sections) [28]. In the late nineteenth century, Naomi J. Halas and her group established the use of plasmonic nanoparticles for the treatment of malignant cells. Plasmonic nanoparticles absorb near-infrared (NIR) light, and due to the molecular vibrations heat is produced, this light-induced heat for treatment is called PTT [29]. There are several more phototherapies like photoimmunotherapy and laser immunotherapy, which are the subcategories of targeted phototherapies. They utilize antibodies or antigens to target the tumor and deliver PS [30]. Using phototherapies for the treatment of cancer accounts for highly controllable procedures, minimally invasive and accurate targeting, and blooming as potential methods for study [31]. Several studies have shown the transition of PDT to clinical use, while very little in the case of PTT. This is because the significant difference between PDT and PTT is the light source. PDT requires low-power lasers, and in some cases, natural sunlight is enough to excite the PS, but PTT requires an expensive setup. PDT is more selective than PTT, and the lack of heat monitoring during PTT can result in adverse effects [32].

PDT mechanism on tumor

In photoreactions, the key stages are light absorption and energy transfer (Fig. 2a). In a typical photodynamic reaction (PDR), the electrons of the PS moiety absorb photons and excite to the singlet state (¹PS) for a short time. Then electrons can decay back to the ground state emitting fluorescence, a valuable asset for imaging and pharmacokinetics [33]; or electrons can move to triplet state (³PS) undergoing intersystem crossing, doing so the spin of the electrons changes making it longer lived than singlet stage species. Additionally, the electrons in triplet state can follow different pathways (Type 1 and Type 2). Type 1 reaction involves the transfer of electrons to substrate to produce cytotoxic radicals and species (H₂O₂, HO^•) to impact severe oxidative damage to the biomolecules. Type 2 reactions involve the transfer of electrons to ground-state triplet oxygen (³O₂) to produce highly reactive singlet oxygen (¹O₂) for cell destruction [13, 33, 34].

Fig. 2.

Schematic illustration of PDT and its mechanisms of tumor destruction. The diagram depicts the process of PDT, where a PS is administered and preferentially accumulates in tumor cells. a Upon irradiation with a specific wavelength of light, the PS is activated, producing ROS. b The ROS induce tumor cell death through three primary mechanisms: (1) Direct cytotoxic effects on tumor cells by damaging cellular components, (2) Destruction of tumor vasculature, leading to nutrient deprivation, and (3) Activation of the immune response, which targets and eliminates tumor cells. The combination of these mechanisms contributes to the effective reduction of tumor size and inhibition of tumor growth

The pathways are simultaneous and highly dependent on the PS used; hence the careful choice of PS is pivotal for effective treatment [35]. Since the radical and singlet oxygen is short-lived (40 ns) the cells at the radius of 20 nm to the cytotoxic species are only affected [34]. While it might feel like these short-lived radicals have less impact on our body, the following tumor destruction mechanism would show their true potential as a therapeutic moiety. Generally, there are three main mechanisms for tumor treatment (Fig. 2b). First is the direct contact of ROS with the tumor cells, the second is the vascular damage, and the third is triggering the immune system to destroy the malignant cells. In most cases, all the mechanism coincides and their ratios are dependent on the PS molecule [35–37].

Direct tumor killing

Cytotoxic radicals can cause direct cellular damage by apoptotic (programmed) pathways and necrosis (non-programmed) pathways. Generally, a high dosage of light can enhance necrosis ablation [34], where sub-cellular and cellular membranes are damaged with the release of cytokines and other lethal chemicals by the cell components like mitochondria, these released chemicals spread to the nearby cells causing bystander effects. On the other hand,  with low doses of light, apoptotic pathways [34] can be initiated causing well-ordered and programmed cell death. Apoptosis has no bystander effect and it also has better selectivity [38]. For instance, the work by Plaetzer et al. studied the effect of light doses on cell death type; they concluded that a light dose of 2.5 to 3.5 J cm − 2 showed signs of apoptotic cell death but a higher dose of light > 6 J cm⁻², showed only necrosis [39]. Still, the efficacy of the PDT treatment is highly influenced by other parameters like the amount of oxygen available near or within the targeted tissue, the rate of PS photobleaching and finally the tumor site [40]. Also, the PS moiety absorbs light of a particular wavelength, mostly 650 nm to 800 nm [41]. This means to perform PDT; the light of a specific wavelength must travel through the tissues and then activate the localized PS moiety for the tumor ablation. Hence, one needs to account for the distance of the applied light, duration of light illumination and intensity of the light source because all these factors affect the cure rate [32]. Only relying only on the direct tumor destruction mechanism is not sufficient for total tumor ablation. Treatment effects can be improved by implementing two techniques: one is the re-oxygenation of the targeted tissues so that oxygen needed for the PDT can be utilized, and the other is lowering the rate of light irradiance while keeping the same fluence delivery, which also reduces the oxygen consumption during the PDT process [42, 43].

Vascular damage

Cutting off the supply of nutrients to the tumor cells is a promising approach to stop the proliferation, therefore targeting the blood vessels connected to the tumor cells, localizing the PS to damage the vessels (vascular effect) and blocking the oxygen supply of the tumor cell is the pathway for long term tumor control. The choice of PS paves the way for different destructive mechanisms like thrombus formation, vascular leakage and blood flow stasis during PDT [33, 40]. For example, if PS like photofrin are used in PDT, it can cause vessel constriction and leakage, thrombus formation, and, in some cases, leukocyte adhesion. But if PS like phthalocyanine and its derivates is used, it can primarily cause macromolecular vessel leakage only [44]. During the PDT both apoptotic and necrosis pathways take place either separately or simultaneously leading to tumor hypoxia and malignant cell death. Hence vascular mechanisms are synergistic with other PDT mechanisms to prevent chances of tumour growth and spread. The circulation of PS moiety through the vessels during the treatment can be due to its solubility, and higher circulation can cause blood flow stasis and acute microvascular damage. Also, higher uptake of PS can result in damaging the endothelial cells [36]. Supportive studies have shown endothelial cells being more sensitive towards PDT when photofrin moiety was used [45]. Interestingly, when the PDT is performed and vascular damages are observed, other mechanisms are also triggered. Leukocytes get adhered to the vessels and trigger immune and platelet activation, ultimately improving the interstitial pressure and thrombus formation. Triggering the leukocytes can prove that all the PDT mechanisms work in synergy. However, the efficacy of the treatment is still limited by light doses, the type of PS moiety, and its solubility [36].

Immune response

Compared to the other two mechanisms the immune response is not spontaneous but it can lead to long-term benefits with prolonged post-treatment effects on the whole body. PDT can either trigger or suppress the immune system. For instance, during PDT treatment, it produces inflammatory molecules to signal the immune system. Neutrophils, which are abundant in the body are the first cells to respond when methylene blue PS is used. Even though they get adhered to the treated site their myeloperoxidase activity doesn’t get altered but their ROS production increases [37, 46]. In the late 1980s, researchers reported the infiltration of lymphocytes and leukocytes in the PDT-treated tissues to enhance the immune response, ultimately preventing the recurrence of the tumor [13, 47]. The inflammatory responses at the treated site are mediated by factors such as the release of cytokines, acute phase proteins, growth factors, ROS and other immunoregulators. These regulators are responsible for signalling tumor location for a regulated infiltration of macrophages and neutrophils at the tumor site. Upon arrival, the macrophages and neutrophils phagocytize photodamaged cancer cells and mediate high immunity responses even after the PDT treatment [48–50]. For example, PDT also increases the expression of other molecules like E-selectin and ICAM1, which aids in adhesion to the blood vessel walls and helps the immune cells to infiltrate the tumor. PDT can also stimulate the T-cells (CD4 + and CD8 +) which are another important class of immune cells that are responsible for killing the tumor cells. In short, post-PDT triggers immune cells like neutrophils, macrophages, NK cells, and T-cells for a coordinated immune attack on the tumor site. Also, this immune reaction is not highly specific to the PDT site, it can also occur in local and distant lymphatic tissues ensuring long-term cancer control and complete treatment [37, 46]. Hence, the key point is that for an improved treatment outcome and long-term tumor control the combination of all three mechanisms: direct PDT effects on the tumor cells and its blood vessels with upregulation of the immune system should take place [33, 40, 48].

Role of photosensitizer

The PS is the main pillar of support to determine the therapeutic outcome of the PDR. Hence, PS with great photophysical, photochemical and better pharmacokinetics are ideal for improving the efficiency of photoreactions. Parameters such as tumor uptake kinetics, toxicity under dark conditions, side effects to nearby healthy cells and amphilicity nature of the moiety are crucial when choosing the PS that will be used during in vivo and in vitro studies [41]. It is understood now that PS absorb light to produce ROS but to improve the production of ROS, the intersystem crossing from the singlet state should be more often and the lifetime of the electron in the triplet state must be longer so that molecular oxygen and other substrates can react frequently. Another challenge is to use PS that absorbs light in the NIR region for two reasons; (1) NIR light has higher tissue penetration. (2) Light wavelengths below 700 nm are absorbed by endogenous molecules [51]. Keeping all the requirements in mind, research toward developing an ideal PS is still ongoing. While numerous PS molecules were reported and some clinically approved, a new generation of PS molecules continues to emerge. As of now, there are three generations of PS and each has its pros and cons (Table 1).

Table 1.

Pros and cons of different generations of photosensitizers

Type of photosensitizers	Pros	Cons
First generation PS	• Simple synthesis • Effective tumor destruction • Negligible dark toxicity • Intravenous dosage formulation Example: Photofrin	• High dosage requirement • Cannot absorb higher wavelength light • Prolong accumulation on the cells • Off-target effects • High photobleaching
Second generation PS	• Longer wavelength absorption • Better selectivity • Less photobleaching • Lower retention on the tissues • Negligible dark toxicity Example: Temoporfin	• Complex synthesis • Limited clinical validation • Limited availability • Limited water solubility
Third generation PS	• Superior selectivity • Improved pharmacokinetics • Longer wavelength absorption • Multimodal functionality • Prevents unwanted side reactions • Personalized and tailored treatment Example: Ce6 + tumor-targeting nanoparticles	• Limited clinical validation • Toxicity due to the carrier used • Complex synthesis • Cost of production • Not well established hence lower availability

First-generation PS are generally tetrapyrrole compounds like porphyrin and hematoporphyrin derivatives (HPD). They are known for their high absorbance of visible light and water solubility. Clinically they are used to treat the brain, lungs, skin, colorectal and other carcinomas. Photofrin (a purified form of HPD) was the first PS to be clinically approved as a photoactive moiety for treating cancer [52]. But photofrin and other first-generation PS have distinct drawbacks like longer skin photosensitivity, low selectivity and photobleaching. Still, they are widely used in clinical treatment [34]. However, these drawbacks enhanced the development of second-generation PS.

Compared to the first generation there has been a lot of improvement in the structure and composition of the second-generation PS. Porphyrinoid compounds and non-porphyrinoid compounds like benzoporphyrins, anthraquinones, chlorins, phenothiazinium dyes, psoraleans, phthalocyanines, protoporphyrin IX (PpIX) and other macrocyclic structures are collectively labelled as second-generation PS. They have advantages such as higher selectivity, lower retention on the skin and less photobleaching than their predecessor compounds [53]. Additionally, the light absorbance shows a minimal shift towards NIR in the case of second-generation PS, for example, modification of porphyrins to produce chlorins shows higher wavelength absorption in the therapeutic region [34].

Some commercially available PS are hydrophilic, while others are hydrophobic. Compounds that contain a porphyrin ring backbone such as chlorin, temoporfin, PpIX, etc. are hydrophobic. PS such as aminolevulinic acid (ALA), and mono-L-aspartyl chlorin e6 (MACE), are highly hydrophilic. However, it’s known that hydrophobic nature allows the PS to penetrate the cell membrane and locate sub-cellular components. High hydrophobicity can cause complications in biodistribution and photophysical properties ultimately lowering its efficacy. Hence the necessary substitution on the frame structure can make the PS amphiphilic to solve the imposed drawbacks [40, 54, 55]. Additionally, the substituents can improve the targeting ability of the PS. For example, anionic substituents like the carboxyl group can improve the targeting of PS towards the cytoplasm, whereas cationic substituents, help in targeting the mitochondrial membrane [34, 51, 56].

To obtain highly selective, reduced adverse effects and improved pharmacokinetics, researchers developed the third-generation PS [57]. Which is an advanced version of the second-generation PS, where it is conjugated with target-specific antigens, antibodies or other targeting moieties. Another type is to use nano or porous materials to encapsulate the PS species. Various nanocarriers such as plasmonic, metal oxides, polymer-based, carbohydrates and many more are used for the incorporation of PS moiety, and each has unique advantages towards targeting and stimuli-responsive PS release [58–61]. Compared to other generations the third-generation PS tend to show better biodistribution and easily bind to biomolecules so that uptake and improved PDR are observed. However nano carriers tend to undergo side deactivation pathways due to variable parameters like temperature and pH, but these issues can be solved with careful and well-designed pharmaceutical formulations to protect the PS from external factors [62, 63].

Limitations and possible improvements of PDT

For a PDR to orchestrate, we need essential components: light, PS and oxygen. However, it is intriguing to observe that the essential components that are responsible for its success can also limit its efficacy. Hence, careful choice of a light source, tumor environment and adequate PS can be the right move towards obtaining potential therapeutic results. Light used during PDR plays an important role in exciting the PS moiety in a light-induced reaction. So, the choice of light source is dependent on the PS used. Generally, the penetration ability of light depends on its wavelength and tissue components. It is to be noted that sub-cellular components and other metalloproteins in the body tend to absorb, scatter, reflect or transmit light [64–66]. Therefore, the light at the wavelength of the so-called phototherapeutic window (600 and 1,300 nm) is used in clinical stages, but above 850 nm the PS cannot excite due to low energy transfer [67]. So, the optimal wavelength range which shows good penetration ability, and ability to excite the PS molecule is 620–850 nm. The choice of a light source depends not only on the absorbance of PS but also on the tumor location, size, and type. Additionally, PDT can be performed with coherent light sources or non-coherent light sources according to the need [68, 69].

Oxygen is a vital component of PDT and the efficacy is also determined by the concentration of oxygen available in the tumour tissues. Hypoxia regions of the tumor where deficiency of oxygen is observed can cause complications during the PDT pathway. Hence combination with hyperbaric oxygen (HBO₂) therapy the hyperoxygenation on the tumor occurs and produces oxygen to aid the PDT [70]. Another method is to tamper with the light source to fix an optimal intensity and fractionate it so that the replenishment of oxygen in the hypoxic region takes place during the treatment process itself [66].

Apart from absorbing light and reacting with oxygen to produce cytotoxic radicals, PS can be used for imaging, targeting and monitoring the tumor [71]. For an effective treatment, factors like purity, dosimetry and pharmacokinetics are important [40, 57]. As discussed previously, developing an ideal PS is highly challenging and satisfying all the ideal conditions is nearly impossible. However, the current first, second and third-generation PS provide a vast option to choose from for specific tumor conditions.

Other than conventional methods to improve the existing limitations, several more novel strategies can be implemented to enhance the PDR outcome. Specific targeting of mitochondria [72], lysosome [73], cell membrane [74], nucleus [75] and other sub-cellular/organelle components can improve the selectivity of the PS for an improved treatment. The advanced strategy of separating the PS activation and singlet oxygen release also known as two-stage photodynamic therapy can be used to address the problems in tumour apoptosis in the hypoxia region [76]. Recent works on multimodal imaging/ monitoring of tumor and utilization of synergistic therapy can further improve the effectiveness of the treatment and prevent tumor recurrence [77].

Role of monitoring and imaging in PDT

The cornerstone of improving the accuracy of PDT is the ability to predict early treatment planning and post and pre-treatment monitoring. As discussed previously imaging tools help in the laboratory as well as clinical research. They aid in predicting the PDT mechanism, PS localization, organ/vascular interactions, cell death monitoring, photobleaching and many more. Generally, the assessment of structural, functional and molecular imaging information regarding the tumor, tumor microenvironment (TME) and biomarkers is great for pre-treatment planning, monitoring and post-treatment follow-ups (Fig. 3). The assessment of the post-treatment outcomes such as blood vessel occlusion and cell death types (apoptosis, necrosis, autophagy) can be observed by comparing the before and after images of the treated sites [18]. Techniques like magnetic resonance imaging (MRI), computed tomography, and fluorescence imaging are often used for imaging the pre- and post-PDT stages at in vitro and in-vivo applications. MRI imaging has excellent spatial resolution, non-invasive with high tissue penetration and in some cases MR contrast agent enhancements represent necrosis. However, MRI suffers from high operating costs, longer protocols and poor sensitivity [78]. Similarly, other techniques like computed tomography and positron emission tomography (PET) can image both necrosis and tumor volume with high resolutions in different locations of the body [79]. The majority of PS compounds have fluorescence properties, meaning they have the inherent ability to serve as an imaging agent for distinguishing normal and tumor cells. However, the lack of imaging tumor in the deep sites is still an issue in all fluorescence-based imaging strategies, it is because the light source used for excitation is generally blue light, which has lower penetration as well as scatters through tissues while propagation. So, a fluorescence-based approach is good for sites on the surface level and not in deep tissue regions [18]. Hence, the use of other imaging techniques along with the inherent fluorescence of PS is appreciated and used to produce reliable and selective observations for guided PDT. This multimodal approach has not been fully explored in terms of synergy between different imaging tools and adaptation of newly emerging imaging techniques. Apart from imaging, the monitoring approach is also important in PDT for obtaining feedback on the treatment responses, and it is to be noted that the efficacy of monitoring circles around the dosimetry or the dose–response of the treatment. During the treatment course, the change in the treated tissues can account for highly valuable information on the outcomes. Depending on the requirements several monitoring strategies like analyzing blood flow dynamics, time-gated fluorescence imaging, PET imaging and other parameters like pH, oxygen and dosage of PS monitoring are used. For example, one of the methods for monitoring oxygen is done by quantifying the diffusion of oxygen levels in the tissues and nearby regions, where the depletion of the oxygen levels is directly proportional to the concentration of the PS and the light irradiance used [80]. Thus, the incorporation of imaging tools for targeted and real-time monitoring of PDT enhances the treatment outcome and also improves patient care in terms of survival. Apart from the conventional imaging tools, SERS is also used for guiding PDT due to its rapid and high selectivity monitoring and imaging approaches (discussed in Sect. 4). Additionally, PDT probes are non-invasive or minimally invasive without the need to cut open the tissue to perform any treatment and most of the PDT probes are injected intravenously or applied on the surface. PDT probes combined with SERS diagnosis can provide non-invasive treatment and diagnosis [81, 82].

Fig. 3.

Schematic matrix diagram depicting the roles of different imaging types and their functions in various steps of treatment. The matrix highlights how various imaging modalities, such as MRI, computed tomography, PET, and ultrasound, contribute to distinct phases of the treatment process, including diagnosis, treatment planning, treatment delivery, and follow-up. Each cell in the matrix illustrates specific applications of imaging modalities. This comprehensive matrix underscores the importance of multimodal imaging in enhancing the precision and effectiveness of cancer treatment [18]

Surface-enhanced Raman spectroscopy for cancer diagnosis

Light scattering can be broadly classified into two types elastic and inelastic (Fig. 4a), Rayleigh scattering is an example of elastic scattering where the energy of the incident photon and the scattered photons are the same, such data doesn’t account for significant use. However, Raman scattering is a form of inelastic scattering which undergoes frequency shift due to a change in the energy of the incident photon upon the interaction with the vibrational modes of the molecules. From Fig. 4a, it shows scattered light lower than the incident light E < E₀ corresponds to Stokes scattering and the scattered light greater than incident E > E₀ is the anti-Stokes scattering. Generally, more Stokes scattering is observed because most of the molecules are in the ground state hence, they absorb light to go into a higher vibrational state and the scattered photons have less energy. Comparatively, anti-stokes scattering is less observed because fewer molecules are only in an excited state at ambient conditions [83]. Raman scattering provides valuable information regarding the molecular structure and composition of biological and chemical species. The major issue for Raman spectroscopy is the order of Raman scatterings which is associated with low sensitivity, time-consuming and poor reproducibility [84]. Fortunately, In 1970 SERS was introduced addressing the issues of sensitivity in Raman spectroscopy and in 1974 enhanced Raman signal of pyridine was first reported by Martin Fleischmann [85]. Intensive research following up on the underlying Raman scattering mechanism was studied and later after three years two mechanisms (i) electromagnetic enhancement theory [86] (ii) charge transfer enhancement theory [87] were accepted by the scientific community. Since then, SERS has been widely used in sensing, imaging and monitoring applications because of its precision, reproducibility, non-invasive and resistance to photodegradation or photobleaching. Huge amount of research has been done on the biomedical applications involving SERS, and numerous reviews have been published in the last five years. For example, Wang et al. [88] reviewed the principle and suitable factors in improving the performance of SERS-based immunoassays for different biological applications. Liu et al. [89] reviewed the biomedical application of SERS tags and provided insights on in-vivo tumor imaging. The bridge between liquid biopsy and SERS microfluidics for the detection of biomarkers has been reviewed by Shanmugasundaram et al. [90]. Diego et al. [91] have reviewed the importance of chemometrics and machine learning in data analysis and also provided insight into the combination with SERS for real-life applications. Feng et al. [92] unraveled to potential of SERS biosensors for the detection of Alzheimer’s disease biomarkers and the path towards clinical studies.

Fig. 4.

Illustration of different mechanisms of light scattering and enhancement in SERS. The diagram demonstrates the following key concepts: a Light Scattering: depicts Rayleigh scattering, where light is scattered without a change in wavelength, and Raman scattering, where light is scattered with a shift in wavelength due to interactions with molecular vibrations. b Electromagnetic enhancement: Shows how localized surface plasmon resonances (LSPRs) on metallic nanostructures amplify the electromagnetic field near the surface, significantly enhancing the Raman signal of molecules adsorbed on or near the nanostructures. c Chemical enhancement: Illustrates the chemical interaction between the analyte molecules and the metal surface, leading to charge transfer processes that further enhance the Raman signal. The combined effects of electromagnetic and chemical enhancements result in the significant amplification of the Raman signal, making SERS a powerful technique for molecular detection and analysis

SERS mechanisms

The first accepted SERS enhancement mechanism is electromagnetic enhancement (EM), which is the major contributor to the Raman signal enhancement factor (10¹⁰–10¹¹ times). In 1985 Martin Moskovits proposed that the enhancement of the Raman signals was due to the collective oscillation of electrons on the metal surface, which in his case was a rough metal surface. Still, he also predicted that the enhancement could be observed even in the metal colloidal solutions [93]. EM can be applied to any analyte but the substrate used plays an important role in enhancing the Raman signals, particularly plasmonic nanostructures via localized surface plasmon resonance (LSPR) (Fig. 4b). When light incidents on a metal nanoparticle or plasmon, the free electrons in the metal surfaces start to oscillate with the frequency of incoming light, at the resonance frequency, the electrons are confined and start to absorb and scatter more. Generally, plasmonic nanoparticle due to their LSPR property is widely accepted in biosensing and diagnostic applications which require high precision and large enhancement, for example, cancer diagnosis from liquid biopsy samples. The electromagnetic mechanism is considered a two-step process coupled with each other, first, it is the excitation of induced dipole and the second is the emission of Raman dipole [94]. It is also to be noted that each nanoparticle feels the effect of the external field plus the polarizing effect of the charges induced in the nearby nanoparticle, so when the dipole moment is perpendicular to the dimer axis, no field enhancement is observed. However, when it is parallel to the axis of dimer then a huge enhancement in the field of nanogaps is observed (hotspots). Analytes closer to the hotspots have better signals than those far away from it which is favourable in the application of single molecule detection, however, while considering other applications such as biosensing and theranostics, researchers work with the average enhancement factor because of reproducibility [95]. Average enhancement is the contributions of both molecules near the hotspot and the many other less-enhancing sites which accounts for 10⁷ or 10⁸ enhancement [96].

At the same time, another SERS enhancement mechanism was also accepted which is the chemical enhancement (CM). The CM is viewed as the change in the Raman polarizability tensor of an analyte upon adsorption onto the metal surface due to the charge transfer (CT) mechanism [97]. Additionally, CT does not contribute to a high enhancement factor; only about 10²–10³ enhancement is observed. Unlike the EM, the CT mechanism depends on the substrate, analyte and laser energy used [98, 99].

(Fig. 4c) illustrates the possible mechanism for CM, Case (1) when metal acts as a perturbation to the electronic structure of the analyte because the analyte couldn’t bind covalently to the metal surface thereby causing a slight change in its electronic distribution ultimately changing in Raman efficiency. Case (2) is the presence of metal complex either by direct or indirect binding, which prominently changes the polarizability of the analyte, herein the polarizability depends on the optical transitions of the molecule, therefore transitions due to overlapping of molecular orbitals provide an enhancement. Case (3) is a photo-driven charge transfer mechanism uncovered by electrochemical cell experiments, where applying an external potential to change the energy between the analyte and metal made enhancement evident at different potentials for different laser intensities. In other words, the difference between the fermi levels (E_F) of the metal surface and the highly occupied molecular orbital (HOMO) or lowest occupied molecular orbital (LUMO) is dependent on the incident laser [97]. It is also believed that according to the chemical enhancement theory, CT contributes to the selective enhancement of Raman bands in contrast to EM. It is to be noted that though both EM and CT mechanisms are not working in the same principle their combination could result in an excellent signal-to-noise ratio and higher reproducibility [100].

SERS substrates and detection strategies

We have discussed previously that substrate plays an important role in the enhancement factor of Raman signals, hence designing an optimal substrate with high reproducibility and reliability for a specific application is always crucial. Over a few decades, researchers have developed various plasmonic substrates and non-plasmonic materials such as carbon-based materials, metal oxides, semiconductors and polymers for sensing and disease diagnosis [101–103]. The first-ever substrate used to observe the SERS signal was a rough surface silver electrode [85], later, it was proved that the silver colloidal solution could also enhance the Raman signals [94]. Following these observations, research on utilizing the LSPR properties of plasmonic nanostructures has been conducted to tune the properties, such as size and shape for better results. Moreover, due to their unique properties, 1D, 2D and 3D metallic substrates have gained a lot of interest.

Colloidal substrates are frequently used in SERS applications due to their low cost and simple fabrication techniques. Additionally, colloidal SERS substrates account for simpler sample manipulation and faster analysis than solid substrates [104]. However, colloidal substrates have critical limitations such as signal instabilities, less control over hotspot distribution, and poor reproducibility [105, 106]. 1D SERS substrate combines LSPR and long-range laser interaction to give a huge enhancement, which is useful for molecular detection. 1D substrates have controlled and hotspot uniformity, which provides extraordinary signal enhancement. Still, they have limitations in terms of synthesizing the materials and having highly ordered arrangements [106]. In the 2D substrates, orderly arrangements yield distinctive advantages such as maximizing the surface density of hotspots and highly enhanced signals due to reduced damping effect [107]. The 2D substrates have enhanced and highly reproducible signals, which are primary for detection and diagnosis. However, they are limited to planar detection, and Surface defects and contamination can affect performance during the SERS analysis [106, 108]. In the case of the 3D substrates, there is a large volume of hotspots and coupling effect of plasmons in the z-axis, which provides an enhancement to detect even single molecules [109]. Due to the increase in hotspots volume in the z-axis of the 3D substrates, they show better performance compared to their 2D counterparts. Xiaoyuan Geng et al. [109] have compared 2D and 3D substrates of gold nanosphere-gold octahedra hybrid and proved that the 3D substrates showed greater enhancement. While it is clear that 3D substrates show higher enhancement compared to all the other dimensionalities, they have the most complex production processes, which increases the manufacturing costs and sometimes pose challenges with uniformity and reproducibility [106]. Hence, colloidal and 1D substrates can be used for general and primary detection needs, and 2D and 3D substrates can be used for complex and highly sensitive applications [110].

On the other hand, the first-ever report by Yamada et al. [111], on the enhanced signal of pyridine by smooth surface metal oxide semiconductors, proved that materials used as SERS substrates were not exclusive to noble metals alone. Recently, materials like carbon and silicon were initially used as a template to synthesise metal nanostructures or hybrid materials to amplify Raman signals, it was found that sole carbon materials can also generate LSP by oscillating π electrons on the surface of the materials under UV light [101]. Powell et al. developed a 3D silicon mesh which is capable of detecting biomarkers of diseases and this mesh-like structure is responsible for trapping the biomolecules onto the surface and the engineered defects are responsible for the SERS effect [112]. Also, semiconducting metal oxides have gained high interest due to their tunable band gap, morphology and engineerable defects and surfaces, which can improve the charge transfer mechanism ultimately increasing the SERS activity [101, 113].

So far, we have seen the types of substrates used for SERS enhancement and now let’s delve into the different types of detection strategies used. There are two ways of detection using SERS: label-free and labelled detection [114].

Label-free detection is a simple strategy where the analyte is in direct contact with the substrate, so the signals for the spectrometer are the original fingerprint of the analyte. This strategy has great scope in biomedical applications such as determining protein structures and monitoring biological mechanisms in the cells [114]. The primary advantages of label-free detection include its simplicity, which enables fast analytical procedures without complex sample preparation steps, making it particularly valuable in time-sensitive and dependent clinical settings. Furthermore, the absence of labelling agents can eliminate or minimize the associated side effects or toxicity of the substrate or nanoprobes. However, label-free detection has several limitations, like the interference of complex systems within the biological fluids can cause a matrix effect and produce less accurate and low reproducible signals. In the complex media where multiple biomolecules are present, the selectivity and specificity of the target molecule are low, hence, we get a spectrum with multiple spectra overlaps, which are very difficult to interpret without the aid of data processing tools [115]. Another major limitation of using the label-free method is the non-intentional binding of unspecific proteins on the surface of the sensing interface, therefore ultimately fouling the surface [116]. However, combining separation techniques and precise chemometrics such as principal component analysis (PCA) and spread spectrum analysis [117] can improve signal reliability. Incorporating these add-ons paves the path for using label-free method prominently in clinical applications.

The second strategy is labelled detection using SERS. Generally, the labelled strategy is subdivided into two types, labelled substrate and SERS tags. The labelled substrate is where the substrate is functionalized with specific biorecognition molecules like aptamers or antibodies to specifically bind to the target analyte for quantitative reports. Whereas SERS tags are colloidal substrates which are nanostructures that are functionalized with Raman reporters and directly mixed with the target analyte [102, 118]. Raman reporters are molecules with intense raman signals, bound to the surface of the nanoparticles via chemical or physical interactions. They have distinct sharp peaks which can aid in detection and diagnosis. Hence, choosing the raman reporter can play a significant role in the signal enhancement and stability of the SERS tags [89, 119]. Label-based detection is an alternative option to detect complex biomolecules which are in complex media where label-less substrates struggle. Additionally, the functionalized labels or biorecognition elements can specifically target the biomarkers, pathogens, biomolecules, cellular components and tumors. This active targeting using labels ensures a higher limit of detection (LOD), improves the signal reliability and better signal-to-noise ratio [120–122]. There are three ways the labelled detection can be used for interpreting biological information [103]. The first way is when the labelled substrate is used for detection and the change in the characteristic peak of the Raman reporter with the interaction of the targeted analyte is used to determine the concentration present. The second way is competition-based detection, here there is a completion between the SERS tags and target analyte to adsorb on the substrate, so when the SERS tags first adsorb onto the substrate, the target analyte replaces the tags which ultimately reduces the signal intensity of the Raman tags to unveiling the concentration of the analyte. Another scenario is when target molecules first adsorb to the substrate via antigen-ligand binding, the SERS tags functionalized with the same antigens are used and then bind to unoccupied substrate ligands, which says the target analyte concentration. The third way is due to chemical reactions that occur between the labelled substrate and analyte the chemical changes include ionic bond formation, hydrogen bond and dipole formation. The changes in the Raman signal account for quantitative information of the analyte, for example, this method can be used to determine COVID-19 positive and negative samples due to alteration in the SERS fingerprint of the reporter [123].

Chemometrics and SERS

Improving signal reliability in SERS is a pivotal need during detection and early diagnosis. Keeping in mind that SERS signals are highly sensitive and can be easily affected by other biomolecules in complex media, hence to solve this issue the use of functionalisation of SERS probes and/or utilization of chemometric tools is implemented. Here more focus is on chemometric techniques since we have already discussed the functionalization of SERS probes previously. Chemometrics are statistical methods which interpret complex data, signal processing, and discriminant datasets [91, 124]. There are significant chemometric methods like PCA, partial least square (PLS), and multivariate curve resolution (MCR) which are highly used in SERS signal interpretation. PCA is used for better-detecting trends, visualization and data interpretation by reducing the dimensionality of the data without losing much information. Generally, PCA uses scaled data to calculate the covariance matrix, which gives the relationship between the pair of variables. Next, the use of linear algebra (eigenvectors and eigenvalues) from the covariance matrix determines the principal components. The principal component is a new variable which contains the maximum information needed from the original data set. Which ultimately reduces its dimensionality. In a typical PCA plot, each data point represents the initial variable and the principal components, the plots also show the underlying variance and similarities between variables in a data set [125, 126]. In SERS the spectra are analyzed with PCA to determine the structural information of samples in complex media [127], signal preprocessing (smoothening and baseline correction) and classification of different analytes from the large set of Raman shift data [128]. In the case of PLS, it is predominately used in omics data analyses and can also deal with multicollinearity. PLS is a supervised method because it uses a labelled dataset (input of both independent and dependent variables) for learning the relation between the datasets to make predictions on new data points [129]. Additionally, the PLS method can be used on high-noise data and also with datasets with missing values. PLS has an upper hand compared to PCA in terms of predictive modelling making it excellent for tasks like regression and discriminant of groups [130]. Its combination with SERS analysis provides predictive modelling and simplifies the complex data and discriminant analysis [131, 132]. Another chemometric method frequently used for separating individual signals from multiple overlapped signals is the MCR technique. MCR is a bilinear model which means it takes the original data set and decomposes it into two low dimensional datasets. One of the low-dimension datasets contains the pure component profile (i.e., the spectral contribution of the individual components to the whole observed spectra) and another is its corresponding concentration. Once the separation is done, MCR provides the spectra of individual species with their concentration in the total sample [133, 134]. MCR for SERS can help in signal noise reduction, signal deconvolution, quantitative analysis and identification [135, 136]. Artificial neural networks (ANN) are a well-known machine-learning technique that can be used to understand the relation between the input raw preprocessed data and output responses, even if the data is non-linear and high-dimensional. In SERS, they are widely used for signal correction, quantitative analysis and target prediction/classification [137]. Xie et al. developed an ANN model to predict different subtypes of breast cancer with the help of SERS detection, their results concluded 100% accuracy compared to other chemometric methods [138]. Another subclass of ANN is the convolutional neural network (CNN), which is good for working with spatially structured data and visual images. CNN uses convolutional layers to detect complex patterns, peaks and even scant edges. In SERS, they can aid in spectral imaging and mapping, quantitative analysis, peak localization and signal correction [139, 140]. Hence, there is no doubt that chemometrics and machine-learning models can also be implemented to improve the selectivity, reproducibility and automatization of the developed SERS-based devices.

SERS for cancer detection

It is clear that the survival rate of patients is high with early detection of cancer, however considering the recent advancements in treatment and diagnosis, most cases are still detected only when the tumour spreads to other parts of the body. There is a need for precise and advanced techniques to tackle this problem. Techniques such as SERS can be used to detect cancer (Fig. 5) due to their high sensitivity, real-time diagnosis and ability to provide unique Raman spectra for different molecules. The SERS detection is of two types, in-vivo and in-vitro. The utilization of biomarkers such as blood-based, genomic, proteomic, metabolomics and exosome are needed sources for both in-vitro and in-vivo detection [141]. Additionally, with the incorporation of different SERS-based immunoassay strategies, the tumor can be diagnosed non-invasively and rapidly [142]. As discussed previously, in the label-free strategy the spectra of biomolecules are produced by their interaction with the nanostructures. While, in the case of the labelled strategy the Raman reporters play a pivotal role in indirect sensing, the change in the intensity of the Raman reporters constitutes the information for cancer detection [142].

Fig. 5.

Schematic diagram illustrating the process of using SERS for detecting cancer. This schematic highlights the specificity, sensitivity, and non-invasive nature of SERS as a powerful diagnostic tool for early cancer detection

In-vitro SERS studies

In-vitro SERS detection has shown promising growth and development due to its molecular fingerprinting ability and highly sensitive detection of cancer biomarkers. Both labelled and label-free approaches can be used for detecting the active molecule in a laboratory setting. Over the past years, research has delved towards developing SERS-active materials for monitoring cell death and detecting overexpressed biomarkers. A few reports (Table 2) have shown the utilization of biomarkers for the detection and monitoring of tumor cells. For example, Anamika et al. [143] fabricated SiO₂@Ag substrate by thin film deposition of silver and decorated with silica nanosphere for sarcosine detection, sarcosine is a potential biomarker for early diagnosis of prostate tumor. With the use of the SERS technique, the author was able to detect up to 1.76 nM. The plasmonic substrates and colloids have complications such as coagulation and selectivity in in-vitro detection. Despite these challenges, Jiamin et al. [144] developed a SERS-based assay where the probe was a DNA-conjugated gold nanoparticle and the substrate was a silver-coated magnetic nanoparticle. The magnetic substrate used was for the separation of the biomarkers from the complex media and silver coating intensified the Raman signals. Furthermore, adding Raman reporters after the separation improved sensitivity and selective detection. The LOD of this sandwich assay was in attomolar concentration and the assay was able to detect multiple miRNAs (miR-122, miR-223, and miR-21) repressed in hepatocellular carcinoma (HCC). With regards to non-plasmonic nanostructures having low enhancement factors, Rupa et al. [145]. Reduced the size to the quantum scale and studied the properties of ZnO-based 3D quantum probes. Due to the size reduction, factors like high surface area, crystallinity, and surface defects improved the SERS signal intensity. The 3D blend nanorods acted like an extracellular matrix for cells to adhere to the surface, and the quantum probes were taken up by the cells for imaging via the endocytosis mechanism. For the in-vitro studies, two cancer cell lines (HeLa and MDAMB231) were compared with one non-cancer cell line (NIH3T3) and by the change in the ratio of Raman intensities with multivariate analysis, cancer and non-cancer cells were distinguished. Though SERS-based non-plasmonic substrates are less explored, they hold promising potential in biosensing and the diagnosis of cancer if engineered to specific needs. Recently, lateral flow immunoassays combined with SERS probes have shown higher sensitivity for cancer detection. Su et al. reported SERS-based lateral flow immunoassays for detecting HER2 and MUC1. The LOD of 3.27 × 10⁶ particles per ml for HER2 and 4.80 × 10⁶ particles per ml for MUC1 was observed. Further, they used MCR for spectral deconvolution and obtained an accuracy of 92% and 88%, respectively [146]. Using magnetic particles can aid in the maximum recovery of the SERS probe after the analysis. Treerattrakoon et al. [147] developed an assay using SERS tags for the detection and separation of mir-29a, using magnetic separation. The assay showed high selectivity for mir-29a with a LOD of 10 pM. SERS in colorectal cancer (CRC) detection has also shown significant progress. Studies using SERS with silver hydrosol substrates, coupled with PCA-linear discriminant analysis (LDA), have achieved high diagnostic accuracy (up to 91.6%) in distinguishing CRC from normal samples. They have observed decreased collagen content in rectal cancer cells, likely due to increased metalloproteinase activity [148].

Table 2.

Examples of SERS-based cancer detection

SERS active nanostructures	Targeted biomarker	Analyte source	Sensing mode	LOD	References
Fe3O4@Ag	microRNA-10b	Exosome	SERS assay	1.86 aM	[160]
SiO2@Ag	Sarcosine	–	SERS	1.76 nM	[143]
anti-B7-H6@ATP@AuNPs	B7-H6	Blood serum	SERS immunosensor	10.8 fg mL−1	[161]
Au/Ag/porous GaN	microRNA (miR-K12-5-5p)	–	SERS	8.84 × 10−10 M	[162]
Au-film deposition	PSA and thrombin	–	SERS-Microfluidics	0.01 ng mL−1 and 0.01 nM	[163]
Q-structured TiOx	EFGR	–	SERS	1 nM	[164]
Cu2O–CuO@Ag	Exosome	Serum	SERS immunosensor	–	[165]
mF-MoS2 NS	microRNA (miR-106a)	Serum	SERS-Electrochemical	67.44 fM and 248.01 fM	[166]
Au@EBP@Au NR arrays and Au@MBN@Ag NPs	5-HT	Serum	Ratiometric SERS	4.9 × 10−9 M	[167]
Au-AgNPs	microRNA (miR-21)	Serum	CHA-SERS	0.15 pM	[168]

This table lists various studies demonstrating the use of SERS for identifying and diagnosing different types of cancer, along with the target biomarkers and key findings of each study. The table emphasizes the versatility and efficacy of SERS in detecting various cancer biomarkers, showcasing its potential for early and accurate cancer diagnosis

In-vivo SERS studies

In vivo SERS detection for cancer offers real-time, non-invasive diagnosis of the tumors. SERS probes designed for in vivo detection offer several advantages like, multiplex detection capacity, stable signals without photobleaching, superior selectivity with active targeting and low auto-fluorescence background [115]. Early studies demonstrated the use of SERS nanotags for in vivo detection of three cancer cells: EGFR, CD44, and TGFβRII. The intratumor injected antibody conjugated nanotags showed maximum signal at 6 h, but the non-conjugated counterpart did not show a signal at 6 h. This study clearly shows the need for active targeting [149]. On the other hand, the clinical potential of SERS is substantial and signal enhancement relies on the metal surface; however, the interaction surface limits its effectiveness in clinical settings. Most clinical experiments have used a laser wavelength of 785 nm to induce light scattering [150]. In 2017, SERS combined PCA-LDA found success in the diagnosis the Oral squamous cell carcinoma (OSCC) and MEC Tan and colleagues conducted a clinical trial involving 135 patients with OSCC and 90 individuals with mucoepidermoid carcinoma (MEC) served as the positive control group for the SERS-based diagnosis of squamous cell carcinoma. Blood serum was used as the sample type for this study. PCA was used to determine the key components, and AuNPs were used as an appropriate substance to capture the signals. The difference in the Raman spectra was observed due to changes from the metabolic alteration [151, 152].

However, the challenges and limitations of using SERS have hindered its widespread adoption for cancer detection and diagnosis in clinical practice. Some common difficulties observed are the lack of established protocols, standardization, and reproducibility of signal intensities [153]. The reason is that parameters such as the nanomaterial structure, instrumental conditions, and choice of analyte have a direct effect on the detection capabilities of the SERS probes. The approach to address these issues would be to first have a standardized procedure and benchmark the developed SERS probes [21]. A recent successful attempt by 15 European laboratories to standardize the SERS protocol marks a primary step towards a simpler and single method for quantification [154]. Still, it requires a deeper insight and effort to reach the desired stages. Another clinical study by wang et al. [150]. showcased to integration of artificial intelligence and deep learning with SERS. They collected serum samples from 729 patients who were suffering from prostate cancer or benign prostatic hyperplasia (BPH) and were able to differentiate them using the machine learning algorithm to identify the distinctive shift in the Raman peaks. This method showed an accuracy of 85% in diagnosis. This type of combination of chemometric tools or convolutional neural networks can be another way to improve the reproducibility and accuracy of the results. Advancements in SERS-based cancer detection lie with the development of a multimodal approach which combines the advantages of SERS with other complementary techniques. Other imaging techniques like MRI, computed tomography, or ultrasound can provide precise information about the tumor location, size and microenvironment [155]. Additionally, the integration of SERS with therapeutic methods like PDT can provide an excellent strategy for diagnosis and treatment.

From the recent advancements in SERS technology, it is clear that SERS has the potential to improve conventional and commercially available methods [156]. However, due to their high cost and complex synthesis, not all reported SERS-based probes can be taken for final device integration. Currently, most SERS-based detection probes are still proof of concepts. However, more research towards the high throughput device engineering, reproducibility and cost-effective methods for bulk preparation is needed to make SERS-based probes move towards industrial-scale production [157–159].

SERS-guided photodynamic therapy

During the PDT studies at in vitro or in vivo stages, SERS can generally help in two ways. The first is by imaging the cancer cell and sub-cellular components, and the second is by in situ monitoring of the treatment or the changes in the other internal factors of the tumor. This section delves the advancements and the role of SERS imaging/monitoring in improving treatment planning and real-time outcome assessment.

SERS imaging of cell and sub-cellular components

Theranostic probes that are used for PDT treatment are functionalized or loaded with PS moiety to generate ROS. Additionally, the complementary fluorescence ability of the PS moiety can be used for fluorescence imaging (FI) of the tumor. However, relying only on FI can hold distinct disadvantages such as poor selectivity, autofluorescence and photobleaching which can affect the reliability of the results. Hence, developing probes with multiple imaging abilities is considered to address the shortcomings of using only a single imaging method. SERS-based imaging can be utilized because of its sensitivity and the ability to generate unique spectral signals [169]. Past few years, multimodal theranostic probes with SERS imaging and PDT ability have been developed. For example, False et al. [170] synthesized the first-ever nanocomposite which can generate ROS and perform both FI and SERS imaging, the SERS imaging was possible due to tagging Raman active dye 3,3′-Diethylthiatricarbocyanine iodide (DTTC) with gold nanostars (GNS) and the FI and ROS production was due to methylene blue (MB) dye. As a proof of concept, they performed preliminary in-vitro studies on BT549 breast carcinoma cells. They found that the nanoprobe showed better PDT efficiency and with 785 nm laser excitation the SERS spectra of the Raman dye were observed which could be useful for tumor cell imaging.

Additionally, the sensitivity of SERS can also improve cancer cell imaging and eliminate false positive results, Zhang et al. [171] developed a multimodal gold nanorod (GNR) system which was capable of performing SERS detection, FI and PDT treatment on HeLa tumor-bearing mice. SERS detection and FI were typically used as confirmatory tools to identify nanoparticle accumulation in tumor-bearing mice (Fig. 6A), FI showed three locations with high intensity but with the help of SERS detection, the false positive signals were eliminated. Seo et al. [172] synthesized core–shell nanoparticles where the GNR was the core and silica was the shell, further the MB dye was loaded into the nanoparticle for the ROS generation and for SERS imaging where the significant peaks at 446, 1380, and 1605 cm⁻¹ were utilized for single cells and aggregated cell imaging. Moreover, factors like the morphology, size and metal used can affect the enhancement of the Raman signals which can directly affect its sensitivity. The same group also developed a silver shell GNR core coated with mesoporous silica and loaded MB dye into it. They studied the LSPR adsorption and SERS effect by varying the silver concentration, the results indicate that compared to the previous study there was a huge increase in the enhancement of the Raman signals upon the addition of Ag metal, as for the in vitro SERS imaging, the developed core–shell nanoparticle showed 15 times stronger SERS intensities from the distributed nanoparticles on the CT-26 cells compared to the previous work [173]. In another work, Su et al. [174] synthesized a hollow graphitic carbon nitride nanosphere self-assembled on copper phthalocyanine (CuPc) and hyaluronic acid (HA), which was capable of producing photo-induced charge transfer to help SERS enhancement, tandem PDT and NIR oxygen evolution reaction. This non-plasmonic substrate had an enhancement factor of 10⁵ due to the chemical enhancement mechanism, the significant peak at 1533 cm⁻¹ of the probe was utilized for time-dependent SERS imaging of nanoprobe distribution on the HeLa cells and in vivo HeLa tumor-bearing mice (Fig. 6B). The probe also showed remarkable ROS generation in both hypoxic and normoxic conditions.

Fig. 6.

A SERS spectra and fluorescence imaging to confirm nanoparticle accumulation on tumor-bearing mice: a) picture of the tumor-bearing mice, b) fluorescence imaging, c) fluorescence spectra at various locations, d) intensity of SERS spectra corresponding to nanoparticle accumulation at different locations. Reprinted with permission from Ref. [171], Copyright

© 2013 WILEY–VCH Verlag GmbH & Co. KGaA, Weinheim. B In vivo SERS imaging of nanoparticle accumulation. A) Raman imaging of tumor-bearing mice at different time intervals and B) Its corresponding SERS spectra with an intense peak at 1533 cm⁻¹. Reprinted from Ref. [174], CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Apart from the metal nanoparticle, the Raman reporter used plays an important role in SERS detection and imaging during the PDT. Conventional Raman reporters like MB, DTTC, DTDC, Rhodamine B (RhB), 2-naphthalenethiol (NPT), and mercaptobenzoic acid (MBA) have sharp Raman signals and when functionalized on the surface of the metal nanoparticle, it can be even used to detect cancer at early stages [175]. However, most of the Raman reporters have their characteristic peaks at the biological region (< 1800 cm⁻¹) which means during in vivo or in vitro detection they are overlapped by the Raman signals of the biomolecules in the complex medium. Addressing this issue, Zhu et al. [176] developed an Au core Prussian blue (PB) shell nanostructure Here PB has multiple roles, it acts as a contrast agent for both MRI and SERS bimodal imaging, it also acts as a PS for the facilitation of PDT in the cellular level. The significant peak of PB at 2156 cm⁻¹, which is in a biological silent region with zero background interference was used for in vivo and in vitro Raman imaging.

Another important criterion for improving the treatment as well as imaging is utilizing the over-expressive biomarker receptors as a target to selectively localize the theranostic probe on the tumor can improve the treatment outcome as well as provide enhanced SERS imaging. However, to selectively target the tumor, functionalization of the nanomaterials with antibodies, aptamers or proteins and functional groups is needed, for example, TAT peptide-functionalization [177] on Silica coated GNS improved the affinity of the nanoprobe to accumulate on the tumor cells. In another work [178], folic acid (FA) was functionalized on SiO₂-coated silver-embedded GNS. Here, Folic acid is used to target Folate receptors which are over-expressive in many cancer cells like HeLa and SK-BR-3 but on the other hand, they are absent in some cells like MDA-MB-468. The results of in-vitro PDT and Imaging studies found that both PDT and SERS imaging were better for HeLa and SK-BR-3 due to the FA functionalization. Zhao et al. [179] synthesized gold nanoparticles and functionalized them with hyaluronic acid-hydrocaffeic acid, pheophorbide (Pheo) and 2-naphthalenethiol (NPT) to get a nanochain-like structure. The targeting ability of the nanoprobe was due to the presence of hyaluronic acid (HA) which can selectively target the CD44 receptor and for the detection of cancer Raman reporter NPT was used and showed excellent enhanced signals because of the SERS enhancement effect by the plasmonic metal.

Additionally, the increase in the length of the self-assembled nanochain improved the SERS enhancement factor increased. Finally, the SERS detection of the nanoparticle localization was done utilizing the peak at 765, 1064, and 1375 cm⁻¹. From their study, a dependency between the Raman signal intensity and incubation time was observed which confirms the superior targeting ability of the developed probe. In another work, the same group [180] synthesized self-assembled gold nano chains with sight modification in the PS and the targeting group used. They functionalized AuNP with folic acid, HA-hydrocaffeic acid, and Chlorin 6 for better ROS generation and improved targeting. SERS signals of NPT were again used for the detection of probe localization on the MCF-7 and A549 cells. Compared to the previous study the PDT efficacy was improved which is attributed to the PS and multiple targeting functionalities used. Similarly, Narayan et al. [181] developed a nano construct consisting of MB-loaded cucurbituril(8)-GNR and anti-Her2 monoclonal antibody. The cucurbituril(8) acted as a glue to linearly align the GNR at a sub-nanometer level for hotspot generation which enhanced the SERS signal of MB. The significant peaks of the MB dye were used for imaging the breast cancer cells MCF-7 and SK-BR-3. The functionalization of the anti-Her2 antibody was able to improve the localization of the nano construct on the cell surface and ultimately improved the imaging of the tumor cell (Fig. 7A).

Fig. 7.

A SERS imaging of SK-BR-3 cells. a) White light image of SK-BR-3 cells, b) SERS mapping of the nano construct distribution, c) Colour-coded representation where the grey corresponds to the nano constructs accumulated on the cell walls, d) 3D representation of nano construct distribution, e) Fourier filter image, f) Histogram of the peak intensity distribution after the incubation of the nano construct with SK-BR-3 cells. Reprinted with permission from Ref. [181], Copyright

© 2021 American Chemical Society. B SERS contrast agent for 3D cellular imaging. A) SERS imaging of the C-26 cell at particular points with an increase in the z-axis, B) SERS spectra correspond to the nanoprobe localization at * spot with change in depth, C) SERS spectra correspond to the nanoprobe localization at ** spot with change in depth. Reprinted with permission from Ref. [182], Copyright © 2015 American Chemical Society

Advanced studies like intracellular imaging and 3D imaging by SERS can be a breakthrough within optical imaging techniques. For example, Simon et al. [182] developed a multimodal theranostic system capable of imaging the nanoparticle distribution using SERS and FI, they synthesized gold nano aggregates encapsulated by Pluronic f127 polymer with MB dye loading. The significant peak of MB at 1623 cm⁻¹ was used for SERS imaging of nanoparticle distribution in C-26 as well as PDT with 660 nm light illumination. After the distribution, the nanoparticles were used as SERS contrast agents for 3D cellular imaging (Fig. 7B). By fixing the XY plane of the cell and imaging the Z plane at different depths was able to show the 3D distribution of the nanoparticle on the cellular components. Tang et al. [183] developed receptor-mediated synergistic PDT/PTT nanodrug which could image the cellular internalization and distribution using the SERS technique. The nano drug contained GNR functionalized with PpIX, MBA and FA, where the significant peaks of MBA peaks at 1583 and 1077 cm⁻¹ were utilized for the internalization and the distribution of the drug onto the HeLa cell. In another work, Wang et al. [184] synthesized a gold nano framework (AuNF) loaded with Doxorubicin (DOX) and functionalized with HA and Raman reporter 4-aminothiophenol (4-ATP). The AuNF possessed a high density of hotspots and also had 4-ATP conjugated on its surface via thiol bonds, with the help of SERS the material capability of the cellular internalization was confirmed, and the significant peaks at 1085 and 1585 cm⁻¹ of the Raman reporter was the key factor in producing the Raman images in both in vitro and in vivo model. Hence SERS-based materials can serve as a contrast agent for imaging of tumor cells and sub-cellular components during PDT treatment.

SERS monitoring

Monitoring the PDT treatment can provide information regarding the treatment effects, biological mechanism, quantitative analysis, outcome assessment and most importantly the changes in the internal conditions of the TME [18, 185]. For example, S. Chen et al. [186] developed a Janus nano motor capable of detecting the presence of H₂O₂ in different cancer cell lines like 4T1, HepG2, HeLa, and B16 using ratiometeric SERS signal analysis. Here the peak at 998 cm⁻¹ is the internal standard which corresponds to the C–C bending of the Raman reporter 3-mercaptophenylboronic acid (3-MPBA) and a new peak at 882 cm⁻¹ which increased with the increase in the concentration of H₂O₂. The ratio between both peaks gave information regarding the quantity of H₂O₂ in the TME. Additionally, the nanomotor also showed both PTT/PDT synergistic tumor treatment, where within 30 min of NIR radiation more than 90% of cell death was observed. In another work, Yue et al. [187] successfully demonstrated the changes in the B16 cell nucleus during the PDT with in-situ SERS monitoring (Fig. 8). During the time-dependent PDT treatment, the change in intensity and shift of the intranuclear SERS spectra of B16 cells provided valuable information regarding the denaturation of proteins, DNA fragmentation and also predicting cell apoptosis.

Fig. 8.

A SERS monitoring of miRNA content in tumor-bearing mice. a) fluorescence imaging at different time intervals, b) and c) shows the miR-21 content in the peripheral blood extracted from the mice on different days, d) Tumor regression of different mice in the group after 15 days, e) Raman imaging of the MCF-7 cells collected from different mice. Reprinted with permission from Ref. [188], Copyright

© 2019 WILEY–VCH Verlag GmbH & Co. KGaA, Weinheim. B SERS monitoring of nucleus changes after the PDT treatment. Inset (A, B, C) corresponds to the intranuclear SERS spectral changes to A- sample, B-sample + light, C- Sample + light + Ce6, D) bright field image of B-16 cell treated with sample + Ce6 + light. Reprinted from Ref. [187], CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). C SERS monitoring of nanoprobe localization to pH changes in the tumor environment. Reprinted from Ref. [190], Copyright © 2022, American Chemical Society

As discussed, earlier plasmonic nanostructures are widely explored in the field of SERS imaging and monitoring however, the recent advancements in 2D non-plasmonic nanostructure have paved the way towards SERS sensing and monitoring of PDT treatment because of their high stability, lower cost and better selectivity. For example, Liu et al. [188] developed hexagonal boron nitride nanosheets conjugated with hairpin G-quadruplex (HG) and copper phthalocyanine(CuPc). Here the CuPc played the roles of PS and in-situ SERS monitoring of MicroRNA (miR-21) and the hexagonal boron nitride substrate showed high enhancement of CuPc SERS signals, together they showed a high LOD of 0.7 fM for miR-21. With the help of ratiometric SERS sensing the peaks at 1530 cm⁻¹ and 1367 cm⁻¹ of CuPu and boron nitride were used to quantify and image the miR-21 of cell lysate in MCF-7. Also, miR-21 content in the peripheral blood of MCF-7 tumor-bearing mice (Fig. 8) was assessed. Finally, the probe could predict the dosage as well as the course of treatment for a better outcome. In another work, Li et al. [189] developed a SERS-based monitoring probe with ROS generation ability for PDT to treat U87 cells. The developed probe was a 2D graphitic carbon nitride conjugated with rhodamine B labelled peptide sequence that was highly sensitive, and selective and showed superior LOD (35 fg mL⁻¹) of metalloproteinase (MMP-9). During in vitro studies the significant peaks 1503 cm⁻¹ of RhB and 1242 cm⁻¹ of 2D graphitic carbon nitride were used for ratiometric assessment of the MMP-9 downregulation.

Recently, transition metal complexes like cyclometalated iridium have been used for SERS sensing and PDT because of their solubility and the ability to generate ROS, for example, Shanmughan et al. [191] synthesized a cyclometalated iridium (III) complex that showcased the bimodal detection of endogenous nicotinamide adenine dinucleotide (NADH) via a fluorescence On-SERS Off sensing strategy. During their study, they observed that there was a hydride shift from NADH to the metal complex which ultimately reduced the intersystem crossing and improved the fluorescence property of the complex. The quenching of a significant peak at 1042 cm⁻¹ corresponding to the C-H bending provided information such as the LOD of NADH. In another work, the same group [192], have done a comparative study on two cyclometalated iridium complexes one with a nitrile functionalized ancillary ligand(Ir-CN) and the other with hydrogen functionalized ancillary ligand (Ir-H). The Ir-CN showed better photoactivation, less cytotoxicity in dark conditions, higher ROS generation and Raman peak at 2129 cm⁻¹ due to CN vibration making it a better SERS probe. Additionally, the probe was able to selectively target the mitochondria in the MCF-7 cells and promote mitochondrial PDT. The outcome assessment of PDT was done by analyzing the changes in the SERS signals of the DNA and other sub-cellular components. Further, the conformation of apoptosis was demonstrated using SERS to monitor the release of cytochrome-C in the early apoptotic phase.

The pH dysregulation in the TME due to the Warburg effect is associated with tumor progression and metastasis. Monitoring the pH changes during the PDT is beneficial to understand the intracellular changes and dynamics mechanisms. SERS-based real-time pH monitoring can provide feedback on the changes in TME pH during PDT, which is valuable information for precise dosimetry. For example, Yue et al. [193] reported a SERS-based pH nanosensor made of gold nanorod functionalized with mitochondria-targeting peptides(MLS) and a Raman reporter 4-mercaptopyridine (4-MPy) to measure the intramitochondrial pH changes after different dosages of PDT. The significant peak at 1095 cm⁻¹ of 4-Mpy was used in pH-responsive studies with three different cell lines and from the dosage study it was found that with excess ROS, cancer cell lines (MCF-7, HepG2) had less change in mitochondrial pH compared to the normal cell line (LO2) which serves as valuable information for early diagnosis. Moreover, research towards combination therapies which utilize pH monitoring and targeting tumor acidity, to design probes capable of performing pH-sensitive drug delivery combined with PDT to enhance the treatment outcome is increasing. For example, Srinivasan et al. [194] synthesized folic acid and DOX conjugates self-assembled silver nanoparticles as a theranostic probe. where the plasmonic nanoparticle plays a dual role in enhancing the Raman signals and generating ROS for PDT. The change in the Raman peak 1453 cm⁻¹ was interpreted for the intracellular DOX release during the pH changes. Similarly, Zhang et al. [190] developed a Janus nanoparticle (JNP) capable of in situ SERS detection and performing chemo-photodynamic therapy (C-PDT). The JNP consisted of a pH-sensitive DNA sequence and an AS1411 aptamer assembled on one side of the AuNP and on the other side ATP aptamer and Raman reporter 5,5-dithiobis-2- nitrobenzoic acid (DTNB) were attached. Additionally, the co-loading of doxorubicin (DOX) and 5,10,15,20-tetrakis-(1-methyl-4- pyridyl)-21H, 23H-porphine (TMPyP4) PS molecule was done. Here JNP was able to selectively target the nucleolin of the cell membrane and was also able to detect the ATP content of the MCF-7 and HEK-293 T cells with an LOD of 2.3 nM. The probe was also able to detect the intracellular pH of the tumor cells with SERS signal intensity changes (Fig. 8C) finally, the in vivo studies on the MCF-7 tumor bearing mice model showed the synergy of C-PDT and SERS detection for improved treatment outcome.

In another work, Jibin et al. [195] synthesized gold and reduced graphene oxide hybrid and functionalized with pPIX, FA and chitosan (CS). Here, the fluorescence of the pPIX and the SERS enhancement of the Au-rGO hybrid were used to assist the intracellular PDT and minimize the side effects. Additionally, the significant peak of DOX at 455 cm⁻¹ was used to precisely monitor the drug release during light illumination and pH change. SERS and fluorescence property of the probe was also utilized to check the pharmacokinetics and biodistribution in the in vivo DLA tumor bearing mice model, the significant peaks 1580 cm⁻¹ and 1340 cm⁻¹ were used to find the time-dependent changes of the distribution at different organs.

Monitoring the intratumor oxygen stress and predicting the catalytic pathways of ROS damage to tumors during the PDT process is vital information required to gain insight into the therapeutic mechanism. However, in some cases, the ROS generated can be neutralized by some reducing species in the TME to establish the equilibrium of the cell. Hence, to tackle this problem nanoparticles with artificial enzyme-like activity capable of producing ROS and depleting reducing species are needed to improve the efficacy of the treatment. Li et al. [196] synthesized a gold core and carbon dot shell nanohybrid capable of mimicking peroxidase and glutathione oxidase-like enzyme activity. where, upon NIR irradiation, the hybrid was able to generate a PTT effect and photoexcited hot carriers by surface plasmon resonance effect which enhanced the enzyme-mimicking activities. The photoinduced peroxidase-like activity (POD) of the hybrid was demonstrated by their ability to oxidize 3,3′,5,5′-tetramethylbenzidine (TMB) with 808 nm laser irradiation. Furthermore, the SERS was used to monitor the difference in the POD activity with and without light illumination, the characteristic peaks at 1192, 1337, and 1611 cm⁻¹ of oxidized TMB were utilized and results confirmed that with light illumination there was an increase in the SERS signals which corresponds to the higher POD activity of the hybrid. Also, the glutathione oxidase-like activity was observed with 808 nm light illumination and was also confirmed with SERS monitoring. For the in vivo study, a group of 4T1 breast cancer tumor-bearing mice was treated with different concentrations of the nanohybrid, H₂O₂ and laser illumination. After 7 days the residual tumor was resected (Fig. 9a). Interestingly, when the resected tumor was treated with H₂O₂, TMB and laser illumination for 3 min, there was a sharp increase in the Raman peaks of oxidized TMB which further justifies the enzyme-like activity of the hybrid (Fig. 9e). However, after 3 min of laser illumination, the SERS signals of the oxidized TMB reduced which was due to the ROS consumption of glutathione over time (Fig. 9f). From this work, it is clear that SERS can be utilized to monitor time-dependent tumor therapy and its reaction mechanism.

Fig. 9.

A a) Treatment procedure used for in vivo study, b) and c) Tumor volume and weight in mice after treatments with variation in the concentration of the sample used, d) Schematic of intratumoral oxidative stress studies using SERS, e) and f) Time-dependent SERS spectra of oxidative stress in tumor and the characteristic bands of oxTMB respectively. Reprinted with permission from Ref. [196], CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). B Ratio of viable/dead cells significant peaks related to a) proteins and b) lipids, B) (right side) a),b),c),d) and e) PCA plots of the 4T1 cancer cells groups: (a) clusters of control and dead, b) clusters of control and viable, c) and d) clusters of control and treated with NE/PS5, and NE/PS10, respectively, (e) combined cluster of all groups. Reprinted with permission from Ref. [197], CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Similar to early detection of cancer, outcome assessments like cell survival and follow-up are both potential information which can be used to predict the efficacy of the treatment and quantify the resistant cells after PDT. However, very few protocols are used to investigate the post-treatment changes of the treated cancer cells. For example, Veloso et al. [197] demonstrated the use of SERS data of cancer cells before and after PDT to check for spectral changes. Further, they used advanced statistical tools to discriminate the data and predict the cell death mechanism. For their in vitro study, they used chloroaluminumphthalocyanine (ClAlP) loaded nanoemulsion where the concentration of CIAIP was varied (5 and 10 mmol L⁻¹) and added to 4T1 cells culture plates followed by irradiation of 660 nm laser for 10 min. After PDT treatment, the SERS spectra collected for untreated 4T1 cells (control) and treated cells (viable and dead) showed slight differences in the intensities and variation in peak positions of the proteins, lipids and nucleic acid. Interestingly, new Raman peaks at 813 and 1565 cm⁻¹ corresponding to phosphodiester bonds in DNA and nucleic bases were seen in the dead cell group revealing that the DNA fragmentation occurs due to the base pair breakages and not phosphodiester bond breakage. Additionally, the ratios of Raman peaks related to proteins and lipids for different PS-loaded nanoemulsions (Fig. 9B) were collected and peaks at 608 cm⁻¹ for protein and 1231 cm⁻¹ for lipids showing the highest difference were used to assess the cell survival after treatment. The PCA analysis separated each cell cluster (Fig. 9B) depending on the biochemical changes before the cell death mechanism occurred. Results showed that the viable cell cluster and dead cell cluster treated with a higher PS dose were closer to the control cluster meaning that they did not undergo many biochemical changes, which further confirms the observation of increased signal intensity of proteins in a lower PS. From their study, they concluded that the increase in the protein signal in treated cells was due to the triggered defence mechanism due to ROS stress during PDT treatment.

SERS and PDT are not restricted to cancer treatment and diagnosis, they are also utilized in pathogen sensing and anti-bacterial/anti-viral PDT as well. For example, Chia et al. [198] used MB dye and AuNP deposited on polylactic acid microneedles for detection and anti-bacterial PDT. The developed template showed excellent sensing of bacterial metabolite (alanine) which has a significant peak at 735 cm⁻¹, and the LOD of bacteria (S. aureus) by SERS detection was 10² CFU cm⁻². In another work, Zhou et al. [199] developed a silver-coated AuNP (Ag@AuNp) which was further covered with silica and functionalized with silicon 2,3- naphthalocyanine dihydroxide (Nc) and Vancomycin (Van). The synthesized nanocomplex was used for SERS imaging and anti-bacterial PDT. Here the SERS was used to image the binding affinity of the nanocomplex toward four different bacterial strains, and the Raman reporter 4-MBA significant peaks were used for SERS mapping. The developed nanocomplex produced observable regression of bacterial colonies both in vitro and in vivo by PDT. Addressing the issues of spectra overlapping of biological analytes and Raman reporters Zhang et al. [200] synthesized Ag-Au-Prussian blue nano jujubes functionalized with vancomycin. Utilizing the peak of PB, the SERS detection of bacteria (S. aureus, E. coli) was performed and the LOD was 56 CFU/ml. Additionally, the developed nano jujubes also showed potential anti-bacterial activity with and without laser illumination, a huge difference in bacterial death was observed, which was attributed to the singlet oxygen generation during the laser illumination. Piantanida’s group [201, 202] worked on boron complexes for tumor cell PDT, anti-viral PDT and SERS sensing. Here they developed tetracationic bis-triarylborane complexes for fluorometric and SERS sensing of DNA and RNA, the results showed that complexes with ctDNA addition observed SERS spectra quenching and the reason was due to the binding of complex and nucleic acid, additionally from their observation they stated that the triple bond of the complex played a major role in the SERS signal enhancement. Next, the best analogues were taken for treating cancer cell lines (HEK293, A549) and anti-viral studies on human adenovirus type 5 (HAdV5) with PDT. The complexes showed promising cellular uptake and effective treatment against both cancer cells and viruses (Table 3).

Table 3.

Summary of research reports published on SERS-guided PDT

Material	Cell line	Role of SERS	Raman reporter	Significant peaks	Remarks	References
Cyclometalated iridium complex (III)	HepG2	Monitoring the NADH concentration Cell death evaluation	–	1042 cm−1	LOD of NADH was 15 pM Apoptotic cell death during PDT was confirmed by SERS	[191]
Cyclometalated iridium complex (III)	MCF-7	To track the apoptotic events Monitoring the cyt c release	–	2129 cm−1	Apoptotic mechanism was confirmed with help of SERS and Annexin V-FITC	[192]
Gold Nanorod (GNR) with FA, PpIX and MBA	HeLa	To observe the nanodrug internalization and image its distribution on the cell	MBA	1077 cm−1 1583 cm−1	The targeting ability of the developed nanodrugs was confirmed by the SERS tracing function	[183]
GNR functionalized with MLS and 4-MPy	MCF-7 HepG2 LO2	Monitoring the intramitochondrial pH change	4-MPy	1095 cm−1	pH monitoring by SERS confirmed the unique spectral response of different cell towards PDT	[193]
GNR functionalized with mPEG-SH NLS & RGD	B16	Monitoring the molecular dynamics of the nucleus during PDT Cell death evaluation	–	602 cm−1 1001 cm−1 829 cm−1 1001 cm−1	PDT time-dependent cell nucleus change mechanism was predicted by SERS	[187]
GNR coated with mSiO2 & conjugated on GO	MDA-MB-231	Imaging cancer cell	–	445 cm−1 1397 cm−1 1624 cm−1	Early detection of cancer cells was achieved using SERS imaging	[175]
SiO2 coated-gold nanostars (GNS) with MB and DTTC	BT549	Imaging of cancer cell	DTTC	–	Theranostic probe with both SERS imaging and PDT was developed	[170]
TAT-SiO2 coated-GNS with PpIX and DTDC	BT549	Imaging of cancer cell	DTDC	1120 cm−1 1150 cm−1	The intracellular imaging was possible using SERS	[177]
SiO2 coated Ag-GNS functionalized FA	HeLa SK-BR-3	Imaging of cancer cell	P MBA	1076 cm−1 1580 cm−1	The targeted imaging ability of the developed probe was confirmed by the SERS	[178]
2D Graphitic carbon nitride with RhB labeled peptide	U87	Ratiometric monitoring of MMP-9 PDT assessment with SERS imaging	RhB	1242 cm−1 1503 cm−1	LOD of MMP-9 was 35 fg mL−1 PDT efficacy was confirmed with ratiometric SERS	[189]
PpIX & DTTC encapsulated silica coated GNR	HeLa	in vivo detection of nanoparticle accumulation	DTTC	508 cm−1	Bimodal detection by SERS and fluorescence was able identify the location of tumor	[171]
CuPu and HG functionalized on 2D hexagonal boron nitride	MCF-7	Monitoring miR-21 in different stages of tumor Imaging miR-21 in cancer cells	–	1530 cm−1 1367 cm−1	LOD of miR-21 was 0.7fM Monitoring of miR-21 in different stages of tumor helped early detection and treatment planning	[188]
MB-Pluronic loaded Au	C-26	To observe the nanodrug distribution on the cell	MB	1623 cm−1	SERS imaging and FI was utilized to detect the nanoparticle in 3D level	[182]
MB-loaded SiO2 coated-GNR	CT-26	Imaging of cancer cell	MB	446 cm−1 1380 cm−1 1605 cm−1	The single cancer cell imaging was possible using SERS	[172]
SiO2 coated MB-loaded Ag@GNR	CT-26	Imaging of cancer cell	MB	450 cm−1 502 cm−1	The intracellular imaging was possible using SERS	[173]
anti-Her2 tagged MB loaded cucurbituril [8]-GNR	SK-BR-3 MCF-7	To observe the nano construct internalization and image its distribution on the cell	MB	452 cm−1 1394 cm−1 1623 cm−1	The prepared construct was able to specifically accumulate on the cell surface which was confirmed by SERS	[181]
CS-PpIX-FA conjugated on Au-rGO hybrid	MDA-MB-231 DLA	in vivo imaging Track DOX release Time dependent changes	–	455 cm−1 1340 cm−1 1508 cm−1	Bimodal detection was able observe the distribution, PDT effect and DOX release to the tumor	[195]
TMPyP4 and DOX coloaded Janus AuNp	MCF-7 HEK-293 T	Monitoring ATP Conc in cancer cells Imaging of cancer cells	DTNB	1327 cm−1	LOD of ATP was 2.3 nM pH dependent drug release was monitored by SERS	[190]
DOX and FA functionalized silver NP	MES-SA SKOV-3	Monitor cellular uptake Monitoring DOX release PDT and cell death mechanism	–	1453 cm−1 1160 cm−1 240 cm−1 1110 cm−1	pH dependent drug release was monitored. Apoptotic cell death during PDT was confirmed	[194]
Graphitic carbon nitride assembled with CuPc and HA	HeLa	Imaging of cancer cell in vivo distribution of nanoprobe	CuPc	1533 cm−1	The in vitro and in vivo imaging was possible using SERS	[174]
Prussian blue shell Au core functionalized with HA	4T1	Imaging of cancer cell	PB	2156 cm−1	The single cancer cell imaging and in vivo imaging was possible using SERS	[176]
Ce6, NPT, FA and HA-HCA functionalized AuNp	MCF-7 A549	To detect the probe’s localization on the cancer cell	NPT	1380 cm−1 765 cm−1 1067 cm−1	The developed SERS active theranostic probe was able to selectively accumulate on the cancer cells	[180]
Pheo, NPT and HA-HCA functionalized AuNp	HeLa	To detect the probe’s localization on the cancer cell	NPT	1375 cm−1 765 cm−1 1064 cm−1	The developed SERS active theranostic probe was able to selectively accumulate on the cancer cells	[179]
Catalase, TAPP and 3-MPBA functionalized JNP	4T1	Ratiometric Monitoring and quantification of H2O2	3- MPBA	998 cm−1 882 cm−1	The in-situ detection of H2O2 by SERS can be used for early detection of cancer	[186]
HA and DOX functionalized Au nano framework	MDA-MB-231	Imaging of cancer cell	4-ATP	1085 cm−1 1585 cm−1	The in vitro and in vivo imaging was possible using SERS and photo acoustics	[184]
Nanoemulsions loaded with ClAlP	4T1	Cell death evaluation	–	608 cm−1 1231 cm−1	SERS and PCA was used to determine the necrotic cell death mechanism	[197]
Gold core and poly(styrene-alt-maleic acid) shell	HeLa	Imaging of cancer cell	MB	1623 cm−1	SERS imaging and FI was utilized to detect the nanoparticle accumulation	[203]
Gold core and carbon dot shell	4T1	Real time monitoring of catalytic process in PDT	TMB	1192 cm−1 1337 cm−1 1611 cm−1	The oxidative stress generated during the PDT was monitored by SERS	[196]
MB and AuNP deposited polylactic acid microneedles	–	Detecting metabolites of bacteria (S. aureus) Monitoring PS concentration	–	735 cm−1 458 cm−1	LOD of MB and was below 200 ppb LOD of S. aureus was 102 CFU cm−2	[198]
Nc and Van functionalized SiO2 coated Ag@AuNp	–	To image bacterial strains (B. Subtilis, E. faecium, E. faecalis, E. coli)	4-MBA	1075 cm−1 1585 cm−1	SERS imaging was able to demonstrate the binding affinity of nanocomplex	[199]
Ag-Au-Prussian blue nanojujubes with Van	–	Detecting bacterial strains (S. aureus, E. coli)	PB	2150 cm−1	LOD of bacterial strains was 56 CFU/mL	[200]
diethynylarene-linked bis(triaryl borane) cations	HEK-293 T A549 WI-38	Sensing ctDNA	–	1557 cm−1 1257 cm−1	SERS sensing of ctDNA was up to 0.1 molar ratio High anti-viral and tumor cells PDT was observed	[202]

This table provides an overview of key studies, including the type of cancer, PS used, SERS substrates, and notable outcomes. This summary highlights the integration of SERS with PDT, showcasing advancements in cancer treatment through enhanced targeting, monitoring, and therapeutic efficacy

Future perspective and conclusion

This article reviewed the scope of combining SERS and PDT for improved cancer theranostics. While it’s clear that PDT can be a potential therapeutic method for treating cancer, it still requires developments addressing the shortcomings of the PS. Additionally, imaging and monitoring of PDT is a vital factor which should be considered to enhance the understanding of the cellular and therapeutic mechanisms. SERS-based imaging and monitoring have been recently exploited due to their sensitivity and non-invasive methodology. Furthermore, the integration of both SERS and PDT offers researchers the distinct advantages of imaging tumor cells, monitoring subcellular changes, and tracking drug release. However, due to the lack of established standardized protocols, and reproducibility of results, SERS use in clinical studies is still in the preliminary stages. Efforts have been made to optimise the experimental setup, detection strategies and result validation using chemometric tools, implementation of these factors can elevate the clinical translation of SERS-based imaging and monitoring approaches [204, 205].

Apart from discussing the importance of combination therapy (SERS and PDT), this article also provided an overview of the fundamental mechanisms, detection strategies and applications in cancer and pathogen detection using SERS-based approaches. Furthermore, the fundaments and limitations of PDT and the role of imaging and monitoring during PDT treatment were also discussed for a better understanding of the reason for combining both PDT and SERS.

Interestingly, multimodal nanoprobes capable of performing multiple therapies and diagnostics are extensively researched due to their ability to provide complementary results which can aid in confirmatory assessment. It is to be noted that the majority of the probes contain PS moiety either loaded or functionalized on the surface of the shell which is useful for generating ROS for treating the tumor, However, the monitoring and imaging with only the fluorescence property of the PS is not sufficient. Hence the development of SERS active PS which can enable SERS detection along with ROS generation with the same wavelength irradiation. This approach can be useful for real-time monitoring of the PS photobleaching which can improve the PS dosimetry [206]. Additionally, these nanoprobes generally contain a plasmonic core which can be used for Raman signal enhancement making it an effective SERS contrast agent for imaging applications.

The efficiency of the SERS for in vivo imaging and monitoring can be improved by utilizing more advanced Raman detectors which are capable of deep tissue imaging and wide field view. Furthermore, combining SERS with chemometrics and artificial intelligence can help discriminate the SERS spectra to refine the complex biochemical changes during the treatment [207].

The immune system can hinder the SERS-guided PDT nanoprobes therapeutic intervention by assuming they are some foreign substances or even pathogens, hence developed nanoprobe must be engineered with strategies to evade immune system detection. Surface modification with coatings such as PEGylation, carbohydrate and other Biomimetic polymers has shown the cloaking ability to evade the immune system. Additionally, size and morphological changes can play a significant role in evading the immune system. Studies have shown immune suppression (avoid being phagocytosed) while using rod-shaped morphology compared to spherical-shaped. Similarly, a high surface charge on the probe either anionic or cationic is targeted by the immune system. By keeping all these points in mind, the systematic and efficient development of SERS-guided PDT therapeutic probes can perform better [208].

From our perspective, we believe that the fruitful outcome of the combinational approach of incorporating SERS with PDT active theranostic probes would be when the SERS probe can generate the Multiplexing capability of the biomolecular matrix and in situ monitoring of their changes for understanding the destruction mechanism. Additionally, if the PS moiety used is SERS active then the dosimetry of the PS used can also be calculated. Lastly incorporating chemometric tools and machine learning algorithms improves the reliability and accuracy of the results obtained, developing such SERS active PDT probes and methodologies could be the way forward.

In conclusion, SERS can undoubtedly improve the efficacy of the theranostic probe. While it cannot directly enhance the effectiveness of the treatment on the tumor, it can aid the imaging and detection aspect of the tumor cells by monitoring and imaging the subcellular changes. Additionally, it can also help in monitoring the drug delivery and in some case monitor the PS photobleaching during the PDT treatment. Hence this information is crucial and can be used for developing precision and advanced cancer theranostic probes.

Author contributions

TJ and RBA Conceptualized the idea. TJ, SG, HS, RR, NR, and SS did a literature survey and worked on the Manuscript framework. TJ, SG, and HS wrote the initial part of the manuscript. RBA, JP and JS revised the draft manuscript. JP and RBA proofread and supervised the preparation of the original manuscript. All the authors reviewed the manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent to publish declaration

Not Applicable.

Clinical trial number

Not Applicable.

Competing interests

The authors declare no competing interests.