Nano-biosensors for circulating tumor markers: advancing liquid biopsy toward precision cancer diagnostics.

Nano-biosensors are well-suited for detecting CTMs because nanomaterials (1–100 nm) confer high surface area and unique electronic/optical properties that amplify biorecognition signals (El-Tanani et al. 2025 ; Wang et al. 2020 ; Roy et al. 2024 ). For example, gold nanoparticles, carbon nanotubes (CNTs), graphene, quantum dots (QDs), and nanocomposites are widely used to construct electrochemical, plasmonic, and fluorescent sensors with ultralow detection limits. These platforms can achieve sub-femtomolar sensitivity and high specificity for the target circulating tumour markers while remaining portable and low-cost (Roy et al. 2024 ). Electrochemical nanobiosensors, in particular, offer “high selectivity, high sensitivity, and simple equipment” at low cost, making them ideal for point-of-care liquid biopsy (Sfragano et al. 2023 ; Khaksari et al. 2023 ; Zha et al. 2023 ). Together with advances in device miniaturisation and wireless/readout integration, nanostructured biosensors are bringing precision oncology diagnostics closer to the patient (Roy et al. 2024 ). The nanobiosensors can be classified into 2 categories based on the mechanism of action, namely, Electrochemical and Optical nanobiosensors (Turner 2013 ) (Fig.  3 ). Below, we discuss the two mechanism-based nanobiosensors and their differences. Electrochemical nano-biosensors form a major class of CTM detectors. These devices transduce biomolecular binding (e.g. DNA hybridization or antigen–antibody interaction) into electrical signals (current, voltage or impedance). Electrochemical biosensors have shown consistent results in detecting cancer biomarker proteins such as CEA, CA125, and HER2. These electrochemical biosensors are enzyme-integrated, immune-based platforms that can produce electroactive products. These nano-integrated platforms have enabled the detection of CEA at concentrations of 1.2 pg/mL (Nadeem-Tariq et al. 2026 ). Incorporation of nanomaterials greatly enhances their performance. For example, gold nanoparticles (AuNPs) and gold nanocages are commonly conjugated to carbon electrodes (graphite, CNT-, or graphene-coated) to enhance conductivity and binding capacity (Roy et al. 2024 ). A graphene oxide nanosheet sensor modified with phospholipid-PEG has been used to isolate and detect EpCAM‐positive lung cancer CTCs with high efficiency. Similarly, CNT‐decorated electrodes and AuNP–aptamer assemblies have yielded reusable, label‐free electrochemical aptasensors for ctDNA and cell detection. These devices are prized for portability and ease of use: as summarised in a recent review, electrochemical nano‐biosensors are “user‐friendly, portable, and cost‐effective”, with rapid response and very low detection limits. In practice, they have demonstrated impressive metrics: for example, an electrochemical DNA sensor using Au‐CNT nanocomposites on a printed carbon electrode achieved a ctDNA limit of detection (LoD) of 42.8 fM across a wide 10 − 7 to 10–14 M range (El-Tanani et al. 2025 ; Wang et al. 2024a , b , c ). Likewise, an aptamer‐functionalized PDA nanoparticle interface on a pencil graphite electrode captured A549 lung cancer cells with a sensitivity of 25 cells/mL over 102–105 cells/mL. In general, electrochemical sensors combine very high sensitivity with simple potentiostatic hardware, label-free real-time readouts, and multi-channel formats (e.g., 96-well EFIRM microarrays). Their limitations include potential nonspecific binding or fouling by complex blood matrix and the need for surface regeneration or calibration. Nevertheless, these are being addressed by antifouling coatings and redundant microfluidic flow paths (Khaksari et al. 2023 ; Zha et al. 2023 ). Optical nanobiosensors are another key modality that uses light–matter interactions for CTM detection. Surface plasmon resonance (SPR) sensors, for instance, employ metal (usually gold) nanofilms to detect changes in refractive index upon biomolecule binding. SPR devices are inherently label-free and can provide real-time kinetic data. They are valued for their high sensitivity and stability, and recent designs have achieved the detection of single-nucleotide ctDNA mutations. However, SPR systems typically require precise optical alignment and can be bulky or expensive to construct. Surface-enhanced Raman spectroscopy (SERS) sensors provide an alternative optical approach: analyte molecules are adsorbed onto roughened metal nanoparticles or nanostructures, dramatically amplifying their Raman scattering. SERS biosensors can achieve ultra-sensitivity and allow multiplexed “fingerprint” readouts, since each molecule gives a unique spectrum. For example, SERS nanotags have been used to quantify multiple EX microRNAs and protein markers in plasma simultaneously. The cost is greater instrumentation complexity and potential variability in the nanoprobe substrates. Fluorescence‐based sensors, often using fluorescent nanomaterials such as QDs or fluorescent dyes, are widely used for CTM analysis. QDs offer high photostability, broad excitation spectra, and bright emission, enabling highly sensitive detection and imaging of target DNA or EX. In one design, QDs served as FRET donors in multiplexed assays to detect cancer‐related DNA and protein markers with femtomolar sensitivity. Fluorescence methods can suffer from background autofluorescence and typically require labelling, but their high signal‐to‐noise makes them attractive for many assays. Lateral‐flow immunoassays (paper‐based strips) with gold or QD labels also fall in this category: these provide qualitative or semi‐quantitative results rapidly at the bedside. Overall, optical platforms tend to achieve very high sensitivity and can easily multiplex, but at the expense of more complex optics or labelling steps compared to electrical sensors (Zha et al. 2023 ). Although both mechanisms aim to make CTM detection easier, they have their own specifications. Electrochemical sensors are better suited to low-cost operations, but optical systems will always require sensitive reading equipment. As such, electrochemical sensors can be widely utilised but run the risk of non-standardised information and low device reliability. For example, a comparative study of the two systems’ effectiveness in detecting CA-125 found that optical nanobiosensor-based platforms offered greater selectivity. In contrast, electrochemical nanobiosensor-based platforms showed better sensitivity and greater ease of use. Optical nanobiosensor platforms proved more expensive and struggled with multiple analytes. Electrochemical nanobiosensor-based platforms showed false positives. Both platforms lacked clinical validation (Pourmadadi et al. 2023 ). Throughout these strategies, shared themes arise. Nanomaterials, including graphene, CNTs, AuNPs, and QDs, become central figures in transduction – e.g., gold provides easy biofunctionalisation and surface plasmon properties, and carbon nanostructures permit rapid electron transfer and excellent loading of nucleic acid probes. Representative performance figures are remarkable: detection limits are achievable at femtomolar DNA or single-cell concentrations, and multiplex assays can quantitate dozens of markers in parallel. For instance, multichannel EFIRM electrochemical arrays detect EGFR and PIK3CA ctDNAs from saliva within 3 h, and SERS immunoassays have revealed multiple EX protein markers by distinct spectral tags. Biocompatibility is achieved through the use of inert nanomaterials (gold, silica, carbon) and by embedding sensors within hydrogels or soft substrates. Drawbacks include nonspecific protein adsorption by serum, matrix effects in blood, and regulatory standards for clinical translation. Nevertheless, the advantages – ultra-sensitivity, multiplexing, speed, and ultimately personalised diagnostics – are driving rapid progress.(Khondakar et al. 2023 ; Beg et al. 2022 ; Singh et al. 2023 ; Kim et al. 2024 ). In practical terms, nano-biosensors have demonstrated the ability to detect virtually all classes of CTMs. For example, aptamer‐based FET sensors have quantified specific ctDNA mutations (EGFR, KRAS) at sub‐pM levels, and electrochemical immunosensors have measured tumour protein markers such as CEA and CYFRA21‐1 at pg/mL levels. EXs have been profiled using plasmonic chips and magnetic-bead microfluidics to monitor biomarkers such as EGFR and PD-L1. CTCs have been enriched on nano‐textured surfaces and counted by impedance or optical imaging down to tens of cells per mL. In each case, nanostructuring dramatically enhances capture efficiency or signal readout compared to bulk sensors. As noted in recent literature, integration of such platforms into connected health systems – including wireless links and AI analytics – is now paving the way for precision oncology that is “patient‐centric, real‐time and data‐driven” (Roy et al. 2024 ; Khondakar et al. 2023 ; Beg et al. 2022 ; Singh et al. 2023 ; Kim et al. 2024 ). A comparative analysis of the different characteristics of the nanobiosensor mechanisms is given in Table  2 . Table 2 Table showing the comparative analysis of the Electrochemical nanobiosensors and Optical nanobiosensors Parameter Electrochemical nanobiosensors Optical nanobiosensors Representative references Detection principle Electrical (current, voltage, impedance) Light-based (SPR, SERS, fluorescence) Turner ( 2013 ); Zha et al. ( 2023 ) Sensitivity Very high (fM–pg/mL) Extremely high (single-molecule or fM level) El-Tanani et al. ( 2025 ); Wang et al. ( 2024a , b , c ) Specificity Moderate High Pourmadadi et al. ( 2023 ) Instrumentation Simple, portable (potentiostat) Complex (laser, optics, spectrometer) Khaksari et al. ( 2023 ); Zha et al. ( 2023 ) Cost Low High Roy et al. ( 2024 ); Pourmadadi et al. ( 2023 ) Label requirement Often label-free Often requires labeling (except SPR) Zha et al. ( 2023 ) Matrix compatibility Limited by fouling and nonspecific adsorption Better matrix tolerance but affected by sample preparation Khaksari et al. ( 2023 ); Zha et al. ( 2023 ) Reproducibility Challenging due to surface variability Substrate variability, especially in SERS Pourmadadi et al. ( 2023 ) Clinical readiness Early translational Early translational Nadeem-Tariq et al. ( 2026 ); Zha et al. ( 2023 ) Best application POC and portable detection Multiplexed and high-sensitivity laboratory assays Quesada-González and Merkoçi ( 2018 ); Roy et al. ( 2024 ) Ease of use High Moderate Khaksari et al. ( 2023 ) Best suited for ctDNA, miRNA, proteins ctDNA, miRNA, proteins, exosomes Wang et al. ( 2024a , b , c ); Roy et al. ( 2024 ) Key nanomaterials used AuNPs, CNTs, graphene, PDA nanostructures, Au–CNT hybrids Gold nanofilms, SERS substrates, quantum dots Roy et al. ( 2024 ); Chiorcea-Paquim et al. ( 2023 ) Representative performance ctDNA detection ~ 42.8 fM; A549 cell detection ~ 25 cells/mL SERS nanotags and QD-FRET assays with femtomolar detection El-Tanani et al. ( 2025 ); Wang et al. ( 2024a , b , c ) Limitations Non-specific adsorption; matrix interference; need for surface regeneration Complex optics; labeling requirements Khaksari et al. ( 2023 ); Pourmadadi et al. ( 2023 ) Table showing the comparative analysis of the Electrochemical nanobiosensors and Optical nanobiosensors The electrochemical and optical nanobiosensing technologies have been integrated into 3 major platformsfor detecting CTMs. These include microfluidic-integrated systems (Li et al. 2022 ), wearable biosensors (Ren and Cui 2026 ), and point-of-care devices (González and Merkoçi, 2018 ) (Fig.  3 ). Nano-biosensors with microfluidic capabilities are devices that combine miniaturised fluid handling with detection components to process nanoliter‐sized patient samples in the “lab on a chip” format. By integrating nanomaterial-enhanced sensors into microscale channels, they can detect and differentiate CTMs more effectively. Microfluidic devices require only microliters of blood and can run assays more quickly due to their short diffusion distances. They are highly automatable and amenable to high‐throughput: for instance, a microfluidic chip with 28 parallel electrodes tested samples in under 2 h. Microchannels can be coated with capture antibodies or surfaces functionalized with nanostructures for selective CTC or EX trapping. Microfluidic biosensors offer “high automation capability, fast reaction time, and cost-effectiveness” and can integrate sample pretreatment and detection. An illustrative example is a PDMS microchip containing a CNT‐coated gold array that immunocaptured EpCAM+ tumour cells from whole blood, then measured impedance changes. Another is a nanoplasmonic “nPLEX” chip that sorted EXs and then detected EGFR/PD‐L1 surface proteins by SPR shifts (Roy et al. 2024 ). Such systems achieve LoDs of 103–104 vesicles/mL for EXs or tens of cells/mL for CTCs. The downsides of microfluidics include more complicated fabrication and the potential for channel clogging or bubble formation. Ongoing advances in materials (e.g. hydrophilic coatings) and modular design are rapidly overcoming these challenges. Crucially, microfluidics brings full assay automation closer to point‐of‐care use (Roy et al. 2024 ). Wearable and point-of-care (POC) nanobiosensors represent the newest frontier in cancer diagnostics, enabling decentralised and even continuous monitoring of CTMs. These devices are designed for use outside the laboratory – for example, skin patches, microneedle arrays, or handheld readers that often interface with smartphones or cloud platforms. Wearable biosensors offer non-invasive, real‐time sampling: for instance, sweat or interstitial fluid can be probed for ctDNA fragments or EX biomarkers using on‐patch electrodes or chemical sensors. The same tumour biomarkers tracked by liquid biopsy (ctDNA, RNA, proteins) could, in principle, be monitored by wearable platforms (Hua et al. 2025 ). In practice, prototype wearable cancer sensors have been demonstrated, e.g., an adhesive microneedle array with integrated electrodes was shown to extract and electrochemically detect a ctDNA analogue in simulated interstitial fluid. The aim is to bring cancer surveillance “out of the hospital and into daily life” (Khondakar et al. 2023 ; Beg et al. 2022 ; Singh et al. 2023 ; Kim et al. 2024 ). POC devices more broadly – including simple dipsticks, smartphone‐based microscopes, and paper microfluidics – are also gaining traction. These POC sensors can deliver rapid results (minutes to hours) with minimal training or equipment. The literature stresses that POC biosensors have the potential to provide “rapid, low‐cost, and minimally invasive” diagnostics at or near the patient, addressing the need for early cancer screening across various settings. The incorporation of nanomaterials into POC platforms improves sensitivity (e.g., AuNP‐based lateral flow and nanoparticle‐enhanced fluorescence) without sacrificing portability. Of course, wearable/POC formats may trade some sensitivity versus lab assays and must ensure biocompatibility for patient comfort. However, they present opportunities for regular and prolonged measurements of minimal residual disease, relapse, or therapeutic response (Roy et al. 2024 ). The nanobiosensor platforms discussed above are designed to detect CTMs in cancer patients. However, their intended purposes differ. Microfluidic platforms integrate various body-fluid handling techniques with on-chip detection and can handle multiple analytes. In contrast, Wearable/Point-of-Care nanobiosensors focus only on the detection of a select analyte. Microfluidics platforms prioritise specificity, while Wearable/point-of-care platforms focus on ease of use and real-time monitoring of analytes. Wearable/Point of Care platforms may suffer from data discrepancies because they depend more on end-user interpretation, which may not be adequately trained (Tavakoli et al. 2022 ). A further comparative analysis on these platforms has been mentioned in Table  3 . Table 3 Comparative analysis of the different nanobiosensor platforms CTM / analyte Cancer types detected Clinically validated approaches Proof-of-concept approaches Most practical sensor platform Real-world trade-offs Major limiting constraint Take-home message Representative references CTCs Breast, prostate, colorectal, lung cancers CellSearch-based immunocapture Nanotextured chips, impedance microfluidics, graphene capture surfaces Microfluidic-electrochemical systems High specificity and single-cell characterization, but enrichment is slow, expensive, and technically demanding Rare-cell enrichment and EMT-associated marker loss Integrated enrichment and detection platforms are essential for reliable metastatic cancer analysis Alix-Panabières and Pantel ( 2013 ); Ju et al. ( 2022 ); Lin et al. ( 2021a ); Magri et al. ( 2023 ) ctDNA Lung, colorectal, pancreatic, breast cancers Digital PCR, BEAMing, NGS Electrochemical FET sensors, plasmonic nanoarrays, CRISPR sensors Electrochemical and SPR platforms Electrochemical systems are portable and affordable, whereas optical systems provide superior sensitivity but require advanced instrumentation Extremely low abundance in early-stage cancer ctDNA analysis requires ultra-sensitive amplification-integrated systems for mutation profiling and therapy monitoring Wan et al. ( 2017 ); Bartolomucci et al. ( 2025 ); Connal et al. ( 2023 ); Wang et al. ( 2024a , b , c ) Exosomes / EVs Pancreatic, ovarian, breast, glioblastoma, lung cancers No universally validated biosensor platform nPLEX chips, SERS nanotags, fluorescence nanoprobes Optical-microfluidic systems Optical systems enable multiplex biomarker profiling but increase assay complexity and instrumentation cost Isolation purity and vesicle heterogeneity Optical multiplexing combined with microfluidics is highly advantageous for exosome characterization Yu et al. (( 2022a , b ); Wang et al. ( 2024a , b , c ); Lee et al. ( 2018 ); Liu et al. ( 2021 ) Circulating miRNAs Lung, liver, gastric, renal, prostate cancers RT-qPCR-based assays FRET biosensors, nanozyme amplification systems Fluorescence and electrochemical sensors Fluorescence systems improve multiplexing and sensitivity, but normalization variability limits reproducibility Lack of standardized endogenous controls miRNA panels show strong early diagnostic potential but require analytical standardization Mitchell et al. ( 2008 ); Zendjabil ( 2024 ); Petrou and Ladame ( 2022 ); Alemayehu et al. ( 2024 ) Tumour-associated proteins (CEA, PSA, CA125, HER2) Breast, ovarian, prostate, colorectal cancers ELISA, chemiluminescence immunoassays AuNP immunosensors, CNT immunosensors, plasmonic chips Electrochemical immunosensors Simple and clinically established assays but lower cancer specificity False positives from benign inflammatory conditions Protein biomarkers remain clinically useful when integrated with multi-marker nanobiosensor systems Lee et al. ( 2016 ); Sultan et al. ( 2024 ); Pourmadadi et al. ( 2023 ); Zhou et al. ( 2024 ) Tumour-educated platelets (TEPs) Lung, ovarian, breast, pancreatic cancers No clinically standardized platform RNA-seq profiling, platelet transcriptomics integrated with AI Sequencing-integrated biosensor workflows High diagnostic accuracy but computationally intensive and difficult to standardize Platelet activation and protocol variability TEPs offer promising tumour-origin profiling but require robust standardization and validation Best et al. ( 2015 ); Best et al. 2018 ; Sol and Wurdinger ( 2017 ); Zhu et al. ( 2023 ) Multi-CTM integrated panels Multi-cancer detection Experimental multimodal liquid biopsy systems Hybrid graphene-gold nanochips, multiplex plasmonic arrays Integrated microfluidic-optical-electrochemical systems Highest diagnostic accuracy but complex assay integration and calibration Inter-platform variability and data integration complexity Multi-CTM biosensing represents the future of precision oncology diagnostics Kuligina et al. ( 2024 ); Wang et al. ( 2024a , b , c ); Choi et al. ( 2025 ); Roy et al. ( 2024 ) Comparative analysis of the different nanobiosensor platforms Fig. 3 Showing the basic types of nanobiosensor technologies and the platforms in which the technologies are integrated Showing the basic types of nanobiosensor technologies and the platforms in which the technologies are integrated

The current methods of cancer detection include tissue biopsies, MRI, CT and PET scans. Biopsy-based methods are limited by factors such as observer variability, the lack of observable tissue heterogeneity, and limited availability of tissue samples (Gromek et al. 2024 ). Tissue biopsy remains the current gold standard for cancer detection. Solid tissue biopsy enables molecular characterisation of cancer. Tissue biopsies allow observation of tissue architecture, immunohistology, and the presence of driver mutations. A major drawback involving solid biopsies is their invasive nature, which patients are not comfortable with, and it becomes challenging to collect regularly during the treatment. Additionally, open surgical biopsies can increase the chances of infections and bleeding. Solid tissue biopsies also suffer from an inability to identify tissue heterogeneity, leading to potential sample bias. Tissue biopsies cannot identify clonal evolution as it is not feasible to continuously monitor using invasive procedures (Connal et al. 2023 ). Liquid biopsies are those in which body fluids are used for analysis (Shen et al., 2025 ). The fluids are collected minimally invasively, allowing repeated sampling. Body fluids can include blood, urine, pleural effusion, ascites fluid, cerebrospinal fluid, and saliva (Ma et al. 2024 ). Liquid biopsies enable the detection of circulating tumour cells, circulating tumour DNA/RNA, tumour platelets, and tumour-derived exosomes as major molecular markers of cancer (Boukovala et al. 2024 ). The drawbacks are present at various stages, including sampling, issues of specificity and sensitivity, analysis, cost-effectiveness, and the lack of genetic biomarkers in several cancers (Purkayastha et al. 2023 ). The primary issue with biomarker-based biopsies is that both cancerous and non-cancerous tissues can produce the same biomarkers, even in the absence of cancer, due to tissue injury (Nagpal et al. 2016 ). A case in point is the PSA protein, which is used to detect prostate cancer. The PSA protein can be elevated in men suffering from prostatitis and benign prostate hyperplasia. Similarly, the ovarian cancer biomarker CA-125 can be elevated in menstruation, endometriosis and pregnancy (Connal et al. 2023 ). CA 15 − 3 is a biomarker for breast cancer, but it is elevated in other conditions like hepatitis, pelvic inflammation, and lactation (Zafar et al. 2025 ). This necessitates further exploration of newer biomarkers that are exclusively associated with elevated expression in cancer. For this, CTMs are being evaluated for their potential as cancer-specific biomarkers, which can be used for early cancer detection and monitoring of therapy response. Fig. 1 Overview of cancer biomarker sources, types, and detection strategies, highlighting solid and liquid biopsy approaches Overview of cancer biomarker sources, types, and detection strategies, highlighting solid and liquid biopsy approaches

The integration of artificial intelligence (AI) with CTM-based cancer detection holds potential for more accurate, faster detection of cancer, its type, and stage. AI can play a major role in analysing CTM fragmentation patterns, leading to the rise of “fragmentomics” (Dong et al. 2025 ). AI can help immensely in deciphering the variability and heterogeneity of CTMs, arising from cancer stage, type, and source. The integration of AI and machine learning (ML) with CTM detection will enable the analysis of complex datasets at a higher pace and accuracy, leading to faster determination of the features and genetic makeup of the cancer and thus allowing faster treatment initiation. The Circulating Cell Free Genome Atlas (CCCGA) is an example of combining AI and ML with knowledge of circulating tumour markers to address the challenge of faster, more accurate cancer detection while overcoming the challenges of detecting CTMs. The program aims to integrate ML with cfDNA sequence data spanning the whole genome. ML has been further used in various programs to integrate cfDNA data with cancer type and stage, while preventing false positives from non-cancer samples (Tiwari et al. 2025 ). The integration of CTM composition with neural network-based ML algorithms, such as CNNs, has been examined for their ability to detect metastasis in cancers. The study involved ML-based analysis of various CTM mutations that occur uniquely at the metastatic stage of cancer. The study results showed that integration with ML led to a more than 90% increase in the specificity and accuracy of detecting metastasis in cancer samples. The increased accuracy of the diagnosis could help overcome reliance on solid-tissue biopsy as a conclusive marker of metastasis (Mgbole 2025 ; Tiwari et al. 2025 ). Studies on Non-Small cell lung cancer have shown that ctDNA, like CTM, has high accuracy in predicting disease occurrence, selecting the best therapy regimen by analysing mutations, and detecting treatment response. The integration of ctDNA detection with AI enabled finer detection of NSCLC subtypes, stages, and response to therapy, as AI-enabled analysis of minimal residual disease in NSCLC patients receiving different treatments. This integration allowed the prediction of relapse risk with high accuracy and facilitated timely changes in therapeutic modalities (Thalambedu et al. 2025 ).

Current diagnostic strategies rely on tumour biomarkers obtained through solid and liquid biopsies, with liquid biopsies offering advantages in minimal invasiveness, repeat sampling, and real-time disease monitoring. Emerging circulating tumour markers, such as CTCs, ctDNA, EVs, TAPs and TEPs, provide improved insight into tumour heterogeneity, disease progression, and therapeutic response. Despite their promise, challenges related to low abundance, technical complexity, and specificity of isolation remain. Nanobiosensor-based platforms, including electrochemical, optical, and microfluidic systems, have demonstrated high sensitivity for detecting rare tumour markers from minimal sample volumes. Advances in surface functionalization, affinity ligands, and nanozyme-based signal amplification have further enhanced detection performance. While several nanobiosensor technologies have shown encouraging clinical results, large-scale validation, technological standardisation, and accessibility, particularly in low- and middle-income settings, remain key barriers. Integration with artificial intelligence represents a promising path toward improving diagnostic accuracy and accelerating clinical translation.

The signal amplification and biorecognition strategies discussed above determine the analytical ability and clinical applicability of nanobiosensors for CTM detection. Surface-engineered nanomaterials, plasmonic coupling systems, nanozyme-mediated catalytic amplification, and ligand-based recognition interfaces collectively determine sensitivity, specificity, multiplexing capability, and detection limits. Consequently, the integration of these amplification and recognition mechanisms has enabled the development of nanobiosensor platforms capable of detecting rare CTMs in clinically relevant biofluids (Quesada-González et al. 2018 ). The following section discusses how these engineering strategies have been translated into practical applications for cancer diagnostics, therapeutic monitoring, and real-world clinical validation. Multiple nano-biosensor platforms have undergone preclinical and clinical validation across diverse cancer types. For CTC detection, nanostructured interfaces have demonstrated superior performance compared to existing enumeration systems. A graphene oxide–based microfluidic nanochip demonstrated significantly enhanced CTC capture efficiency in patients with metastatic breast, prostate, and lung cancers, outperforming the FDA-cleared system (Wang et al. 2009 ). Similarly, incorporating gold nanoparticles into microfluidic enrichment devices improved capture efficiency in gastric cancer patients and enabled downstream single-cell molecular characterisation (Sun et al. 2017 ). For ctDNA analysis, nanobiosensors have helped overcome limitations associated with low variant allele fractions. A plasmonic gold nanoarray sensor detected KRAS mutations in plasma samples from pancreatic cancer patients with attomolar sensitivity, surpassing qPCR assays (Fu et al. 2023 ). A complementary approach utilising single-molecule FRET–based nanosensors identified EGFR mutations in plasma from non–small cell lung cancer (NSCLC) patients with higher sensitivity than droplet digital PCR (Choi et al. 2021 ). Additional studies employing nanostructured electrochemical sensors demonstrated reliable multigene ctDNA detection in colorectal cancer cohorts (Kumar et al. 2024 ). Exosome profiling has also benefited significantly from nanomaterials. A graphene oxide–based microfluidic biosensor achieved rapid and accurate detection of ovarian cancer–derived exosomes, with an area under the curve exceeding 0.95 in clinical plasma specimens (Vafadar et al. 2025 ). The nanoplasmonic exosome (nPLEX) platform, which utilises periodic nanohole arrays, detected breast cancer–specific exosomal markers (e.g., HER2, EpCAM) with 100-fold greater sensitivity than standard ELISA (Singh et al. 2018 ). Nano-biosensor–enabled quantification of circulating miRNAs has facilitated sensitive detection of key oncogenic signatures. A quantum dot–based fluorescence biosensor identified serum miR-21 in breast and colorectal cancer patients with substantially improved sensitivity compared with RT-qPCR (Singh et al. 2018 ). A nanozyme-amplified assay similarly enhanced detection of miR-155 in lymphoma patient plasma, correlating strongly with disease burden (Babar et al. 2012 ). TAP detection has also been greatly improved. Gold nanoparticle–amplified immunosensors detected PSA at limits as low as 10 fg/mL in clinical specimens (Suresh et al. 2018 ). Carbon nanotube–based immunosensors yielded enhanced detection of CA-125 in ovarian cancer, demonstrating clinical applicability in patient cohorts (Yang et al. 2025 ). Together, these diverse applications highlight the translational potential of nanobiosensors in early detection, risk stratification, and longitudinal assessment of therapeutic response. Given the complexity and heterogeneity of cancer, multiplexed biomarker detection provides a more comprehensive molecular signature than single-analyte assays. In this context, nano-biosensor platforms offer exceptional capability for integrating multiple sensing modalities. A hybrid graphene–gold nanochip enabled simultaneous detection of CTCs, ctDNA mutations, and exosomal proteins from NSCLC patient plasma, achieving a diagnostic accuracy of 92% in a clinical cohort (Aleixandre andHorrillo 2025 ). In parallel, a plasmonic nanohole array biosensor enabled multiplexed detection of tumour-associated proteins (HER2, EGFR, EpCAM) from breast cancer samples, generating individualized biomarker profiles (Choi et al. 2025 ). These multimodal biosensing systems represent a significant advancement toward personalised oncology, where integrated biomarker signatures may guide therapeutic decision-making with greater precision. Nano-biosensors outperform traditional diagnostic assays across several critical parameters. Their exceptionally low limits of detection, often in the attomolar to femtomolar range, surpass those of ELISA, qPCR, and imaging modalities, enabling detection of early-stage disease that may otherwise remain undiagnosed. Quantum dot–mediated miRNA assays, for instance, routinely exceed the sensitivity of RT-qPCR platforms (Babar et al. 2012 ). Liquid biopsy–based nanosensing is inherently minimally invasive and compatible with serial sampling, enabling real-time monitoring of tumour evolution, therapy response, and the emergence of resistance mutations such as EGFR T790M (Choi et al. 2021 ). The multiplexing capacity of nanostructured microfluidic and plasmonic systems enhances diagnostic accuracy by capturing the biological complexity of tumours through multi-marker detection. Additionally, many nanobiosensor systems offer rapid assay times and low per-test costs, making them suitable for decentralised and point-of-care testing. Despite their promise, significant barriers remain before nano-biosensors can achieve widespread clinical adoption. Biological variability in circulating biomarker levels across patients and cancer types complicates the standardisation of diagnostic thresholds. Sample processing, particularly for rare events such as CTCs, remains technically challenging, potentially affecting reproducibility. Furthermore, most studies to date are limited by small cohort sizes, highlighting the need for multi-centre trials with diverse patient populations. Additional challenges include integrating biosensor platforms into existing clinical workflows, the need for automated data analysis pipelines, and regulatory approval processes. Overcoming these limitations will require collaborative efforts spanning engineering, clinical oncology, regulatory science, and industry partnerships (Sengar et al. 2024 ). A summary of the compatibility attributes of different CTMs across nanobiosensor classes, along with their clinical readiness and restrictions, is presented in Table  4 . Table 4 Comparative analysis on the CTMs and the compatibility with different nanobiosensor class CTM / analyte Cancer types detected Clinically validated approaches Proof-of-concept approaches Most practical sensor platform Real-world trade-offs Major limiting constraint Take-home message CTCs Breast, prostate, colorectal, lung cancers CellSearch-based immunocapture Nanotextured chips, impedance microfluidics, graphene capture surfaces Microfluidic-electrochemical systems High specificity and single-cell characterization, but enrichment is slow, expensive, and technically demanding Rare-cell enrichment and EMT-associated marker loss Integrated enrichment and detection platforms are essential for reliable CTC analysis in metastatic cancers ctDNA Lung, colorectal, pancreatic, breast cancers Digital PCR, BEAMing, NGS Electrochemical FET sensors, plasmonic nanoarrays, CRISPR sensors Electrochemical and SPR platforms Electrochemical systems are portable and affordable, whereas optical systems provide superior sensitivity but require advanced instrumentation Extremely low abundance in early-stage cancer ctDNA analysis requires ultra-sensitive amplification-integrated systems for mutation profiling and therapy monitoring Exosomes / EVs Pancreatic, ovarian, breast, glioblastoma, lung cancers No universally validated biosensor platform nPLEX chips, SERS nanotags, fluorescence nanoprobes Optical microfluidic systems Optical systems enable multiplex biomarker profiling but increase assay complexity and instrumentation cost Isolation purity and vesicle heterogeneity Optical multiplexing combined with microfluidics is highly advantageous for exosome characterization Circulating miRNAs Lung, liver, gastric, renal, prostate cancers RT-qPCR-based assays FRET biosensors, nanozyme amplification systems Fluorescence and electrochemical sensors Fluorescence systems improve multiplexing and sensitivity, whereas electrochemical systems support low-cost rapid testing Lack of normalization standards and extraction variability miRNA detection benefits from highly sensitive nucleic acid sensing and standardized normalization workflows Tumour proteins (CEA, PSA, CA125, HER2) Prostate, ovarian, breast, colorectal, lung cancers ELISA, immunoassays, PSA/CEA testing AuNP immunosensors, CNT immunoplatforms Electrochemical immunosensors Electrochemical systems are inexpensive and portable, while optical assays improve multiplexing at higher cost False positives from inflammation and benign diseases Protein biomarkers remain the closest CTM class to routine clinical implementation Tumour-educated platelets (TEPs) Lung, breast, colorectal, pancreatic, renal cancers None RNA profiling and sequencing-integrated platforms AI-assisted sequencing and biosensor hybrid systems Strong diagnostic potential, but requires sequencing infrastructure, bioinformatics, and standardized preprocessing Lack of standardized platelet isolation workflows TEPs show high promise for tumour typing and mutation profiling but remain technically complex Multi-CTM panels Multi-cancer screening and precision oncology applications None Hybrid AI-assisted multimodal nanobiosensors Integrated electrochemical-optical microfluidics Highest diagnostic accuracy and tumour representation, but computationally intensive and difficult to standardize Data integration and multiplex calibration Multi-CTM systems represent the most promising future strategy for precision oncology diagnostics Comparative analysis on the CTMs and the compatibility with different nanobiosensor class

Early detection and continuous monitoring of cancer remain as a significant clinical challenge owing to the heterogeneity, rarity, and instability of circulating tumour biomarkers. In the last twenty years, Circulating Tumour Markers (CTM) have become important biomarkers for liquid biopsy methods of diagnosing cancer with little or no surgery. CTMs are diverse molecular and cellular particles released from primary tumours and metastatic sites and enter bodily fluids, especially blood plasma and serum (de Abreu et al. 2025 ; Ma et al. 2024 ). Liquid biopsy examines the CTMs released into bodily fluids (such as blood plasma and serum) to monitor cancer in a minimally invasive manner (Bartolomucci et al. 2025 ). CTMs include key entities such as circulating tumour cells (CTCs), circulating tumour DNA (ctDNA), extracellular vesicles (EVs) majority of which is exosomes, tumour-associated RNAs such as microRNAs (miRNAs), tumour-associated proteins, and tumour-educated platelets (TEPs) (Fig.  1 ) (Bartolomucci et al. 2025 ; Ma et al. 2024 ). The use of CTMs in liquid biopsy approaches have some advantages over conventional tissue biopsy due to their non-invasive nature (need for surgery), their ability to capture real-time tumour heterogeneity, the feasibility for repeated longitudinal sampling, and they also reflect the entire tumour genome across all (sub)clones of cancer cells (de Abreu et al. 2025 ; Hinestrosa et al. 2022 ). However, the clinical application of CTM-based diagnostics is impeded by their scarcity in bodily fluids, significant heterogeneity, rapid clearance, and instability, especially in early-stage cancers characterized by minimal tumour burden (Connal et al. 2023 ; Deng et al. 2022 ). This section provides a comprehensive overview of the biological characteristics, clinical significance, detection challenges, and applications of each major CTM type, laying the foundation for understanding how nano-biosensors address these limitations. A comparative analysis for highlighting the major features of the different CTMs has been incorporated in Table  1 (Fig. 2 ) . Fig. 2 Types of CTMs and their abundance Types of CTMs and their abundance Table 1 Comparison of the different aspects of circulating tumour markers CTM Preferred matrix Abundance & stability Clinical use Clinical validation status Reproducibility Key constraint Representative references CTCs Whole blood Extremely low abundance (1–10 cells/mL blood); moderate stability Prognosis, metastasis monitoring, therapeutic response assessment CellSearch clinically validated; most nano-enabled systems unvalidated Moderate; affected by EMT heterogeneity and variable capture efficiency Extremely low abundance; EpCAM loss during EMT; technically demanding enrichment Alix-Panabières and Pantel ( 2013 ); Ju et al. ( 2022 ); Lin et al. ( 2021a , b ); Magri et al. ( 2023 ) ctDNA Plasma / serum Low abundance in early-stage cancer; highly fragmented with short half-life Mutation profiling, minimal residual disease detection, therapy monitoring PCR and NGS clinically validated; nanobiosensor systems largely experimental Moderate to high under optimized conditions Very low abundance in early-stage cancer; interference from normal cfDNA Wan et al. ( 2017 ); Hinestrosa et al. ( 2022 ); Bartolomucci et al. ( 2025 ); Zhou et al. ( 2024 ) Exosomes / EVs Plasma / serum Highly abundant (~ 10⁹ vesicles/mL); highly stable due to lipid bilayer protection Early diagnosis, therapy monitoring, molecular profiling Mostly proof-of-concept studies Low to moderate due to EV heterogeneity and isolation variability Difficult separation of tumour EVs from normal vesicles and serum contaminants Yu et al.( 2022a , b ); Wang et al. ( 2024a , b , c ); Liu et al. ( 2021 ); Lopez et al. ( 2023 ) Circulating miRNAs Plasma / serum / exosomal fraction Moderate abundance; highly stable due to protein/EV association Early diagnosis, prognosis, therapy-response monitoring Not clinically validated Low due to normalization and extraction variability Lack of standardized endogenous controls; inter-study variability Mitchell et al. ( 2008 ); Zendjabil ( 2024 ); Kornienko et al. ( 2024 ); Petrou and Ladame ( 2022 ) Tumour proteins (CEA, PSA, CA125, HER2) Serum / plasma Moderate to high abundance; relatively stable Routine cancer screening, prognosis, disease monitoring Clinically validated protein biomarkers exist High for conventional assays; moderate for nano-enabled systems False positives due to inflammation, pregnancy, and benign diseases Lee et al. ( 2016 ); Sultan et al. ( 2025 ); Desai and Guddati ( 2023 ); Zhou et al. ( 2024 ) Tumour-educated platelets (TEPs) Whole blood / platelet fraction High abundance of platelets; RNA profiles relatively stable after isolation Cancer detection, tumour typing, mutation profiling Proof-of-concept only Low due to platelet activation and protocol variability No standardized isolation workflow; influence of inflammation and medications Best et al.( 2015 ); Best et al. ( 2018 ); Sol and Wurdinger ( 2017 ); Varkey and Nicolaides ( 2021 ) Multi-CTM panels Whole blood / plasma Variable depending on analyte combination Precision oncology, integrated diagnosis, patient stratification Not clinically validated as integrated systems Moderate; affected by inter-platform variability Complex data integration; assay standardization; multiplex calibration Kuligina et al. ( 2024 ); Wang et al. ( 2024a , b , c ); Connal et al. ( 2023 ); de Abreu et al.( 2025 ) Comparison of the different aspects of circulating tumour markers CTCs are intact, viable cancer cells that have dissociated from primary tumours and metastatic sites, subsequently entering the bloodstream, thereby constituting a critical element of the metastatic cascade (Alix-Panabières and Pantel 2013 ; Allen 2024 ; Ju et al. 2022 ). CTCs were first described in 1869 by Ashworth, they originate from a process involving epithelial-to-mesenchymal transition (EMT), intravasation into blood vessels, survival in the hostile circulatory environment, and possible extravasation into distant tissue beds (Alix-Panabières and Pantel 2013 ; Ju et al. 2022 ; Lin et al. 2021a , b ). CTCs show how dynamic tumour biology is because, as they circulate singly or in aggregated clusters (microemboli) and provide a real-time view into tumour heterogeneity, dissemination dynamics, and therapeutic resistance (Allen 2024 ; Sánchez-Herrero et al. 2022 ). The biology of CTCs is intricate; tumour cells frequently undergo epithelial-to-mesenchymal transition (EMT) to intravasate, and CTCs may migrate individually or as multicellular aggregates. CTC clusters (microemboli), constituting approximately 1% of detected CTCs, exhibit significantly greater metastatic potential, with estimates indicating a 20 to 100-fold increase in efficiency for seeding new tumours compared to individual cells (Chu et al. 2021 ; Paoletti et al. 2019 ). Clustering likely confers survival advantages, tumour cells in clusters stick together via adhesion molecules (such as homophilic CD44 interactions, which activate survival signals through PAK2/FAK and OCT4) and are shielded by platelets and galectin/MUC1 complexes that protect against shear stress and immune attack (Paoletti et al. 2019 ). In many cancers, higher CTC counts are linked to more advanced disease and a worse prognosis (Chu et al. 2021 ). Despite their scarcity, CTCs are of great interest because they can directly seed metastases. Most CTCs die quickly, and only a small number of them live long enough to spread to other organs, however, those that do may initiate new tumours (Ju et al. 2022 ). Consequently, CTCs are regarded as a “seed” of metastasis, and their identification and characterization provide insights into metastatic potential (Ju et al. 2022 ; Lin et al. 2021a , b ). In practice, this means that even a single CTC in a blood sample can carry prognostic significance. CTCs serve as a real-time “liquid biopsy” because they carry the tumour’s genetic and proteomic information, offering a dynamic window into tumour heterogeneity and evolution that complements traditional tissue biopsy. Studies on breast cancer revealed that ~ 4% of cases exhibited exclusively epithelial CTCs, while ~ 26% had predominantly mesenchymal CTCs and ~ 6% had mixed states (hybrid E/M), necessitating multiple markers (Topa et al. 2025 ). This diversity has clinical consequences: mesenchymal CTCs are generally more resistant to chemotherapy and more aggressive than epithelial CTCs (Allen 2024 ; Lin et al. 2021a , b ; Topa et al. 2025 ). Indeed, single-cell RNA sequencing of CTCs has demonstrated that highly mesenchymal CTCs frequently downregulate translational genes and display stem-like characteristics (e.g., CD44 + CD24 – /low, ALDH1) along with multidrug-resistance profiles, which correlate with a markedly poor treatment response and reduced progression-free survival (Topa et al. 2025 ). CTCs are also physically larger and less deformable than blood cells, a property exploited by filtration and microfluidic devices (Ju et al. 2022 ). One of the hardest challenges in CTC detection is that they are extremely rare in peripheral blood (Hong and Zu, 2013 ). In non-metastatic breast cancer (early-stage cancer), there are usually less than 1 CTC per 10 mL of blood, only 1–5.9% of cases showed a levels 5 CTCs/7.5 mL or higher. In blood, there are usually 1–10 CTCs among ~ 5 billion red blood cells and ~ 1 million white blood cells per millilitre of blood (Ilie et al. 2014 ; Ju et al. 2022 ; Lin et al. 2021a , b ). Diagnosis of cancers using CTCs depend upon the expression of tumour markers like epithelial cell adhesion molecule (EpCAM), cytokeratins (epithelial intermediate filament proteins) and other tumour antigens (HER2, EGFR, PSMA, etc.). However, CTCs often undergo epithelial-mesenchymal transition (EMT), resulting in the loss of certain epithelial markers and thus the CTCs lose their ability to relate with and diagnose cancers (Ju et al. 2022 ; Lin et al. 2021a , b ). The relatively low numbers of CTCs makes the process of isolation and detection of specific markers in CTCs highly technical and dependent on advancements in the techonology associated with immunocapture, microfluidics, and single-cell sequencing techniques that are aimed at identifying CTCs (Ma et al. 2024 ). As such, CTC based detection has remained restricted to proof of concept and limited cohort studies. In practice, platforms combine these above discussed features for enrichment, for instance, CellSearch uses magnetic beads coated with anti-EpCAM followed by cytokeratin staining and CD45 exclusion to isolate CTCs (Mayo et al. 2013 ). But because many CTCs downregulate EpCAM during EMT, newer technologies use cocktail antibodies or microfluidics to catch CTCs that have low or no EpCAM. Devices that work based on size, like membrane filters or spiral chips, take advantage of the fact that tumour cells are usually bigger and stiffer than blood cells (Yu et al. ( 2022a , b ). Following capture, isolated CTCs may be enumerated and subsequently analyzed through immunofluorescence, culture, or single-cell sequencing. Genomic analyses, specifically single-cell DNA sequencing, demonstrate that CTCs undergo continuous evolution characterized by frequent copy-number alterations and mutations that are distinct from the primary tumour. This provides insights into emerging resistance mechanisms that were not identified in previous tumour biopsies, establishing CTCs as a dynamic biomarker for tumour evolution (Lin et al. 2021a , b ; Topa et al. 2025 ). Combining CTC enumeration with markers like CEA or detection of CTC clusters further enhances prognostic precision. CTC clusters amplify the danger posed by CTCs, further highlighting why sensitive, multimodal capture methods are vital. In practice, patients with a higher count of CTCs generally experience poorer outcomes. For instance, in metastatic breast, prostate, and colorectal cancer, the FDA-approved CellSearch® system (which captures EpCAM-positive CTCs) has shown that in metastatic breast cancer, ≥ 5 CTCs per 7.5 mL (CellSearch) predicts worse outcomes, while in metastatic colorectal cancer, ≥ 3 CTCs correlate with shorter progression-free and overall survival (Magri et al. 2023 ). Consequently, the enumeration of CTCs and molecular analysis (via immunostaining or single-cell sequencing) can aid in staging, assess therapeutic response, and possibly direct treatment (for instance, by identifying actionable mutations in CTC genomes). Serial CTC monitoring often outperforms single measurements; patients whose CTC counts drop to zero during treatment experience significantly longer remissions than those with persistent or rising CTC levels (Magri et al. 2023 ). In short, both the presence and trajectory of CTCs during therapy provide powerful prognostic information. ctDNA comprises short DNA fragments (approximately 130–200 base pairs) released into the bloodstream by tumour cells undergoing apoptosis, necrosis, or active secretion (Hinestrosa et al. 2022 ; Hitchen et al. 2025 ). It constitutes a small fraction of total cell-free DNA, usually less than 1% in early-stage cancer, and their abundance increases with the increase in tumour burden and stage (Hinestrosa et al. 2022 ; Zhou et al. 2024 ). Since ctDNA originates from cancer cells throughout the body, it reflects tumour heterogeneity in a single sample and often carries tumour-specific somatic mutations, copy-number changes, methylation signatures, or even tumour-associated viral DNA (Hinestrosa et al. 2022 ; Leung et al. 2016 ), providing a composite “liquid biopsy” snapshot of the tumour genome (Bartolomucci et al. 2025 ).A key feature of ctDNA is its rapid turnover with half-life ranging from minutes to a few hours in plasma (Gao et al. 2022 ), as nucleases and organ filtration (liver, spleen, kidney) swiftly clear the fragments (Alexander et al. 2024 ). This means that ctDNA levels is highly dynamic, for example, effective treatment can make ctDNA levels drop quickly, while tumour recurrence can make ctDNA levels rise before other clinical signs. ctDNA is also highly fragmented and is usually associated with nucleosomes. Patterns of fragment lengths and end motifs may also hold diagnostic signals (Bartolomucci et al. 2025 ). However, the short half-life means precise timing of blood sampling is critical for detection (Gao et al. 2022 ). The level of ctDNA in plasma is usually correlates with tumour load and disease stage (Hinestrosa et al. 2022 ; Zhou et al. 2024 ), though this relationship can be affected by factors like tumour proliferation and cell death rates. Patients exhibiting elevated ctDNA levels typically present with more progressed disease and poorer survival rates (Bartolomucci et al. 2025 ). Importantly, elevated ctDNA frequently precedes radiologic or clinical indicators of progression by several weeks to months, for instance, increased ctDNA has been detected ~ 11.5 months prior to radiologic relapse in colorectal cancer and ~ 3 months in lung cancer (Bartolomucci et al. 2025 ). The detection rates for ctDNA differ depending on the type of cancer. For example, preoperative ctDNA is present in about 50–85% of early-stage cancers but more than 70% of metastatic cancers. Most patients who eventually have a recurrence have detectable ctDNA after surgery, while most patients who do not have a recurrence do not have ctDNA (Hyugen-Hoang et al. 2025 ). ctDNA presents certain challenges for real-time applications. The very low levels of ctDNAs make it difficult to detect potential mutations in critical genes (Wan et al. 2017 ). Currently, several methods, such as Tagged-amplicon deep sequencing (TAm-Seq) and cancer personalised profiling by deep sequencing (CAPP-Seq), are being tested and optimised for ctDNA isolation and identification. However, none have been validated using large-scale clinical samples. The low ctDNA levels make it challenging to suppress background noise, which affects the assays’ sensitivity to distinguish ctDNA sources in cases of tumour heterogeneity and cfDNA from non-tumour tissues (Wang et al. 2024a , b , c ). Because ctDNA is so rare, we need very sensitive methods to find it. Digital PCR, BEAMing, and error-corrected NGS enable the identification of mutant ctDNA molecules at levels below 0.1%. This limits ctDNA analysis to centres with a strong technical and financial infrastructure (Bartolomucci et al. 2025 ). In clinical settings, ctDNA assays are utilized for mutation genotyping and therapy monitoring. The FDA-approved plasma tests identify EGFR mutations in lung cancer to facilitate targeted therapy, and serial ctDNA tracking can indicate the development of resistance (e.g., new KRAS or EGFR mutations). These assays often outperform traditional biomarkers, detecting relapse earlier than imaging or serum markers (Parums 2025 ). However, sensitivity is limited by the tumour’s DNA shedding rate; very low or low-shedding tumours may produce undetectable ctDNA, leading to false negatives (Parums 2025 ). Early cancer screening using ctDNA is notably difficult, as tumours at early stages release negligible amounts of DNA (frequently < 0.1% of cfDNA) (Bartolomucci et al. 2025 ; Connal et al. 2023 ). To achieve 95% sensitivity in screening, very large blood volumes (about 150–300 mL) would need to be analyzed (Connal et al. 2023 ), which is why researchers are exploring combined multi-analyte approaches. In general, ctDNA is the most widely used liquid biopsy analyte to date. ctDNA offers a minimally invasive insight into tumour genotype and burden, establishing it as a fundamental component of precision oncology for customizing treatment and assessing outcomes (Bartolomucci et al. 2025 ). Extracellular vesicles (EVs) are nano-sized membrane-bound particles released by virtually all cells, including exosomes (30–120 nm; a major subtype), microvesicles (100–1000 nm), and apoptotic bodies (500–3000 nm) (Dilsiz 2024 ). In cancer, tumour cells release EVs into the bloodstream at extraordinarily high levels (approximately 109 particles per mL), significantly surpassing the number of circulating tumour cells (Yu et al. ( 2022a , b ). This abundance is a key advantage for biomarker applications: EVs can be harvested in bulk from a single blood sample. Furthermore, EVs exhibit inherent stability in biofluids, as their lipid bilayer protects enclosed nucleic acids and proteins from circulating nucleases and proteases (Yu et al. ( 2022a , b ), facilitating highly sensitive downstream analysis. Exosomes (EX) from tumours contain many functional biomarkers, such as DNA, RNA (mRNA, miRNA, lncRNA), proteins, and lipids, that reflect the molecular state of the cancer that produced them (Ghosh et al. 2024 ; Javdani-Mallak et al. 2025 ). Proteomic profiling shows that exosome protein content varies by tumour type; for example, specific plasma exosome signatures can distinguish pancreatic cancer from lung cancer (Yu et al. ( 2022a , b ). EXs have specific surface markers, such as the tetraspanins CD9, CD63, and CD81. They also often carry tumour antigens or oncoproteins(Wang et al. 2024a , b , c ; Yu et al. ( 2022a , b ). The nucleic acids contained within EXs reflect the tumour genotype. For example, glioblastoma patient EVs have been found to contain mutant EGFRvIII mRNA (Figueroa et al. 2017 ), and cancer-related microRNAs (such as miR-15a-5p in endometrial cancer and miR-1247-3p in metastatic liver cancer) are very common in patient EVs (El Hayek et al. 2024 ; Zhou et al. 2021 ). These findings underscore that EXs accurately show tumour status and can reveal dynamic changes in tumour biology. These characteristics render EXs an appealing “liquid biopsy” reservoir for cancer biomarkers, allowing detection of signals from tumour EV DNA, RNA, and proteins with a single blood draw. Researchers are developing multiplexed assays to exploit this reservoir. For example, circulating EXs miRNAs (miR-21, miR-155, etc.) and lncRNAs have been proposed as diagnostic markers in many cancers, and tumour proteins such as mutant EGFR, prostate-specific antigen, or tumour-specific glycoproteins may be detected on or within EVs (LohajováBehulová et al. 2023 ). Preliminary studies show that this method has a lot of potential in real life: EVs enriched with the tetraspanin TSPAN1 exhibited a sensitivity of 75.7% in distinguishing colon cancer patients (Lee et al. 2018 ), while EXs containing PSA and CD81 were exclusively identified in prostate cancer patients, demonstrating 100% sensitivity and specificity Similarly, EVs displaying phosphatidylserine (PS) on their surface have been identified as significant cancer biomarkers, with notably increased PS-positive EV levels documented in early-stage ovarian, breast, and pancreatic cancers (Jin et al. 2025 ; Logozzi et al. 2019 ). These examples showcase the diverse tumour signals encapsulated by EXs. Despite this promise, significant technical hurdles remain. The main problem is separating the rare tumour-derived EVs from the huge number of normal vesicles (Berti et al. 2024 ). Standard techniques such as ultracentrifugation, size-exclusion chromatography, and others can collect large numbers of EVs. However, due to low specificity, preparations are often contaminated with abundant serum proteins, lipoproteins, and non-tumour vesicles (Li et al. 2025 ). Immunoaffinity capture, which uses antibodies against EV surface proteins, can selectively enrich tumour vesicles, but it can only work with known markers (Lopez et al. 2023 ). In practice, the nanoscale size and heterogeneity of EVs often mean that co-purified debris, such as lipoproteins, cell-free DNA, and normal EVs, overwhelm the tumour signal (Yu et al. ( 2022a , b ), thereby reducing assay sensitivity and specificity. Nonetheless, the accumulated evidence highlights the diagnostic value of EXs. Meta-analyses of EV-based biomarkers in thousands of patients show that the pooled sensitivities for finding cancer are about 62%, and the specificities are about 76% (summary AUC ≈ 0.88) (Liu et al. 2021 ), indicating a very high potential accuracy. Tumour-derived EXs combine the high abundance and stability of a liquid biopsy reservoir with tumour-enriched molecular content (Yu et al. ( 2022a , b ). They have already yielded highly specific cancer biomarkers in proof-of-concept studies and achieving their full clinical potential will necessitate standardized, ultra-sensitive techniques for EV isolation and analysis. miRNAs are short noncoding RNAs (~ 22 nucleotides) that post-transcriptionally regulate gene expression and are frequently dysregulated in many cancers (Komatsu et al. 2018 ; Mitchell et al. 2008 ). In contrast to mRNAs, circulating miRNAs are remarkably stable, despite the high RNase activity in blood (Allegra et al. 2012 ; Mitchell et al. 2008 ). This stability is due to their protection within protein-RNA complexes (such as Argonaute1/2) or their encapsulation in extracellular vesicles (Kornienko et al. 2024 ). Quantitative studies have identified specific miRNAs to be abundant in plasma, reflecting their stability (Katayama et al. 2025 ; Mitchell et al. 2008 ). Tumour-associated proteins (TAPs) comprise another class of circulating markers. Classic serum biomarkers, prostate-specific antigen (PSA), carcinoembryonic antigen (CEA), alpha-fetoprotein (AFP), CA-125, and CA19-9, are secreted by tumour cells (Desai and Guddati 2023 ; Zhou et al. 2024 ). These proteins are utilized in clinical settings but frequently exhibit insufficient sensitivity in early stages, like CA-125, which is elevated in approximately 50% of stage I ovarian cancer patients, while a significant proportion (50%) of early-stage cases go undetected (Lee et al. 2016 ). For example, CEA protein remains at < 3ng per mL of blood in healthy individuals, while in patients suffering from lung and breast cancer at the early stages, the levels of CEA can increase upto 200 ng/mL, while in metastatic cancers it can reach concentrations > 1000 ng/mL (Duffy et al. 2007 ). Individual circulating miRNAs serve as cancer biomarkers, while miRNA panels improve specificity. Many types of tumours have higher levels of established oncomiRs like miR-21 and miR-155 (Cicatiello et al. 2025 ; Mihai et al. 2024 ). In prostate cancer, serum miR-141 was among the first miRNAs to show differential expression and distinguish between patients from healthy individuals (Sapre and Selth 2013 ). A panel of four miRNAs: miR-21-5p, miR-150-5p, miR-145-5p, and miR-146a-5p, had an AUC of about 0.94 for renal carcinoma compared to benign disease (Chen et al. 2021 ). Patients with gastric cancer exhibit elevated levels of miR-17-5p, miR-17-3p, miR-18a, miR-19a, miR-19b, miR-20a, and miR-92a as diagnostic biomarkers (Li et al. 2017 ). Exosomal miRNAs, originating from tumour-derived vesicles, have been characterized for diagnostic and prognostic purposes. In prostate cancer patients, increased levels of exosomal miR-1246 and miR-4644 have been observed, while exosomal miR-196b-5p has been associated with chemoresistance in colorectal cancer (Jafari et al. 2023 ; Ren et al. 2017 ). In general, circulating miRNAs provide a convenient multi-gene signal in liquid biopsy and can be detected by qRT-PCR or sequencing (Naranbat et al. 2025 ). miRNAs associated with cancers can be present in the range of 10–50 ng per mL blood (El-Daly et al. 2023 ). Proteogenomic projects like CPTAC have identified proteins that are linked to cancer genome alterations and tumours (Savage et al. 2024 ). Combining multiple protein markers improves diagnostic performance; for example, adding periostin (POSTN) to CA15-3 and CEA enhances breast cancer detection (Jia et al. 2022 ; Zhou et al. 2024 ). Tumour cells also release mutated or overexpressed proteins into the blood, sometimes in EVs. EX proteins, including mutant EGFRvIII (glioblastoma), glypican-1 (pancreatic cancer), and EV-associated HER2/EGFR (breast and lung cancers), have been identified in patients (Huang et al. 2022a , b ). Even extracellular matrix enzymes, such as metalloproteases, and cytokines released by tumours may serve as indirect markers of tumour activity (Huang et al. 2022a , b ; Unver and McAllister 2018 ). miRNA-associated results exhibit considerable variability in the literature. This variability can be attributed to the methodologies, quantification, and result normalisation used in the miRNA isolation and detection protocols. The lack of a clinically established endogenous control for quantifying miRNA levels largely hampers relative quantification analysis of miRNAs and thereby undermines its prospects for clinical reproducibility (Zendjabil 2024 ). Regarding tumour proteins, the major issue is that they are not cancer-specific. For example, PSA is a marker for prostate cancer, but PSA levels can also be elevated in conditions like prostatitis. Other factors, such as age, race, body mass index, and medication history, can also influence PSA levels. CA-125 is a marker for ovarian cancer, but its levels can also be elevated in situations like pregnancy and menstruation. These factors lead to the generation of false positives in the diagnosis of cancer (Connal et al. 2023 ). Circulating miRNAs in cancer show potential as CTMs. miRNA signatures have been observed in cancers originating from tissues such as the lung and liver. Compared with cancer-specific proteins, miRNAs show greater sensitivity for detecting early cancer stages. For example, the panel of miR-21, miR-210, and miR-155 has demonstrated greater sensitivity in detecting early-stage lung cancer than the protein marker CEA (Mitchell et al. 2008 ). In hepatocellular carcinoma, the miR-122 and miR-21 panel showed greater sensitivity for detecting early-stage hepatocellular carcinoma than carcinoembryonic antigen (Alemayehu et al. 2024 ). miRNAs can be a good measure of cancer patients’ response to therapy and can provide real-time feedback. However, the lack of large clinical studies has hampered the implementation of miRNAs in clinical settings (Petrou and Ladame 2022 ). In general, circulating TAPs tend to be less sensitive and specific than nucleic acid markers, but they complement other CTMs by reflecting different aspects of tumour physiology (Tan et al. 2025 ). Modern biosensors can specifically target these proteins using antibodies or aptamers. At the same time, assays can identify tumour-derived proteins on isolated cells or vesicles (e.g., immunostaining for PD-L1) to yield supplementary functional data (Wu et al. 2021 ). An emerging liquid biopsy biomarker is tumour-educated platelets (TEPs), which are platelets whose RNA profiles alter when they come into contact withtumours (Liefaard et al. 2023 ; Xiao et al. 2022 ). In early stages of cancer, 30–70% of platelets can be associated with tumour, with the number elevating to 90% in metastatic stages of cancer (Best et al. 2015 ).TEPs contain tumour mRNAs and have shown remarkable potential for disease diagnosis. A pan-cancer classifier trained on TEP mRNA profiles achieved 98.8% accuracy (AUC 0.999) during training and 95.4% accuracy (AUC 0.972) during validation (Xiao et al. 2022 ). In renal cell carcinoma, a 68-gene TEP signature demonstrated 100% accuracy in training (AUC 1.000) and approximately 95.9% overall accuracy (Xiao et al. 2022 ). Even though mutant transcripts are not directly detected in platelets, global TEP mRNA profiles can distinguish between patients with KRAS-mutant tumours and EGFR-mutant tumours, as well as between HER2-amplified and triple-negative breast cancers (Best et al. 2018 ). This indicates that TEPs encompass broad tumour-associated transcriptional alterations via specific splicing activity and the assimilation of tumour RNAs (Best et al. 2018 ; Liefaard et al. 2023 ). The ability of the cancer environment to influence TEP RNA expression enables the use of TEPs to understand the tumour microenvironment. At the clinical level, TEPs can be studied for their RNA profiles, as they are expected to differ between healthy individuals and cancer patients. TEP RNA profiles can also vary by cancer type and stage, which can be useful for determining disease progression (Kuznetsov et al. 2012 ). For example, TEPs were able to locate the primary tumour origin with 72% accuracy in cases of cancers originating in the lung, liver, breast, colon, and pancreas. TEP mRNA analysis also revealed driver mutations such as MET or ERBB2 Positivity, mutant KRAS, EGFR, or PIK3CA in cancers. TEP mRNA analysis achieved 96% accuracy in detecting metastasis in cancer patients and distinguishing them from healthy individuals. Thus, TEP mRNA profiling can yield real-time information on a cancer patient’s metastatic status and help determine the treatment course based on the driver mutations identified (Best et al. 2015 ). Although TEPs hold promise as CTMs, certain challenges remain. A major hurdle is the lack of a clinically established, standardised method for isolating TEPs. These differences in methods ultimately result in variations in results and interpretations (Best et al. 2015 ). Cohort studies have shown that non-cancer-related factors, such as associated diseases, medications, age, and inflammation, can influence the TEP mRNA profile. All of these can cause variations between individuals, making it difficult to establish a baseline value for mRNA expression levels related to cancer. Additionally, isolation of mRNA from TEP and their genetic analysis is a highly technical process and thus will have a limited base of end users (Sol and Wurdinger 2017 ). TEPs have shown potential in cohort studies for cancer detection. As such, they can have clinical applications. TEP has potential as a minimally invasive or noninvasive method for cancer detection. TEP-derived mRNA profiling has been shown to have diagnostic specificity, distinguishing cancer patients from healthy individuals and determining tumour origin (Varkey and Nicolaides 2021 ). Studies in ovarian cancer have shown that TEP mRNA profiling has better diagnostic performance than the biomarker protein CA-125 (Zhu et al. 2023 ). The further development of TEP profiling will involve integrating TEP mRNA profiling with automated quality controls and machine learning models to optimise TEP’s functionality as an established CTM (Sol and Wurdinger 2017 ). Combining different types of CTMs or biomarkers within the same class often enhances diagnostic performance. Multi-CTM analysis provides a better understanding of the tumour microenvironment and its complexity. The advantage of multi-CTM analysis lies in the ability of individual CTMs to highlight specific aspects of cancer development. CTCs can provide information on the genetic and protein make up of the cells, ctDNA can highlight the mutations in the major driver genes of cancer, exosomes can provide genetic, transcriptomic and proteomic data, miRNAs can track the progress of cancer and its response to therapy, proteins often provide existing clinical signatures while TEPs can provide information on the tissue of origin of cancer and its subtypes (Kuligina et al. 2024 ; Wang et al. 2024a , b , c ). Several studies have shown that multi-CTM analysis yields better diagnostic efficiency than single-CTM analysis. For example, simultaneously identifying elevated CTC counts and increased serum CEA significantly increased the odds ratio for colorectal cancer recurrence diagnosis compared with using either marker in isolation (Xiao et al. 2022 ). Similar synergistic effects are observed when combining EX biomarkers with traditional markers or integrating ctDNA analysis with imaging (Wang et al. 2022a , b , c ; Zhou et al. 2024 ). Researchers are actively developing multi-analyte liquid biopsy panels to make them as robust as possible. Despite progress, challenges remain, such as CTMs often being low in abundance in early-stage disease, and heterogeneity and temporal instability complicate detection (Connal et al. 2023 ; Deng et al. 2022 ; Liu et al. 2021 ), while the lack of standardisation of collection and analysis increases variability. Future directions include integrating machine learning with nanobiosensors and signal amplification techniques to enhance sensitivity and support patient stratification(de Abreu et al. 2025 ; Ma et al. 2024 ). With continued innovation in multi-marker assays, CTM-based liquid biopsy holds promise for precision cancer diagnostics.A comparative summary of the attributes of the different CTMs is presented in Table  1 .

The major step in early cancer detection is conducting regular screenings of both vulnerable and healthy populations. Screening can detect cancers before they become metastatic. Metastasis is a major cause of tumour recurrence and treatment failure. Tumour biomarkers are critical for the regular screening of populations for Cancer. Tumour biomarkers are obtained from blood and urine, preferably. Further tests can be done on cerebrospinal fluids, followed by solid tissue biopsies (Schiffman et al. 2015 ). Biomarkers can be proteins; for example, Prostate Specific Antigen (PSA) for the detection of prostate cancer and Cancer antigen 125 (CA-125) for ovarian Cancer. Mutations in genes such as the Breast Cancer Susceptibility gene (BRCA) 1/2 and the Epidermal Growth Factor Receptor (EGFR), in the case of breast and lung cancer, respectively, can serve as biomarkers. Circulating tumour DNA (ctDNA), metabolites, and antibodies against pathogens such as human papilloma virus are candidates for early cancer detection (Tenchov et al. 2024 ). Below, we discuss some of the traditional biomarkers for cancer detection. Prostate cancer detection has remained reliant on measuring PSA in tissue and fluid samples from individuals. PSA levels are significantly elevated in prostate cancer patients. PSA can also give information on disease progression and response to therapy. Elevation of PSA is attributed to disruptions in the epithelial tissues of the prostate, which are responsible for producing the PSA glycoprotein (Sultan et al. 2025 ). Recent studies have focused on improving the sensitivity of PSA-based prostate cancer detection by combining PSA-based assays with mpMRI and analysing additional markers such as PTEN, PMSA, and urinary zinc (Stopka-Farooqui et al. 2025 ; Amparore et al. 2025 ). CA-125 is one of the most widely used biomarkers in the detection of ovarian cancer. Serum levels of CA-125 generally increase with cancer stage and progression. CA-125 can be used both as a diagnostic marker for ovarian cancer detection and for monitoring response to therapy and disease-free survival (Lee et al. 2016 ). To improve CA-125’s diagnostic capability for predicting ovarian cancer, it is being combined with other biomarkers, such as HE4 protein, and digital imaging technologies. Additionally, machine learning and artificial intelligence have been utilised in cohort studies to analyse biomarker-based assays involving CA-125, demonstrating significantly better diagnostic and therapeutic evaluation of ovarian cancer (Xu et al. 2025 ). The tumour suppressor genes BRCA1 and BRCA2 are critical in maintaining genetic stability. The occurrence of pathogenic mutations in these two genes has been associated with a 70% increase in the risk of developing breast cancer and 40% risk of developing ovarian cancer. Investigation of the mutational status of BRCA1 and BRCA2 can predict the development of breast cancer, allow clinical planning in risk reduction and choosing the treatment protocol in breast cancer diagnosed patients (Dossa et al. 2025 ). To improve the diagnosis of breast cancer, the investigation of the mutational status of BRCA1/2 is combined with other marker proteins, such as TPS, CA153, and CEA. Additionally, genomics studies and integration of clinical and genetic data with artificial intelligence-based applications are being studied to provide the best prediction, prognosis, therapy and monitoring the response to therapy (Ahn et al. 2023 ). EGFR is a transmembrane protein that promotes cell proliferation and survival. In the diagnosis of non-small cell lung cancer, the EGFR protein has proven to be a clinically significant marker for predicting and detecting the disease. Deletion of exon 19 and substitutions in exon 21 of the EGFR gene have been observed in non-small cell lung cancer patients. The detection of EGFR mutations and their type allows the prediction of disease likelihood and the determination of the type of therapy for the individual patient with non-small cell lung cancer (Kim et al. 2022 ). Recent developments have shown that ctDNA obtained from body fluids can be utilised to carry out the detection of EGFR mutations and determine the type of mutation and the associated EGFR inhibitor that can be administered as therapy (O’Leary et al. 2020 ).

The menace of cancer has plagued the health scenario across the world. Although improvements in cancer therapeutics have occurred, overall survival rates depend on the stage at which these therapies are administered. Late administration of cancer therapeutic drugs does not yield the expected response in the patients. As such, the early detection of cancers is as critical in the response of cancer patients to cancer therapeutics as the therapy itself (Gromek et al. 2024 ). The importance of early cancer detection is evident in the fact that people whose cancer is detected at early stages have lower treatment morbidity, improved quality of life, and better care (Whitaker 2020 ). The 2024 report by the National Cancer Institute, USA, showed that early detection of cancers led to significantly better 5-year survival rates than detection at metastatic stages. For example, early detection of breast cancer allowed 99% of the patients to have a 5-year survival period as compared to a 32% percent 5-year survival rate in the case of patients whose cancers were detected late. In lung cancer, early detection is associated with a 64% survival rate, while late detection is associated with a 9% survival rate over 5 years. Similar trends were recorded against several other cancer types like liver, prostate, thyroid, uterine and oral cancers from the Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute (NCI), USA (Surveillance, Epidemiology, and End Results Program 2024 ). Treatment of cancers detected at early stages has the added advantage of simpler, less expensive therapeutic regimens than those for late-stage cancers. In late-stage cancer, treatment must be multi-drug to target the tissue of origin and metastatic cancer cells in other organs, which adds complications and increases treatment costs. At the same time, the quality of life is less hampered for patients whose cancers are detected early than for those whose cancers are detected late (Aguiar-Ibáñez et al. 2025 ). In India, the treatment cost of early-detected oral cancers was USD 1528, while that of late-stage oral cancers was USD 2717. On average, there is a 17% increase in the treatment cost as the oral cancer moves from stage I to stage IV of disease progression. The economic cost of oral cancer treatment is projected to reach USD 3 billion in India if regular screening and early detection are not carried out extensively. Early detection can save India USD 30 million annually by enabling cancer care to be delivered at earlier stages. Thus, early detection of cancers is critical from a healthcare and economic perspective, especially for growing economies like India (Singh et al. 2021 ).

The identification of CTMs at extremely low concentrations necessitates biosensors that integrate effective signal transduction, strong biomolecular recognition, and compatibility with biofluids. Designing nanobiosensors for CTM detection integrates the functioning of three major components : (i) nanomaterial surface engineering for signal enhancement, (ii) biorecognition interfaces based on antibodies, aptamers, or ligands for selective target capture, and (iii) catalytic signal amplification strategies using nanozymes or plasmonic enhancement systems (Fig.  4 ) (Quesada-González et al. 2018 ). Fig. 4 Design components of nanobiosensors aimed at signal recognition and amplification Design components of nanobiosensors aimed at signal recognition and amplification The surface of nanomaterials can be modified with chemical groups to enhance CTM capture and signal readout. Functionalization increases the effective surface area, providing abundant anchor sites for bioreceptors and thereby amplifying sensitivity. For instance, the nanomaterial MIL-156 MOF@COF, a hybrid metal-organic framework (MOF) with a covalent organic framework (COF), was decorated with gold (Au) nanoparticles that offered two advantages: the porous MOF–COF shell provided a high surface area and functional groups. At the same time, the AuNPs served as efficient anchors for anti-CA15-3 antibodies via amine coupling (Liu et al. 2025a , b ). This miniature editing further revealed a well-defined response to the cancer antigen (30–100 nU/mL) and a very low limit of detection (∼2.6 nU/mL) (Liu et al. 2025a , b ). Graphene nanomaterials and carbon nanotubes are other nanomaterials with a large surface-to-volume ratio and are also equipped with dense probe loading (Prabhu and Liu 2025 ). Similarly, thiolated aptamers or proteins are attached via Au–S bonds or via carbodiimide (EDC/NHS) chemistry to create robust sensing interfaces (Wang et al. 2025b ). Functionalization of the nanomaterial also stimulates the signal. For instance, gold films and gold nanoparticles with surface plasmonic resonance (SPR)activities can greatly enhance optical signals through metal–metal coupling. In one SPR sensor, attaching gold nanoparticle (AuNP) tags to an aptamer-coated Au surface resulted in a 10,000-fold increase in sensitivity over conventional ELISA (Prabhu and Liu 2025 ). While surface functionalization is appealing, it remains challenging. Uncontrolled adsorption or aggregation is likely. Further, non-specific binding must be blocked, which can be achieved with blocking agents. Nevertheless, with cautious control towards maintaining reproducibility and reduced noise, tailored nanomaterial coatings provide the basis for high-sensitivity CTM biosensors with maximized probe loading and electron transfer (Liu et al. 2025a , b ; Wang et al. 2025b ). Although aptamers and antibodies improve analytical sensitivity through highly specific target recognition and increased capture efficiency, they primarily function as biorecognition elements rather than as direct signal amplification mechanisms. To capture tumour antigens, the surface of nanomaterials (biosensors) is functionalized with biorecognition elements, which are generally antibodies, aptamers, or other ligands (Wang et al. 2025b ). Antibodies have long been the gold standard: their multi-domain 3D structure (Fab regions, complementarity-determining regions) confers very high affinity and specificity for target antigens (Prabhu and Liu 2025 ). Antibodies as biosensing agents on the surface yield excellent analytical accuracy, even at low analyte concentrations. However, these systems have notable drawbacks: producing high-quality antibodies is expensive and time-consuming. Further, antibodies may get denatured or vary between batches. Their large size also limits how densely they can pack on a surface, which can slow binding kinetics and reduce signal output (Atabay et al. 2025 ; Bohunicky and Mousa 2010 ; Lin et al. 2021a , b ; Prabhu and Liu 2025 ). Aptamers, which are often considered “chemical antibodies,” are short single-stranded DNA/RNA sequences selected in vitro for tight target binding (Prabhu and Liu 2025 ; Wang et al. 2024a , b ). They are with several positive sides: (a) ease of synthesis at large scale without the requirement of animals as production host, (b) stability in a wide pH and temperature range, (c) nano-size enabled high surface density and fast diffusion, and (d) ease of modification for enhanced conjugation with other sensors (Hassan et al. 2016 ; Prabhu and Liu 2025 ). Furthermore, aptamers allow for creative designs that can be used to display conformational changes upon target binding to modulate signals (Prabhu and Liu 2025 ). Nevertheless, use of nanoparticle conjugation with aptamers as a biosensing agent also has challenges: they are not as specific as antibodies and truly high-affinity aptamers are difficult to isolate or synthesize (Wei et al. 2022 ). Antibody- or aptamer-conjugated nanobiosensors have been ultra-sensitive in detecting circulating tumours or tumour antigens of various types (Wu et al. 2015 ). For instance, 6-carboxyfluorescein-labelled aptamers (FAM-apt) (FAM-labelled aptamer) adsorbed Iron doped porous carbon nanozymes have been investigated for the detection of prostate-specific antigens, wherein the bound aptamer enhanced the nanozyme’s mediated oxidation of the chromogenic substrate Tetramethylbenzidine (TMB). In contrast, the aptamer’s own green fluorescence was quenched. In the presence of PSA, the aptamers dissociate from the nanozyme and exhibit their own fluorescence, thereby facilitating PSA detection (Liu et al. 2025a , b ). In another illustration, a polydopamine-functionalized gold nanoparticle (Au@PDA-NP) functionalized with a capture aptamer was used for the detection of exosomal proteins in hepatocarcinoma, with surface plasmon resonance (SPR)- based detection employed. (Liao et al. 2020 ). These strategies show how aptamers tethered to nanostructures enable multimodal, amplified sensing (Prabhu and Liu 2025 ). Antibody-based biosensing nanomaterials are also common. For instance, Manganese oxide–mesoporous silica nanoparticles conjugated to PSA-specific antibodies can detect prostate cancer via magnetic resonance imaging (MRI) (Du et al. 2020 ). Anti-Glypican-1 antibody-conjugated gold nanocages modified with hyaluronic acid can be employed for MRI-based sensing of pancreatic cancer (Popovtzer et al. 2008 ). Similarly, an immunosensor for anti-CA15-3 antibodies (Cancer antigen 15 − 3, a breast cancer marker) on an AuNP/MOF surface via Au–amine linkages can be used for the detection of breast cancer (Liu et al. 2025a , b ). In all cases, the ligand’s orientation and density on the nanomaterial influence binding kinetics and sensitivity. While both aptamer- and antibody-functionalized nanobiosensors are important and popular tools for theragnostic applications in cancer, they have their own advantages and limitations. While antibodies provide specificity and clinical validation, aptamers tend to be stable and multiplex. Both approaches can detect tumour antigens at low levels when coupled with signal-amplifying agents. However, antibodies may reveal batch variability and activity loss at times, and aptamers may require protection in complex environments. The selection of a suitable, optimal antibody or aptamer (on the grounds of cost, ease of labelling, required stability, and regulatory factors) and subsequent conjugation with nanomaterials provides a strong and reliable biosensing agent (Prabhu and Liu 2025 ). Nanozymes are nanoparticles that can act as enzymes and provide powerful signal amplification when detected by sensors. Nanozymes, when combined with aptamers, antibodies, or other ligands, reveal enhanced analytical performance. Generally, towards efficient detection of cancer/tumour biomarkers such as CTCs, EXs, and CEAs, two types of sensing strategies- single-mode and dual-mode sensing are considered, wherein nanozymes primarily act as a probing agent (Wang et al. 2025a ). Nanozymes may catalyse reactions similar to those of peroxidases, oxidases, and catalases and are accordingly categorised as oxidase, peroxidase, hydrolase, superoxide dismutase, and catalase mimics (Wang et al. 2025a ). Metal oxide nanoparticles, such as CuO, which mimic both catalase and peroxidase, can reduce H₂O₂, thereby enhancing the performance of electrochemical sensors for the detection of CTCs in breast cancer (Tian et al. 2018 ). Other metal oxide nanoparticles, such as Fe₃O₄, MnO₂, and CeO₂, exhibit peroxidase-like activity and can act on H₂O₂ to generate radicals that oxidise chromogenic substrates such as TMB (Wang et al. 2025a ). In another design, a Fe–N–C nanozyme served as an oxidase mimic: it oxidised TMB directly (without added H₂O₂) upon stimulation by a tethered aptamer (Liu et al. 2025a , b ). An example of a “dual nanozyme” sensor is the combination of Cu-HHTP-MOF nanosheets (2,3,6,7,10,11-hexahydroxytriphenylene-metal-organic framework nanosheets rich in Cu sites) and gold nanoparticles (AuNPs). In this system, Cu-HHTP provides abundant catalytic Cu–O₄ centres for H₂O₂ oxidation, and AuNPs enhance electrical connectivity. This Cu-HHTP/Au system delivered an extremely high catalytic current for H₂O₂, enabling H₂O₂ sensing down to ~ 5.6 nM for the detection of colon cancer cells (Huang et al. 2022a , b ). Unlike proteins, nanozymes do not denature easily, offering excellent stability. The size and composition of nanozymes can be readily modulated. They are less costly and provide scope to be integrated with other amplifiers and detection technologies such as photochemical, electrochemical, fluorescence, and colourimetric methods(Wang et al. 2022a , b , 2025a ; Das et al. 2021 ; Jiang et al. 2023 ; Li et al. 2019 , 2023 ). However, nanozymes often lack the high specificity of natural enzymes: they may act on any suitable substrate, leading to potential background signals if interfering species (endogenous peroxides, reductants) are present. Further, they are dependent on pH and the requirement of co-substrates (Wang et al. 2025a , b ). Nevertheless, when well-integrated, for example, coupled to an aptamer or an antibody (for specific recognition), nanozymes have enabled ultrasensitive CTM detection, possibly at femtomolar or even lower LODs through cascade signal amplification (Wang et al. 2025a ).