In general, technologies for identifying CTCs involve three critical steps: Capture and enrichment, detection and identification, and release for further analysis (30). As noted by Wang et al (31), capture and enrichment methods rely primarily on the interaction between specific materials and certain physical, chemical or biological properties of CTCs. After collection, various identification methods can be used, such as flow cytometry, fluorescence microscopy and DNA sequencing. Released CTCs are typically utilized for downstream analyses, such as genomics, transcriptomics, proteomics, and immunofluorescence assays of cultured cells (31). Enrichment methods for CTCs are based on certain biological and physical properties (32,33). Since most tumors originate from epithelial tissues, epithelial cell adhesion molecules (EpCAMs) are the most commonly used biomarkers for CTCs. Physically, CTCs can be differentiated by properties such as size, density, charge, and deformability (34–37). In recent years, detection techniques that integrate both the physical and biological properties of CTCs have emerged, offering enhanced efficiency and sensitivity (38,39) (Fig. 2).

In addition to size-based enrichment methods, density-based enrichment is a workable strategy with distinct advantages. One commonly used density gradient centrifugation technique involves the use of a sucrose polymer with a high synthetic molecular weight (Ficoll-Paque; GE Healthcare) as the primary density gradient medium (49). The high centrifugal force applied in this method creates a density gradient and the size and density of the cells determines their sedimentation rate. The cell loading and harvesting procedures must be determined empirically as they are significantly influenced by the Ficoll-Paque, the rotor, and the type of centrifuge tube used. Excessive cell loss and low separation efficiency for CTCs isolated by Ficoll-based density gradient centrifugation have been attributed to density-related toxicity (38,50). The OncoQuick gradient system addresses these problems by minimizing cell loss and contamination, thereby achieving higher CTC enrichment rates than the traditional Ficoll-Paque technology does (38,51).

The dielectrophoretic array (DEPArray) system, developed in 2013, is a semiautomated platform capable of selecting and isolating individual CTCs and small populations of cells (52). The primary advantage of the DEPArray system lies in its ability to identify multiparameter immunofluorescence staining characteristics of individual cells before isolation. Additionally, it can be integrated into a continuous, semiautomated workflow platform, such as with the CellSearch System. However, the system's advanced capabilities for imaging and isolating individual CTCs have significant drawbacks, including high costs and an average cell loss of ~40% during sample preparation and separation (52). To address these challenges and achieve higher separation resolution, researchers developed an integrated dielectrophoretic (DEP) and magnetophoretic (MAP) microfluidic chip in 2022. This integrated MAP-DEP approach enables the separation of various types of CTCs on the basis of their size and dielectric properties (53).

In 2019, an investigator introduced a novel multiflow microfluidic (MFM) system designed for the high-purity separation of CTCs. By utilizing the inherent differences in the inertial migration of cells flowing within microchannels, the MFM system demonstrated a sensitivity that achieved over 87% purity in label-free separations. However, the system has serious limitations. Its sensitivity and recovery rates are compromised when CTCs exhibit significant size overlap with other cell types. Additionally, compared with other label-free separation systems, the device operates at a relatively low flow rate. To address these limitations, a new class of collectors based on similar principles, known as inertial microfluidic labyrinth devices, was developed (54). These devices incorporate 12 distinct arrays of microfilters and combine inertial focusing with Dean flow dynamics. This approach achieves size-based cell separation by continuously focusing all cells while selectively isolating CTCs from smaller blood cells at the device's outlet. Current CTC and CTC cluster separation techniques often rely on biomarker-dependent antibody-based capture. However, compared with individual CTCs, CTC clusters pose unique challenges: They have a lower surface-area-to-volume ratio (55) and often exhibit mixed epithelial-mesenchymal phenotypes, resulting in reduced antibody binding regions (56). Consequently, these methods may result in lower enrichment rates for CTCs and CTC clusters with stem-like or mesenchymal characteristics. By contrast, physical property-based separation methods, such as inertial microfluidic labyrinth devices, avoid the biases inherent in molecular marker-dependent systems. These advanced devices offer a more effective solution for separating CTCs and CTC clusters, including those with diverse phenotypes, because they rely on size-based rather than molecular-based characteristics (54).

In summary, CTC enrichment methods based on physical properties are label-free and relatively simple to perform. However, they often suffer from limited enrichment purity and a risk of cell loss. As a result, they are best suited for preliminary enrichment, and their combination with immunoaffinity methods may offer a more effective strategy to increase detection sensitivity and specificity.

CTCs are typically enriched through positive selection using antibodies targeting tumor-associated antigens. These antigens include epithelial cell adhesion molecule (EpCAM), mesenchymal markers, cytokeratins (CKs), mucin-1, human epidermal growth factor receptor 2 (HER2) and epithelial growth factor receptor (EGFR), among others (57,58). Alternatively, CTCs can be enriched through negative selection by removing unwanted leukocytes using antibodies against markers such as CD45; however, the EpCAM-dependent positive selection technique is the most widely used (59,60).

Currently, several methods are popular for the positive enrichment of CTCs. The CellSearch system, which uses an EpCAM-based immunoassay, is the first and only FDA-approved CTC assay (61) for colorectal, breast, prostate and other tumors (62). The CellSearch system's principle involves the use of ferromagnetic fluid nanoparticle-labeled antibodies that specifically bind to EpCAM, which is highly expressed on the surface of a wide range of tumor cells. CTCs are then separated from other blood cells by means of a magnet. The detailed collection process is illustrated in Fig. 3. However, EpCAM-antibody interactions can cause CTCs to adhere to the device's surface, making their release difficult. To address the shortcomings of the CellSearch system, researchers developed MagSweeper, an immunomagnetic cell separator that enriches CTCs while eliminating nonmagnetic particle-bound cells (63).

Advances in microfluidic chip technology and nanomaterials have further improved the sensitivity, specificity, and cost efficiency of CTC detection (32). A key feature of microfluidic chips is their flexible combination and large-scale integration of multiple functional units on a miniature platform. These chips represent an interdisciplinary achievement combining physics, chemistry, material science, medical science, mechanics, optics, mechanical engineering and bioengineering. Microfluidic devices utilizing antigen-antibody interactions can be categorized into two types: Microcolumns (requiring chemical modification of the monoclonal antibody coating) and devices using antibody-coated surfaces (38). The CTC microarray isolates target cells by enabling the interaction of CTCs with an array of 78,000 microcolumns coated with EpCAM antibodies under precisely controlled laminar flow conditions. This method does not require pretreatment of the samples and preserves cellular activity (22).

The size-determined immunocapture chip, a deterministic lateral displacement (DLD) microarray, captures CTCs with high sensitivity, efficiency and purity. This chip features microcolumns that are surface modified with EpCAM antibodies. Relatively large CTCs are deflected laterally as they collide with microposts, prolonging their flow time, while smaller blood cells advance with the laminar flow (64) and selectively increase the probability of interaction with the microposts. This method achieves capture efficiencies exceeding 92%, with a purity of 82% (65,66). Another approach employs microfluidic encapsulated microarrays with a staggered herringbone design on channel ceilings, increasing cell-surface contact points. This design significantly enhances capture efficiency while preserving cellular activity, enabling downstream analyses such as cell culture and genomic studies. However, CTCs undergoing epithelial-mesenchymal transition (EMT) exhibit increased invasive and metastatic potential (67). During the EMT, the expression of tumor markers on the cell surface is downregulated (68,69), posing significant challenges for devices that rely exclusively on anti-EpCAM antibodies for CTC capture (70). A synergistic, dual-antibody chip based on the DLD principle partially addresses this challenge. In this approach, anti-ASGPR and anti-EpCAM antibodies are simultaneously modified in parallel on a microfluidic synergistic chip, with a capture efficiency of more than 87% per channel (71).

Over time, advancements in research have transitioned CTC capture technology from first-generation chips, such as the CTC-Chip, to second-generation chips, such as the HB-Chip, GO-Chip and GEM-Chip. Currently, a third-generation microfluidic chip, called the CTC-iChip, has emerged as a recognized innovation in the field (72–74). Compared with first-generation CTC microarrays, second-generation microarray technology, specifically surface capture devices, is simpler to fabricate and more efficient at capturing CTCs. However, these devices have serious limitations, as the captured CTCs are immobilized on the device surface, making their recovery for downstream genomic analyses challenging. While trypsin digestion can be used to release the captured cells, this process may remove some receptor proteins from the surface of the CTCs and interfere with subsequent analyses.

Rostami et al (75) proposed overcoming this limitation by developing a novel microfluidic chip incorporating smaller magnetic pores and utilizing aptamers as an alternative to conventional antibody-based CTC capture. Aptamers provide several advantages, including high binding affinity, low cost, facile chemical modification, and straightforward release mechanisms (75). To increase the capture rate of specific CTCs, researchers have developed novel capture platforms with specially designed surfaces conjugated with engineered capture molecules. A representative example is the TDN-nanogold click chip developed by Sun et al (76), which features a gold-modified surface functionalized with DNA nanostructures. This advanced design, combined with a microfluidic chamber, enhances specific CTC capture. Traditional methods often have difficulty capturing CTCs that have undergone EMT because of the consequent reduction in the expression of common capture targets such as EpCAM. However, the TDN-Nanogold Click Chip uses click chemistry (77) to effectively and irreversibly capture CTCs with a capture rate as high as 91.3±6.7% (76).

Third-generation microfluidic chip technology, such as the inertial focusing-enhanced microfluidic system (CTC-iChip), enables the efficient negative depletion of anuclear RBCs and platelets. This is achieved through the hydrodynamic separation of nucleated leukocytes and CTCs in plasma, followed by inertial focusing into a single stream. Magnetic bead-labeled leukocytes are then efficiently removed by magnetophoresis, resulting in a solution enriched with CTCs with both epithelial and mesenchymal features (78). However, this system has notable drawbacks, including high initial cost a long setup time, limited ability to analyze single-cell molecules, reliance on multiple manually attached chips, and challenges in clinical implementation (32).

To reduce the cost of capturing CTCs, a recent study proposed the use of EpCAM-resistant nanofiber cellulose scaffolds with immobilized Fe₃O₄ NPs for harvesting CTCs (25). Cellulose nanostructures can be categorized into two main types: Cellulose nanocrystals and cellulose nanofibers (CNFs). These nanomaterials offer several advantages, including relatively high aspect ratios, large specific surface areas, biodegradability, environmental sustainability, biocompatibility and non-toxicity (25). Compared with fibrous scaffolds, rigid nanocrystal scaffolds interact more efficiently with CTCs (25); however, CNFs have a higher aspect ratio, which increases the probability of intermolecular interactions. Kumar et al (79) developed a CNF-based microfluidic chip with a capture efficiency of 79–98%. However, its average flow rate of 5 µl/min poses a significant limitation for processing large volumes of blood, as it reduces the point of contact necessary for effective CTC capture (79). By contrast, a microchip developed by Cui et al (80) and integrated with ZnO nanowires achieved a high CTC capture efficiency of 91.1±5.5% while maintaining excellent cell viability (96%). Notably, the platform also enabled damage-free release of captured cells. This advancement in microarray technology facilitates single-cell analysis and holds considerable potential for advancing personalized medicine, particularly in the development of therapeutics targeting cancer stem cells (80). Researchers have also recently developed a nanoplatform called NICHE for determining the relationships between host genes and the microenvironment in living cells (81). This innovative platform combines microfluidics and nanofluidics to efficiently capture CTCs from blood samples. The platform then integrates stable multichannel fluorescent probes into live CTCs, enabling rapid in situ quantification of target gene expression. The NICHE microfluidic device is unique in that it allows simultaneous gene expression and phenotypic analyses on individual cells in situ. It also has the potential to generate predictive indices for screening patients who are likely to benefit from ICIs.

The decreased expression of EpCAM on the surface of cells undergoing EMT results in a reduced capture rate when an EpCAM-dependent positive enrichment method is used (38). To overcome these shortcomings, label-free and negative enrichment methods have been developed. Similar to the previously mentioned CTC-iChip, a 3D microfluidic device has been developed and fabricated for the negative enrichment of CTCs from whole blood samples (82). The device captures unlabeled leukocytes on its inner surface using antibodies targeting leukocyte-specific membrane antigens and separates red blood cells and platelets from CTCs according to their small size. This method is compatible with subsequent molecular, immunocytochemical and functional analyses.

In conclusion, positive enrichment methods offer superior specificity and purity, making them effective for enriching CTCs with known markers. However, their reliance on specific markers may cause them to miss certain CTC subtypes. Conversely, negative enrichment methods can capture a more diverse range of CTCs but generally result in lower enrichment purity and a risk of cell loss. In simple terms, positive enrichment methods are best suited for applications requiring high sensitivity and specificity, whereas negative enrichment methods are more appropriate for capturing a heterogeneous CTC population, particularly when markers are unknown or inconsistently expressed. A combined approach may provide a more thorough CTC detection strategy, despite being more complex to implement. In addition to the above-described platforms, numerous potential CTC enrichment methods have been developed in recent years (Table SI).

This technique involves labeling enriched cells after fixation using fluorescent markers such as anti-human pancytokeratin (CK), anti-mouse CD45, DAPI, and a cocktail of antibodies targeting cell surface proteins such as EpCAM, EGFR and HER2. CTCs are identified as DAPI⁺/CK⁺/CD45⁻ cells (83). The primary advantage of this method is its convenience in observing the phenotype and morphology of cells under a fluorescence microscope (84). However, the expression of protein markers often lacks stability and cannot be used to distinguish between viable and non-viable cells (83,85).

RT-PCR can be used to detect target cells by reverse transcription, in which the enzyme reverse transcriptase converts RNA into complementary DNA using a specific primer (86). While this method allows for the genetic identification of target cells, it cannot distinguish non-viable cells from viable cells, limiting its application in analyses requiring live cells, such as assessments of cellular deformability and drug responses.

This method enables the detection of specific proteins secreted by CTCs. It involves culturing enriched CTCs and tagging secreted proteins with fluorescently labeled antibodies. Each immunofluorescent spot represents a living cell, offering high specificity and stability (87). However, this technique requires sufficient accumulation of secreted proteins to form immune spots, which can be time-consuming.

FISH can be used to detect CTCs by labeling abnormal chromosomal numbers or structures specific to tumor cells. This genetic method offers high sensitivity and specificity for identifying CTCs but it can suffer interference from abnormal chromosomes in the patient's blood and the limited scope of chromosomal probes, which can detect only a limited number of specific markers (54,88).

NGS provides insights into genetic structures and expression states at the single-cell level, allowing for the genotyping of CTCs and the detection of tumor-specific mutations (89). While this method offers unparalleled precision, highly effective upstream enrichment techniques and stringent QC are needed to ensure the purity and abundance of CTCs (90) (Table SII).

Numerous studies have shown the usefulness of CTCs in clinical prognosis; however, their role in predicting the response to tumor immunotherapy has been relatively unexplored. Research has demonstrated that the presence of CTCs serves as an independent prognostic indicator of shorter overall survival (OS) in patients with various cancers, including non-small cell lung cancer (NSCLC) (97). Lin et al (98) reported a significant reduction in the number of CTCs in the peripheral blood of patients with stage IV NSCLC who were treated with NK cell immunotherapy on Days 7 and 30 of treatment. These findings suggest that CTCs may serve as valuable biomarkers for evaluating treatment efficacy (98). A different study revealed that the presence of CTCs was a predictor of poor durable response rates following immune checkpoint inhibitor (ICI) therapy. A durable response is defined as stable disease or a partial or complete response without disease progression at six months (99). The prognostic and predictive value of CTCs was further demonstrated in a phase 1b study using lysosomal virus immunotherapy, where a reduction in CTC counts was observed in baseline-positive patients who responded effectively to treatment. Moreover, a study by Castello et al (100) revealed that the mean CTC count was significantly lower in patients with NSCLC treated with ICIs than in untreated patients. The results of the aforementioned study also revealed that combining the mean CTC count with the median metabolic tumor volume was associated with PFS and OS after 8 weeks of ICI treatment (100). Importantly, CTC counts were identified as independent predictors of both PFS and OS (101,102). Gu et al (103) developed a risk model to predict prognosis on the basis of aberrantly methylated DNA in CTCs from patients with lung adenocarcinoma (LUAD). In the aforementioned study, patients were categorized into high-risk and low-risk groups on the basis of their median risk score. High-risk patients were associated with poor prognosis, whereas low-risk patients demonstrated hyperimmunocompetence, which correlated with favorable prognosis. Univariate and multivariate Cox regression analyses confirmed that the risk score was an independent prognostic factor for survival in patients with LUAD (103).

Antitumor immunotherapeutic strategies have emerged as promising approaches for the treatment of various cancers. The expression level of PD-L1 is a crucial indicator of the immune status of patients with cancer and a key marker for predicting the efficacy of immunotherapy (104). Studies have suggested that PD-L1 expression serves as a predictive biomarker for immunotherapy response in certain solid tumors, including NSCLC, melanoma and renal cell carcinoma (95,105,106). For example, in a study of immunotherapy for NSCLC, patients with PD-L1-negative CTCs at baseline, as well as at 3 and 6 months of nivolumab treatment, achieved favorable clinical benefits at the 6-month mark. By contrast, tumor progression occurred in patients with PD-L1-positive CTCs. These findings suggest that PD-L1 expression on CTCs could serve as a predictive marker for the immunotherapy response (107). Dall' Olio et al (108) reported that pretreatment with PD-L1-positive CTCs was associated with improved survival, whereas posttreatment with PD-L1-positive CTCs was correlated with worse survival in patients with advanced NSCLC receiving PD-L1/PD-1 inhibitors as second- or third-line therapy (108). In a phase 1 trial of the PD-1 inhibitor IBI308 for the treatment of gastrointestinal tumors, the presence of CTCs with high PD-L1 expression was associated with significantly better disease control rates. The abundance of PD-L1-positive CTCs at baseline was also found to predict PFS, suggesting that monitoring CTC dynamics could provide an early indication of therapeutic response.

The authors' hypothesis that CTCs can serve as predictive biomarkers of tumor immunotherapy efficacy was also supported by a study on metastatic genitourinary tumors. PD-L1-positive CTCs, elevated CTC counts, and specific CTC morphological subtypes were associated with shorter OS from immunotherapy for metastatic genitourinary cancers (109). Combining an assessment of CTC cluster abundance with single CTC counts significantly improved prognostic accuracy (110). Thus, CTCs not only are useful for assessing the prognosis of patients with cancer, but they also serve as important predictive biomarkers for monitoring the efficacy of immunotherapy and developing robust personalized treatment strategies. Most studies have focused on NSCLC, colorectal cancer, breast cancer, prostate cancer and melanoma; the predictive value of CTCs in other cancer types requires further exploration.

Tumor metastasis and recurrence remain the primary causes of cancer-related death and present difficult treatment challenges. The development and successful application of ICIs represent major breakthroughs in tumor therapy in recent years. While it is not yet clear whether CTCs are more susceptible to interventions targeting PD-L1 than cells in solid tumors are, activating the immune system by blocking PD-L1 expression on CTCs and reducing their numbers in peripheral circulation could decrease the likelihood of malignant tumor recurrence and metastasis, representing a promising new therapeutic approach (8,111). In addition to PD-L1, CD47 is recognized as a crucial immune checkpoint that is highly expressed on tumor cells. CD47 interacts with signal-regulated protein alpha on macrophages, transmitting inhibitory signals that suppress phagocytosis (112). Thus, PD-L1 acts as the key ‘don't find me’ signal for the adaptive immune system, whereas CD47 functions as the ‘don't eat me’ signal for the innate immune system and regulates the adaptive immune response (113,114). Targeting both PD-L1 and CD47 simultaneously with specific antibodies has been shown to be more effective at suppressing lung metastases than targeting PD-L1 or CD47 alone (114) (Fig. 4).

CTCs often express multiple immune checkpoints, including PD-L1 and CD47 (115). Lian et al (115) demonstrated a significant reduction in tumor growth and metastasis by using antibodies specific to PD-L1 and CD47 expressed on the surface of CTCs. Additionally, Yu et al (116) reported on O-TPNVs, a type of nanovesicle created by fusing TIGIT-expressing cell membranes with platelet-derived membranes (TPNVs) and encapsulating oxaliplatin (OXA). The platelet-derived membrane components enable O-TPNVs to specifically target postsurgical wounds and interact with CTCs. OXA not only directly eliminates residual tumor cells and CTCs, but it also induces immunogenic cell death and activates the immune system. Furthermore, O-TPNVs exhibit synergistic chemotherapeutic and immunotherapeutic effects, effectively preventing the recurrence and metastasis of triple-negative breast cancer (4T1) after surgery (117). In addition, clinical trials are actively exploring the mechanisms involved in blocking tumor immune escape through ICIs. For example, Jacot et al (118) evaluated the clinical and pathological relevance, as well as the prognostic value, of PD-L1-positive CTCs in a cohort of 72 patients with metastatic breast cancer (MBC) as part of a prospective clinical trial. PD-L1-positive CTCs were associated with survival outcomes in patients with MBC (Clinical Trial Registry: NCT02866149) (118). In conclusion, immunotherapeutic targeting of CTCs is likely a therapeutic strategy for preventing tumor metastasis and recurrence.