Demystifying EV heterogeneity: emerging microfluidic technologies for isolation and multiplexed profiling of extracellular vesicles

Abstract

Extracellular vesicles (EVs) are heterogeneous lipid containers carrying complex molecular cargoes, including proteins, nucleic acids, glycans, etc. These vesicles are closely associated with specific physiological characteristics, which makes them invaluable in the detection and monitoring of various diseases. However, traditional isolation methods are often labour-intensive, inefficient, and time-consuming. In addition, single biomarker analyses are no longer accurate enough to meet diagnostic needs. Routine isolation and molecular analysis of high-purity EVs in clinical applications is even more challenging. In this review, we discuss a promising solution, microfluidic-based techniques, that combine efficient isolation and multiplex detection of EVs, to further demystify EV heterogeneity. These microfluidic-based EV multiplexing platforms will hopefully facilitate development of liquid biopsies and offer promising opportunities for personalised therapy.

---

1. Introduction

The increasing incidence and mortality rates of cancer pose a serious challenge globally. Cancer is one of the leading causes of death in the world population, burdening both society and families. Due to factors such as lack of awareness of health, uneven distribution of medical resources, and insufficient cancer screening, many patients are diagnosed with cancer only at advanced stages, resulting in compromised treatment efficacy.¹ Therefore, early diagnosis and dynamic monitoring of cancer are crucial for patient treatment and prognosis. Personalized precision medicine represents a new trend in cancer treatment, offering tailored treatment plans to patients.^2,3 However, meeting the requirements of precision medicine, such as minimizing iatrogenic damage, reducing medical costs, and systematizing pathological research, poses great challenges. Traditional tissue biopsies have limitations such as sampling difficulties, invasiveness, incomplete representation, and high risks of complications. Existing methods like blood biomarker detection and imaging exams suffer from insufficient sensitivity and specificity, leading to false-positive or false-negative results that impact diagnostic accuracy. These shortcomings make it difficult for these methods to serve as efficient and accurate detection tools meeting the needs of clinical precision medicine.⁴

In recent years, the emergence of non-invasive diagnostic techniques for cancer, specifically liquid biopsies, marks a significant advancement in the pursuit of precise cancer treatment. Liquid biopsy is a novel cancer diagnostic technology for detection of circulating tumour cells (CTCs), cell-free DNA (cfDNA), extracellular vesicles (EVs), and other biomarkers in bodily fluids to diagnose and monitor cancer, representing the forefront of malignant tumour diagnostics.^5–7 EVs, as an important clinical biomarker in liquid biopsies, hold important guiding significance in unravelling cancer mechanisms, cancer diagnosis, and postoperative assessment. Compared to traditional tissue biopsy methods, EV-based liquid biopsy technology offers advantages such as comprehensive tumour molecular information, minimal invasiveness, easy sampling, lack of radioactive contamination, and low cost. It has become one of the most promising non-invasive diagnostic and real-time monitoring methods for cancer, accelerating the realization of precise cancer treatment.^8–10

EVs are diverse entities which encompass nanoscale vesicles released by various cells. EVs are found in a wide range of bodily fluids, such as blood, saliva, cerebrospinal fluid, bile, and breast milk, functioning as “couriers” involved in material exchange and local, as well as long-distance communication, in physiological and pathological processes.^11–14 As shown in Fig. 1, EVs are typically categorized into exosomes (40–160 nm) and microvesicles (100–1000 nm) based on their sizes. Exosomes mainly originate from endosomal multivesicular bodies (MVBs), which fuse with the plasma membrane to release exosomes into the extracellular environment.

Fig. 1 EV biogenesis and heterogeneity. The two major categories of EVs are microvesicles and exosomes. Microvesicles are released through plasma membrane budding and are in the size range of ∼100 nm to 1000 nm. Exosomes originate from the endosomal pathway by the formation of the endosomal multivesicular bodies (MVBs). When MVBs fuse with the plasma membrane, exosomes are released (size range ∼40 to 160 nm). EVs can be a highly heterogeneous population in terms of size and molecular composition, including proteins, nucleic acids and glycans.

The heterogeneity of EVs extends beyond their size to include the cargo they carry. This variation in size and content poses challenges for researching the diverse biological functions of EVs. Studies have indicated that heterogeneous EV populations, laden with distinct cargos, may significantly influence both local and systemic transmission of EV-mediated phenotypic changes, including the oncogenic transformation of normal cells.^15–19 For instance, distinct subpopulations of EVs derived from neuroblastoma cells exhibit exclusive expression of either the transmembrane tetraspanin CD63 or the amyloid precursor protein, endowing them with the capacity to target different cell types selectively.²⁰ Furthermore, the differential presentation of surface glycans on EVs of varying sizes contributes to the heterogeneity in cellular internalization rates.²¹ While bulk analyses have been instrumental in certain scenarios, it is crucial to acknowledge that such assays obscure the substantial heterogeneity present at the structural, compositional, and functional levels of individual EVs or at the single-cell level. For instance, these methods may fail to discern variations in reaction pathways or the molecular states of individual proteins and nucleic acids within the vesicles, potentially leading to the misinterpretation of bulk analysis outcomes.²² Generally speaking, the primary constraints in understanding EV heterogeneity are largely attributed to the lack of sufficient tools to reveal the mysteries of EV heterogeneity among overlapped populations.

Efficient isolation is a crucial prerequisite for unlocking the heterogeneity secrets of EVs and apply them to clinical practice. EVs exist in almost all kinds of body fluids, like blood, ejaculates, urine, cerebrospinal fluid, saliva, and breast milk. In such a large and complex biological matrix, it is difficult to achieve highly sensitive and specific EV separations using conventional methods such as ultracentrifugation (UC), ultrafiltration, and density gradient centrifugation. Nevertheless, these methods have drawbacks, such as complexity, high costs, and low isolation efficiency, thereby limiting their widespread application in EV isolation, especially in clinical settings.^23–29

Within the contemporary landscape of precision oncology, the analysis of multi-dimensional molecular alterations within EVs holds significant promise.³⁰ Such analyses not only shed light on the molecular mechanisms specific to cancer but also facilitate the identification of novel biomarker combinations for early cancer detection. Consequently, as miniaturized systems undergo rapid evolution and become more cost-effective, the molecular characterization of EVs could offer valuable technical insights for precision cancer management. This advancement would enable a comprehensive understanding of both intra- and inter-individual heterogeneity, thereby enhancing our capacity to address the complexities of cancer at a molecular level.

With advantages of low sample consumption, high analysis throughput, rapid reaction rates, and high automation levels,^31–34 microfluidics is also emerging as a promising tool for the multiplexed analysis of EVs. This miniaturized technology enables the efficient separation, detection, and characterization of EVs and their molecular cargo within small sample volumes. Therefore, we aim to present a comprehensive review that delineates the pivotal methodologies employed in microfluidic isolation and the design of multiplexed microfluidic platforms. Applications include analysis of intact EVs at the individual or single-cell level, along with their derived proteins, nucleic acids, and glycans, for further elucidation of the importance of EV heterogeneity in clinical applications. The review also addresses the potential challenges and future directions in unravelling EV heterogeneity with this promising technology.

2. Microfluidics isolation of EV

Microfluidics has become a prominent tool in biological research due to recent advancements in microfabrication technology.³¹ These advancements enable microsystem devices to efficiently separate and isolate micron- or nanoscale particles within small fluid volumes, making them particularly valuable for on-chip biological analysis. This section provides a detailed overview of microfluidics-based techniques for isolating EVs. We will focus on microfluidic sorting methods based on surface component affinity, acoustics, filtration, and hydrodynamics (Fig. 2).

Fig. 2 Schematic view of the microfluidic-based methods for EV isolation: surface component affinity separation (A), acoustic separation (B), filtration separation (C), and hydrodynamic flow separation (D).

2.1 Surface component affinity separation-based microfluidics

Surface component affinity (SCA)-based separation of EVs within microfluidic devices is facilitated by either the integration of microchannels lined with surface component affinity ligands like antibodies, peptides, aptamers, and lipid ligands or the introduction of ligand-coated magnetic beads into microchip systems.^35–38 These sophisticated methods are particularly effective for the discrimination of specific types of exosomes from the broader EV subpopulations. Recent scholarly work has indicated that microfluidic affinity-separation techniques offer a compelling alternative to traditional tumour biopsy procedures, underscoring their potential in advancing cancer diagnostics. Mun et al. engineered a microfluidic chip incorporating three-dimensional nanostructures in a herringbone configuration, demonstrating exceptional efficacy in the capture of specific EVs.³⁹ These nanostructures, fabricated by accumulating silica nanoparticles, were designed to augment interactions between exosomes and the substratum. The herringbone arrangement facilitates enhanced mass transfer via micro mixing, thus optimizing the capture process. The integration of anti-HER2 antibodies into the nanostructures enabled the nanochip to achieve approximately 97.7% capture efficiency of HER2-positive EVs.

The inherent limitation of certain EV isolation techniques lies in their dependence on a single antibody for both capture and identification, which constrains the ability to differentiate between various EV subtypes. To overcome this limitation, Mun et al. developed a microfluidic chip-based system, designated as the MEIS-chip, which utilizes magnetic nanoclusters (MNCs) with distinct magnetization levels for the selective isolation of exosome subtypes.⁴⁰ The MEIS-chip integrates CD63-LMC (low saturation magnetization MNC conjugated with CD63) for capturing general exosomes and HER2-HMC (high saturation magnetization MNC conjugated with HER2) for isolating HER2-overexpressing exosomes. Differential magnetization facilitates the stratified capture and separation of exosomes within the MEIS-chip through the application of magnetic fields, thereby enabling the segregation of distinct EV populations.

DNA aptamers, selected through the SELEX process,⁴¹ offer a distinct advantage over antibodies and peptides in the recognition and isolation of EVs displaying specific proteins.³⁷ These nucleic acid fragments exhibit high-affinity binding to their targets. Compared to antibodies, aptamers are distinguished by their low production cost, exceptional stability, and straightforward synthesis. Consequently, the aptamer-based technique may surpass traditional methods, providing an enhanced approach to EV isolation.

To overcome the hydrodynamic resistance near the surface, Niu et al. developed a fluid nanoporous micro interface named FluidporeFace within a herringbone microfluidic chip. By encapsulating supported lipid bilayers (SLBs) on nanoporous herringbone microstructures, not only was mass transfer improved, but multivalent recognition of aptamers was also achieved, leading to enhanced affinity reactions on multiple scales.⁴² The enhanced affinity was approximately 83 times that of non-fluid interfaces. Furthermore, a microfluidic chip was specifically engineered to create dynamic multivalent magnetic interfaces, which bolstered the kinetics and thermodynamics of biomolecule recognition for the effective separation of tumour-derived EVs.

While most affinity-based EV separation techniques are directed towards capturing proteins present on the EV surface, the lipid bilayer that encapsulates these vesicles presents an alternative and underexplored target for the development of lipid-affinity separation strategies. By harnessing the unique properties of the lipid composition, it is conceivable to engineer lipophilic probes that could offer a novel approach to EV isolation, thereby expanding the scope of current methodologies. For example, Zheng et al. designed an alternating teardrop-shaped micropillar array to help Tim4-modified magnetic beads (Tim4 beads) on a chip capture tumour-derived exosomes.⁴³ The microfluidic chip employing Tim4 beads demonstrated several advantageous properties, including minimal nonspecific adsorption and the capacity for rapid, straightforward, and hassle-free exosome isolation. Moreover, the capture efficiency of these Tim4 beads was remarkably high, achieving a rate of up to 84.9%. In contrast to conventional affinity-based separation methodologies that are contingent on the presence of specific surface antigens and are influenced by the size variability of EVs, the lipid probe offers a distinctive approach. This probe captures EVs in an antigen-agnostic and size-independent manner, thereby circumventing the potential loss of EVs due to surface heterogeneity and size-related biases inherent in traditional separation techniques.

2.2 Acoustic separation-based microfluidics

The acoustic technique employs a pair of interdigital transducers (IDTs) on a lithium niobate (LiNbO₃) piezoelectric substrate to create an acoustic field within the fluid.⁴⁴ When a sinusoidal signal is applied to the IDTs, it generates surface acoustic waves (SAW) across the piezoelectric layer. This results in a standing SAW field within the microchannel. The pressure oscillations caused by these waves create regions of high and low pressure: pressure nodes (areas of minimal pressure) and pressure antinodes (areas of maximal pressure).⁴⁵ These pressure variations generate acoustic radiation and viscous forces within the fluid. The acoustic radiation force, which is proportional to the volume of EVs, causes larger EVs to migrate more quickly toward pressure nodes or antinodes compared to smaller EVs.⁴⁶ Particles with positive acoustic contrast factors, such as cells and vesicles, move towards the pressure nodes, while those with negative acoustic contrast factors, such as certain lipoproteins, are attracted to the pressure antinodes.⁴⁷ By adjusting the position and number of pressure nodes, researchers can fine-tune the separation process to target specific EV sizes and ensure precise isolation. This capability allows for the efficient separation and isolation of EVs based on their physical properties.

Zhang et al. have introduced an innovative acoustic–hydrodynamic device, termed acoustic nanoscale separation by wave column excitation resonance (ANSWER).⁴⁴ This technique generates a series of virtual acoustic wave columns within a confined microchannel filled with liquid, utilizing the synergy of SAW and excitation resonance. As particles of varying dimensions traverse these virtual columns, they experience disparate magnitudes of acoustic radiation forces, causing them to follow distinct trajectories within the channel. The capability to alter the acoustic parameters on the fly allows for a dynamic modulation of the acoustic radiation force, thereby enabling the selective tuning of the separation cut-off size for particles. Consequently, this method achieves a one-step process for the rapid and high-purity separation of subpopulations of small EVs of varying sizes from human plasma.

To address the limitations of current EV isolation methods, which frequently require costly consumables, expensive equipment, and skilled personnel, while being susceptible to contamination, Naquin et al. have engineered an innovative device, the ASCENDx, which employs an acoustic-driven microfluidic disc for the efficient separation and concentration of EVs from plasma samples.⁴⁸ The ASCENDx platform incorporates a rotating microfluidic disc that utilizes SAW to drive disc rotation on a spinning droplet, enabling rapid separation and analysis of EVs.

2.3 Filtration separation-based microfluidics

Filtration is a versatile technique for the continuous isolation and enrichment of EVs without requiring any active components. This passive approach leverages the size-based separation of EVs, allowing for their continuous and efficient isolation from complex biological samples.^49–51 Ramnauth et al. have developed a novel microfluidic device capable of rapidly filtering exosomes from small volumes of 50–100 μL of patient plasma.⁵² The device employs a filtration method that leverages the principle of size exclusion to collect exosomes. In this setup, a 2 μm pore size filter is used to retain larger biomolecules present in the biological sample. The exosomes are then isolated using 15 or 30 nm filters. Moreover, Li et al. have engineered a set of cascaded microfluidic circuits for the pulsed filtration of EVs directly from whole blood samples.⁵³ The system comprises a cell removal module featuring an external polycarbonate membrane filter with 600 nm pores, and an EV isolation module with an anodized aluminium oxide (AAO) membrane filter having 20 nm pores. Drawing an analogy with electrical circuits, the microfluidic design is crafted to produce a pulsating flow across the porous membrane, which dislodges particles from the surface, thereby preventing filter clogging and particle clumping. This microfluidic pulsed filtration technique enables the rapid, high-yield, and high-purity extraction of EVs from whole blood within 30 minutes.

2.4 Hydrodynamic flows-based microfluidics

Hydrodynamic separation techniques operate on the principle that, at low Reynolds numbers – characteristic of microfluidic scales – the centres of particles tend to follow – fluid streamlines. By strategically manipulating the flow through various inlets, channel geometries, and outlet configurations, it is possible to effectively separate and sort particles based on their size. This approach enables precise control over the movement of particles within the fluid, facilitating targeted separation according to specific dimensions.^54,55 The current methods for separating exosomes using on-chip hydrodynamics are typically categorized into three main categories: nanoscale deterministic lateral displacement (nano-DLD), viscoelastic microfluidics, and flow-field-flow separation.

Deterministic lateral displacement (DLD) is a microfluidic technique effectively used for separating exosomes from biological fluids, such as conditioned media and serum from liposarcoma cells. This method involves integrating an array of micropillars with a specific deflection angle into a microfluidic chip. These micropillars create fluidic forces and obstacles that influence particle flow. Particles larger than a certain size threshold, known as the critical diameter (D_c), are laterally displaced along the array upon collision with the micropillars, while smaller particles follow the streamlines without displacement.^56,57 The critical diameter is determined by the geometry and arrangement of the micropillar array. Wunsch et al. were pioneers in developing nanoscale DLD arrays with micropillar gaps as narrow as 25 nm, enabling the separation of particles in the size range of 20 to 110 nm.⁵⁶ However, the low flow rate within these chips presented challenges for sample handling. Smith et al. subsequently improved the design by integrating 1024 parallel nano-DLD arrays into a single chip, which significantly increased the processing rate up to 900 μL h⁻¹ while using gaps as narrow as 25 nm.⁵⁷ This enhanced nano-DLD chip successfully isolated exosomes from urine and serum with recoveries of approximately 50%, offering high throughput and rapid processing.

Viscoelastic microfluidics offers a label-free and straightforward technique for the efficient sorting of exosomes using synthetic polymers such as polyvinylpyrrolidone (PVP), polyethylene oxide (PEO), and polyacrylamide (PAA) as viscoelastic media. Exosomes suspended in these viscoelastic media experience size-dependent elastic lift forces within microchannels, leading to variations in their migration velocities and trajectories.^58,59 For characterization, Liu et al. implemented a high aspect ratio microchannel chip for handling biological fluids containing 0.1% PEO.⁶⁰ Initially, fluid samples are directionalized along the microchannel sidewalls using a sheath fluid, resulting in larger EVs migrating swiftly towards the channel centerline due to their elevated elastic lift. Smaller EVs are inclined to remain proximate to the sidewalls. Under optimized conditions, the method enabled the successful isolation of exosomes with a purity rate of 94% and recovery rate of 80%. This advanced viscoelastic microfluidic approach constitutes a powerful method for exosome separation without intricate manipulations. However, it may exhibit limitations in distinguishing exosomes by proteins or other diminutive biomolecules.

Asymmetric Flow Field Flow Fractionation (AF4) is an adaptation of the versatile Flow Field Flow Separation (FFFS) technique.⁶¹ AF4 has emerged as a prominent tool for the separation of exosome subpopulations, offering a high-resolution approach to discern their heterogeneity. The AF4 system utilizes a thin and flat microchannel equipped with a semi-permissive membrane at the bottom, designed to retain particles that exceed a certain size threshold. The channel generates two distinct flow patterns: a parabolic laminar flow that progresses from the inlet to the outlet, and a staggered flow that crosses from the top wall to the bottom wall, perpendicular to the laminar flow. The interplay of these flows results in a unique separation mechanism where particles with larger hydrodynamic sizes and reduced diffusion coefficients settle near the channel bottom and exit at later time points due to lower laminar velocities. Conversely, particles with smaller hydrodynamic sizes and higher diffusion coefficients are subject to higher laminar velocities and exit the system earlier.⁶² Utilizing AF4, researchers have successfully characterized distinct exosome subpopulations, such as Exo-L (90–120 nm), Exo-S (60–80 nm), and smaller exosomes (∼35 nm), each exhibiting unique molecular expression profiles and distinct biological distribution patterns within organs.⁶³ The method is able to isolate these subpopulations with high efficiency, gentle handling, and rapid processing, thus making AF4 an invaluable technique for exosome research, particularly in studies focused on exosome heterogeneity and biological function.

2.5 The integration of microfluidics-based isolation

While microfluidic-based systems have made strides in exosome separation, they are often hindered by limitations such as extended processing times, intricate operations, and insufficient purity levels.⁶⁴ To counter these issues and improve efficiency, there have been significant efforts in the integration of diverse methodologies. Within this framework, the EXODUS system has been introduced as an innovative, ultrafast separation technology for the automated and label-free isolation of exosomes from a variety of biological fluids, such as plasma, urine, and cell culture media.⁶⁵ The EXODUS system employs a combination of negative pressure oscillations and a filter membrane, integrated with a dual-coupled harmonic oscillator, to enhance the exosome separation process. This approach offers several distinct advantages over traditional techniques, including increased throughput, decreased processing time, user-friendly operation, and enhanced reproducibility. Remarkably, the system achieves high-purity exosome separation, estimated at approximately 90%, by effectively reducing membrane clogging through the oscillation of the nanoporous membrane and column.

Recently, an innovative micro- and nanofluidic chip was designed for the isolation of EVs from the conditioned media and serum of liposarcoma cells.⁶⁶ This device integrates cross-flow filtration with an immunoaffinity-based trap, ensuring the recovery of over 76% of EVs from a minimal sample volume of 300 μL within 60 minutes. An oscillating viscoelastic microfluidic system has been proposed for exosome sorting, providing the advantages of a sheathless design and rapid processing. This system employs an internal electronic circuit to generate an oscillating flow that concentrates particles at various channel positions, thereby improving sorting efficiency. Additionally, a novel microfluidic technique for the simultaneous isolation and preconcentration of exosomes has been introduced.⁶⁷ This method utilizes electrophoresis to establish an electric field, which alters the lateral flow path of the particles. By incorporating ion-depleted ion-selective membranes, this approach enables the concurrent preconcentration of exosomes and filtration of cellular debris.

In summary, each microfluidic-based separation methods are discussed in terms of its utility in basic research and potential clinical applications, along with their limitations (Table 1). For instance, while surface component affinity separation offers high specificity, it may be limited by the need for specific ligands and the potential for low throughput. Acoustic separation provides a label-free method but may require expensive equipment. Filtration methods, although cost-effective and high-throughput, may be limited by sample clogging issues. Viscoelastic flow and hydrodynamic flow methods score well in scalability but may require to optimize processes and reduce costs. However, the aims of these technologies are to enhance the clinical potential of EVs and introduce new possibilities for clinical applications. The development of such microfluidic devices represents a significant advancement in molecular diagnostics and precision medicine within the realm of EV research.

Table 1 Comparison of microfluidics-based EV isolation methods

Isolation technique	Isolated EV size (nm)	Sample volume (μL)	Sample type	Throughout	Cost	Scalability	Ref.
Surface component affinity	30–300	20–100	Plasma; culture medium	Low, typically suitable for small-scale samples	High, due to the potential need for specific surface modifications or antibodies	Limited, dependent on specific affinity pairings	39, 68–71
Acoustic flow	30–1000	<300	Whole blood; urine	Low to moderate, depending on the precision of acoustic manipulation	High, requiring precise acoustic equipment	Scalable, but cost increases with scale	72, 73
Filtration	20–600	<100	Plasma	High, capable of continuous sample processing	Low to moderate, depending on the filter membranes and equipment	Good, suitable for scaled-up production	74, 75
Viscoelastic flow	30–200	<100	Plasma; culture medium	Moderate, limited by fluid properties	Moderate, may require additional polymers to adjust the fluid's viscoelasticity	Scalable, but consistency in fluid properties is challenging	58–60
Hydrodynamic flows	20–110	<500	Plasma; culture medium; plasma	High, suitable for high-throughput analysis	Moderate, requiring precise microfluidic chips	Excellent, easy to integrate and automate	61–63

---

3. Microfluidic profiling of EVs at bulk level

In order to delve into the dynamics and diversity of biomarkers on EVs, molecular analysis lays the foundation for clinical diagnostics. Multiplex analysis of EVs refers to the ability of the assay platform to detect a variety of EV-derived analytes (typically proteins, nucleic acids, and glycans) in a short period of time. Microfluidics-based methods for EV analysis offer multiple advantages such as precise control of fluid flow, reduced sample volume requirements, and increased sensitivity and specificity for EV detection and analysis. By utilising tandem biological microscale channels and structures, it is possible to perform in-depth analysis of EVs and identify biomarkers associated with specific biological processes or diseases. Simultaneous detection and quantification of multiple EV biomarkers can enhance the comprehensive characterisation of EVs, leading to a better understanding of EV heterogeneity and functional roles. Therefore, microfluidic-based EV profiling promises to improve our understanding of EV biology and facilitate the development of novel diagnostic and therapeutic strategies for various diseases. By taking advantage of the unique capabilities of microfluidics, researchers can gain deeper insights into EVs and harness their potential for clinical applications.

3.1 Microfluidic-based profiling of EV proteins

Conducting multiplexed profiling of EV surface protein biomarkers offers the potential to acquire abundant information that more accurately reflects the molecular signature of parental cells. This comprehensive approach holds promise to enhance the accuracy of cancer diagnostics and therapy monitoring. Since EVs are derived from parent cells, the surface protein markers on EVs can provide valuable insights into the molecular characteristics of the originating cells.⁷⁶ Therefore, analysis of a diverse panel of EV surface protein biomarkers enables a more accurate representation of the complex molecular profile of the parent cells, compared to the assessment of a single biomarker in EVs. This wealth of information has the potential to improve the precision of cancer diagnostics by facilitating the identification of specific biomarker patterns associated with different cancer subtypes or stages.⁷⁷ The various microfluidic-based biosensors which enable the simultaneous detection of multiplexed EV surface proteins are summarized in Table 2. These innovative technologies offer the capability to analyse multiple EV biomarkers concurrently, providing a comprehensive snapshot of the molecular profile of EVs. By leveraging the sensitivity and specificity of microfluidic-based biosensors, the accuracy and efficiency of EV surface protein analysis for diagnostic and therapeutic applications have been improved.

Table 2 Summary of the microfluidics-based biosensors for profiling of EV proteins

Sensor type	Target biomarkers	Cancer type	Detection of limit	EVs separation	Diagnostic performance	Merits	Demerits	Ref.
Note: N/A: not applicable; PC: prostatic cancer; OC: oral cancer; BC: breast cancer; CRC: colorectal cancer; GC: gastric cancer; LUC: lung cancer; HCC: hepatic cell carcinoma; AD: Alzheimer's disease; MC: melanoma cancer; OSC: osteosarcoma; UC: ultracentrifugation; SCA: surface component affinity.
Electrochemical	PSMA, EpCAM	PC	N/A	UC	Accuracy: 80%	Rapid response; low limit of detection; low cost; high sensitivity and specificity	Low to moderate reproducibility; requires specialized device; low to moderate throughput	78
	EpCAM, CD24, CA125, HER2, MUC18 and EGFR	OC	3.0 × 10^4 particles per mL	SCA	N/A			79
	MUC1, HER2, EpCAM, and CEA	BC	N/A	SCA	AUC = 1.00			80
	EpCAM, EGFR, CD133, GPA33, CD24 and CD63	CRC	10^4 particles per mL	SCA	Sensitivity: 94%; specificity: 100%; accuracy: 96% (n = 48)			81
	CD63, HER2, EpCAM, and PDL1	BC	2.54 × 10^4 particles per mL	SCA	Sensitivity: 100%; AUC = 1.00			82
	CD63, CD9, EGFR, and EGFRvIII	GC	5.0 × 10^2 particles per mL	SCA	AUC: 0.94 (n = 20)			83
Fluorescence	CD9, CD63, EGFR, HER2, CA125, FRα, CD24 and EpCAM	OC	21 particles per mL	SCA	Accuracy: 100%; AUC = 1 (n = 20)	Low-cost; high sensitivity; visual detection; moderate to high reproducibility	Low to moderate throughput	84
	CEA, Cyfra21-1 and ProGRP	LUC	N/A	SCA	N/A			85
	CD9 and CD63, MMP14-E, and MMP14-A	BC	16 particles per mL	SCA	Accuracy: 92.9% (n = 70)			86
	CD81, PSMA, and EpCAM	PC	1.1 × 10^6 particles per mL	SCA	N/A			87
	PD-L1, EpCAM, HER2	BC	1.01 × 10^4 particles per mL	UC	N/A			88
	CA 15-3, CA 125, CEA, HER2, EGFR, PSMA, EpCAM, and VEGF	BC	3.8 × 10^7 particles per mL	Thermophoretic enrichment	Accuracy: 91.1%;			77
	CD63, PTK7, EpCAM, LZH8, HER2, PSA and CA125	BC	3.3 × 10^6 particles per mL	Thermophoretic enrichment	Sensitivity: 95%; specificity: 100%; accuracy: 68% (n = 102)			76
	CD81, EpCAM, and HER2	BC	10 particles per mL	SCA	N/A			89
SPR	CD9, CD63, CD82, CD41b, EpCAM, E-cadherin	HCC	4.87 × 10^7 vesicles per cm^2	SCA	N/A	Real-time detection; high throughput; high sensitivity; label-free; high reproducibility	High cost; requires specialized device (nanohole arrays)	90
	EGFR, EpCAM, HSP70, HSP90, CD63, TSG101	OC	10^4 particles per mL	SCA	N/A			91
	CD63, CD24, EpCAM, and MUC1	CRC and GC	1.5 × 10^3 particles per mL	Filtration	AUC = 0.970 (n = 20)			92
	CD63, EGFR, EpCAM and MUC1	LUC	10^3 particles per mL	Filtration	AUC = 0.982 (n = 76)			93
	CD63, CD9, CD81, NCAM, L1CAM, and CHL-1	AD	10^2 particles per mL	N/A	N/A			94
	EpCAM, CD24, CA19-9, CLDN3, CA-125, MUC18, EGFR, HER2	OC	3.0 × 10^3 particles per mL	N/A	Accuracy: 97% (n = 30)			95
SERS	MCSP, MCAM, ErbB3, and LNGFR	MC	N/A	SCA	N/A	Moderate to high throughput; high sensitivity; molecular vibration fingerprint, easy operability	Moderate to high cost; low to moderate reproducibility	96
	CD63, CD9, and CD81	MC, LC and BC	N/A	SCA	N/A			97
	MCSP, MCAM, CD61 and CD63	MC	N/A	SCA	AUC = 0.95 (n = 41)			98
	CD63, VIM and EpCAM	OSC	2 particles per mL	SCA	Sensitivity: 100%; specificity: 90%; accuracy: 95%; AUC = 0.971 (n = 30)			99
	CD63, HER2, EpCAM, PDL1, CEA, and MUC1	BC	2.0 × 10^4 particles per mL	SCA	AUC = 1 (n = 30)			100
	CD63, PDL1 and EGFR	LUC	4.46 × 10^2 particles per mL	SCA	N/A			101
	CD63, MUC1, EGFR, and TNC	LUC	N/A	SCA	AUC = 1 (n = 76)			102
	N-cadherin, E-cadherin, THBS1 and ABCB5	MC	10^5 particles per mL	SCA	N/A			103
FCM	CA125, STIP1, CD24, EpCAM, EGFR, MUC1, and HER2	OC	N/A	N/A	Sensitivity: 92.6%; specificity: 100%; accuracy: 94.2%; AUC: 0.973 (n = 69)	High throughput; capable of single-particle analysis; high reproducibility	High equipment and operational costs; requires high sample stability and uniformity	104

---

3.1.1 Microfluidic-based electrochemical biosensors. Microfluidic-based electrochemical biosensors offer a promising avenue for detecting EVs due to their high sensitivity, automated operation, and small-scale design. By integrating electrodes into microfluidic chips, EVs can be analysed on-site. However, the key to enabling clinical diagnosis lies in constructing low-cost, yet highly sensitive, chips. These devices utilize electrodes that are tailored with specific capture reagents, such as antibodies or aptamers, to selectively recognize the surface proteins of target EVs. This binding generates measurable electrical signals, facilitating the rapid and efficient simultaneous detection of multiple biomarkers. Advances in manufacturing technology have expanded the capabilities of microfluidic-based electrochemical biosensors by introducing multi-channel electrodes using different materials and designs.^105,106

For instance, Zhou et al. introduced a multiplexed electrochemical sensor utilizing a microfabricated chip with multiple gold electrodes to detect exosomes derived from prostate cancer.⁷⁸ Through the electrooxidation of labelled magnetic nanoparticles (MNPs), EpCAM and PSMA proteins expressed by exosomes can be identified, producing distinct electrochemical signals. To enhance the efficiency of protein detection, advanced platforms for multi-target and highly sensitive analysis have emerged. For example, Kilic et al. have devised an innovative microfluidics-based electrochemical system known as iPEX (impedance profiling of extracellular vesicles), designed for rapid and multi-channel analysis of EV protein profiles (Fig. 3A).⁸³ The iPEX system leverages impedance measurements to evaluate the electrical properties of EVs and their protein biomarkers. By analysing impedance changes resulting from EV–protein interactions, this approach enables the simultaneous detection of multiple EV surface proteins (up to four) on a single chip. Additionally, Park et al. integrated a microfluidic electrochemical sensor with a 96-well plate, significantly enhancing detection throughput.⁸¹ The total readout time for all 96 probes is less than 2 minutes, markedly improving the detection efficiency. Multiplexed profiling holds the potential to enhance diagnostic accuracy, reduce missed diagnoses, and minimize the need for costly procedures, ultimately optimizing healthcare resource utilization and lowering medical expenses. Despite the numerous advantages of electrochemical methods, challenges exist, including issues with surface functionalization, sample matrix effects, and reproducibility. Ongoing efforts are dedicated to refining these strategies by exploring robust surface functionalization techniques and ensuring electrode stability.

Fig. 3 Schematic illustrations of representative microfluidic platforms for multiplexed profiling of EV surface proteins. (A) Illustration of simultaneous detection of multiple proteins on the EV surface by label-free microfluidic electrochemical impedance of iPEX chip based on quadruplicate measurements. Reproduced with permission.⁸³ Copyright 2022, American Chemical Society. (B) Schematic illustration of fluorescent microfluidic biosensor for detection of EV matrix metalloproteinases and their proteolytic activity. Reproduced with permission.⁸⁶ Copyright 2020, The American Association for the Advancement of Science. (C) Schematic view of SPRi microfluidic biosensor for profiling of EV transmembrane proteins. Reproduced with permission.⁹¹ Copyright 2018, American Chemical Society. (D) Schematic view of SERS microfluidic biosensor for profiling of EV proteins. Reproduced with permission.¹⁰² Copyright 2024, American Chemical Society. (E) Illustration of single-wavelength algorithm-aided microfluidic imaging biosensor for profiling of EV proteins. Reproduced with permission.¹⁰⁷ Copyright 2021, Springer Nature Publishing. (F) Schematic view of aptamers-assisted nanoflow cytometry for profiling of EV proteins. Reproduced with permission.¹⁰⁴ Copyright 2024, Wiley-VCH Verlag.

While these advancements are commendable, it is essential to acknowledge the inherent limitations and ongoing challenges within this field. The sensitivity and specificity of these microfluidic-based electrochemical biosensors are significantly influenced by the stability of the biomolecular elements and the effectiveness of the electrode modifications. Additionally, the incorporation of these sensors into practical diagnostic devices encounters hurdles, particularly concerning the reproducibility and reliability across diverse testing environments. Therefore, despite the progress made in microfluidic-based electrochemical biosensors, the path toward their routine clinical application remains fraught with challenges that require sustained research and innovative solutions.

3.1.2 Microfluidic-based optical biosensors. Due to their high sensitivity, speed, and high throughput advantages, microfluidic-based optical biosensors have been used for multiplex analysis of EV surface proteins. An optical biosensor consists of a biometric element that is specific to the analyte and a sensor element that converts the biological signal into a quantifiable optical signal. According to the signal detection device and transduction mode, microfluidic-based optical biosensors can be divided into fluorescence biosensors, surface plasmon resonance (SPR) biosensors and surface-enhanced Raman spectroscopy (SERS) biosensors.

Microfluidic-based fluorescence biosensors using fluorescent dyes, quantum dots (QDs) or metal particles are popular strategies in multiplexed profiling of EV surface proteins.^108–110 Bai et al. reported a bead-based microarray for exosome isolation and multiplexed tumour marker detection.¹¹¹ CD63 antibody-functionalized beads were utilized for immunocapture separation of exosomes. The exosomes were subsequently labelled with antigen-positive fluorescent QDs. The beads were uniformly trapped and aligned among micropillars on the chip. This design improves the fluorescence observation by spreading the signals evenly across each bead, thereby preventing optical interference and leading to more accurate test results. However, this method directly labels the exosome surface proteins without amplifying the fluorescence signal, which may result in a lower detection sensitivity.

In order to enhance the efficiency of capture and fluorescent signal intensity of exosomes, Zhang et al. developed a microfluidic exosome analysis platform using three-dimensional (3D) herringbone nanopatterns (Fig. 3B).⁸⁶ The interface with the 3D herringbone patterned chip provides a large specific surface area, which significantly improves the immune capture efficiency of exosomes and effectively permits drainage of the boundary fluid to reduce near-surface hydrodynamic resistance.

Based on this nano-interface, an exosome ELISA assay via signal amplification by SβG enzyme catalysis of FDG substrate was developed with an LOD of 10 particles per μL. While smart nanomaterials combined with fluorescence have found extensive use in EV detection, they still face challenges such as quenching effects and background fluorescence interference. The development of new materials with minimal background signal and resilient fluorescence is essential. Fluorescence biosensors offer rapid detection times and a straightforward mechanism, particularly when combined with microfluidic devices and nanochips, thereby enhancing the robustness of EV analysis.

Microfluidic-based SPR biosensors are frequently utilized for detecting analytes and characterizing molecular interactions, including antibody–antigen, proteins, and small molecules. These sensors detect changes in the local refractive index caused by binding of target substances to a sensing surface. This results in an optical resonance shift, enabling the label-free detection of target molecules captured by immobilized ligands on the sensor surface. Additionally, SPR sensors have narrow sensing ranges of 10 to 300 nm from the surface.¹¹² This range aligns well with the size of most EVs, such as exosomes (40–160 nm), which fall within the evanescent field of surface plasmons. This advanced technology provides real-time, label-free detection of biomolecular interaction, enabling the simultaneous analysis of multiple EV biomarkers with high sensitivity and specificity. For instance, Im et al. developed a nano plasmonic assay to profile exosomes based on their membrane proteins and lysate proteins.⁹⁵ This method employed transmission SPR and involved the use of regularly arranged nanohole arrays functionalized with antibodies. Multiple profiling can be conducted by using different antibodies. This study represents the first instance of performing multiplex analyses of EVs with SPR technology.

To improve the analysis throughput, Zhu et al. developed an SPRi (SPR imaging) technology capable of simultaneously analysing four EV markers.⁹⁰ While label-free SPR techniques allow for multiplexed profiling of EVs, there are still sensitivity challenges in analysing source-specific EVs that are less abundant in the circulating system. In order to achieve highly sensitive multiple profiling of EVs, Park et al. labelled the EV markers with gold nanoparticles (AuNPs) of a size similar to that of the EVs (Fig. 3C).⁹¹ This labelling method amplifies electromagnetic fields through plasmonic coupling between particles and the device surface.

SPR microchips offer significant benefits in the quantification and subpopulation analysis of EVs. Nevertheless, the fabrication of most of these chips necessitates the use of focused ion-beam milling to produce the nanoholes, leading to a time-consuming and costly preparation process. Furthermore, nonspecific adsorption hinders detection accuracy, which could be mitigated by enhancing the frequency and duration of washing and by optimizing the capture interface.

In contrast to fluorescence-based techniques, which are hindered by the issue of overlapping spectra and photobleaching of fluorophores, microfluidic-based SERS biosensors offer the ability to efficiently monitor multiple targets on a single substrate due to their distinctive molecular fingerprint properties.¹¹³ This capability is particularly important when dealing with very small sample volumes and low concentrations, as well as samples containing multiple analytes within a single target, such as EVs. Recently, Wang et al. created a microfluidic chip called EPAC that consists of a single substrate modified with specific capture antibodies, gold nanoparticles (AuNPs) linked to target antibodies, and various Raman reporters.⁹⁶ By harnessing alternating current electrodynamics to guide lateral fluid flow, this biochip promotes effective encounters between antibodies and antigens, minimizing the presence of nonspecific molecules. This innovative approach allows for the accurate depiction of EV phenotypic diversity and patient response to treatment. The same group further coupled a microfluidic structure containing six independent flow channels with an asymmetric circle-ring electrode to enhance sample mixing and improve the detection performance of microfluidic-based SERS sensors (Fig. 3D).¹⁰² Unlike the horizontal fluid microfluidic chips discussed previously, Su et al. developed a vertical flow microfluidic-based SERS biosensor for multiplexed EV detection and found that this approach reduced cross-reactivity and false-negative results.¹¹⁴ However, the simultaneous multiplexed detection capacity is limited by steric hindrance from multiple binding on the EV surface. Our group has designed a magnetically driven tandem chip integrated with exonuclease I-based strategy to eliminate steric hindrance and amplify the SERS signal of multiple protein biomarkers on EVs, allowing simultaneous multiplexed profiling of the six EV biomarkers within 1.5 h.¹⁰⁰ Despite the wide application of SERS approaches, it is evident that the repeatability of the SERS approach is subpar. This issue could potentially be addressed by optimization and development of SERS tags.

Generally, microfluidic-based optical biosensors present several advantages in the detection of EV biomarkers, which include their high precision for identifying low-level biomarkers, making them suitable for early diagnosis. Additionally, their capacity to simultaneously detect multiple biomarkers using distinct fluorophores enables a more holistic analysis. Furthermore, their capability for real-time monitoring of biomarker levels is particularly beneficial for dynamic research and the ongoing management of patients. However, several disadvantages are noteworthy. The requirement for advanced instrumentation, such as spectrophotometers and lasers, may restrict their applicability in settings with limited resources. The degradation of fluorophores over time, known as photobleaching, can diminish signal intensity and potentially compromise data fidelity. Moreover, the multiplexing potential of these biosensors is constrained by spectral overlap in optical elements.

3.1.3 Other microfluidic-based biosensors. Although many research groups have utilized electrochemical or optical methods for multiplexed profiling of EVs, some have explored alternative strategies. For instance, Jahani et al. integrated a microfluidic system with three separate flow channels with an antibody-functionalized patterned dielectric meta-surface (Fig. 3E).¹⁰⁷ This meta-surface enhances the local electric field, generating a robust signal upon interacting with target analytes for multiplexed analysis of EVs. Their approach aims to enable quick detection of EVs derived from breast cancer while maintaining analytical performance and reproducibility.

The Flow Nano Analyzer fills a crucial gap in traditional flow cytometry, offering the ability to detect particles smaller than 200 nm. This breakthrough technology provides a new window into the nanoscopic world for flow detection. By providing high-resolution, high-selectivity, and high-throughput detection of individual nanoparticles (ranging from 7–1000 nm) in terms of particle size, distribution, concentration, and biochemical properties, the Nano Flow Detector offers an invaluable tool for life sciences and biomedical research.¹¹⁵ Li et al. introduced an aptamer-based nanoflow cytometry (nFCM) detection platform for molecular diagnostics of EVs (Fig. 3F).¹⁰⁴ This platform facilitates the swift analysis of seven vital protein markers from ovarian cancer cells, enabling the molecular detection and classification of ovarian cancer with an impressive accuracy of up to 94.2%.

3.2 Profiling of EV nucleic acids

EVs are known to contain various functional nucleic acids such as miRNAs, mRNAs, lncRNAs, and DNA, which carry genetic information transferred from parent cells.⁷² These nucleic acids have demonstrated significant effects on the development, physiology, and pathology of recipient cells or tissues.^116,117 Importantly, encapsulation within vesicles enhances the stability of RNA compared to free nucleic acids.^118,119 Consequently, capturing EVs and quantifying the encapsulated miRNAs and mRNAs has emerged as a promising non-invasive approach for detecting cancer-related biomarkers. Due to the relatively low quantity of nucleic acids within each EV,¹²⁰ sensitive, simple, accurate, rapid, and cost-effective microfluidic biosensors have been developed for detecting specific EV nucleic acids (refer to Table 3). Signal amplification technologies address the limitations of traditional analytical methods, which often fail to detect substances at very low concentrations.^121–123 These technologies aim to enhance the sensitivity and expand the detection range of current methods. The biosensors are categorized into polymerase-dependent and polymerase-independent microfluidic biosensors.

Table 3 Summary of the microfluidics-based biosensors for profiling of EV nucleic acids

Sensor type	Target biomarkers	Detection of limit	Detection type	EVs separation	Diagnostic performance	Merits	Demerits	Ref.
Note: N/A: not applicable; NSCLC: non-small-cell cancer; GBM: glioblastoma; LUC: lung cancer; PC: prostatic cancer; HCC: hepatic cell carcinoma; BC: breast cancer; UC: ultracentrifugation; SAC: surface component affinity.
Polymerase-dependent-based	miR-21, miR-378, miR-200 and miR-139	1.68 fM	NSCLC	UC	N/A	High throughput, capable of analysing multiple targets simultaneously	High cost, require specific polymerases; limited scalability, as it requires specific primers designed for particular targets	124
	miR-21, miR-1246 and miR-155	0.17 pM (for miR-21)	BC	SCA	AUC = 1 (n = 28)			125
		0.24 pM (for miR-1246)
		0.11 pM (for miR-155)
	mRNA: ARC, SYT17, SOX11, CYP1B1, TGFBI, LOX, SLCO3A1, PDLIM4, PERP, PLAUR, COL1A2 and ALDH1A3; miRNA: let-7b-5p, 17-5p, 21-5p, 27a-3p, 29a-3p, 30a-5p, 34a-5p, 221-3p, 222-3p and 223-3p	<10 RNA copies	GBM	SCA	AUC = 0.897 (n = 60)			126
Polymerase-independent-based	miR-223, miR-210, miR-146b and miR-127	N/A	N/A	SCA	N/A	Low to moderate cost; high scalability, as it allows for the design of aptamers or probes suitable for different targets; moderate cost	Low throughput, typically analyzing one or a few targets at a time	127
	GSTπ1, MGMT, APNG, ERCC1, ERCC2, MVP, ABCC3, CASP8, IGFBP2, CD63, EGFR, PDPN and EpHA2	N/A	GBM	SCA	Accuracy: 90% (n = 32)			128
	GAPDH, SLC9A3-AS1 and PCAT6	10 copies per μL	LUC	SCA	AUC = 0.811 (n = 62)			129
	CD63, CK18, CK19, DCN, Lgals1, Erbb3, GAPDH, ODC1, KRAS, CD45, ARG1 and H3F3A	N/A	PC	UC	AUC = 1 (n = 10)			85
	miR-486-5p and miR-21-5p	N/A	NSCLC	SCA	AUC = 0.835 (n = 43)			130
	mRNA: AFP, GPC3, ALB, APOH, FABP1, FGB, FGG, AHSG, RBP4 and TF	N/A	HCC	SCA	AUC = 0.87 (n = 95); sensitivity: 93.8%, specificity: 74.5%			131
	miR-18a-3p, miR-136-5p and miR-4685-3p	N/A	BC	SCA	N/A			132
	miR-375, miR-221, miR-210 and miR-10b	0.36 fM	BC	Thermophoretic enrichment	AUC = 0.94 (n = 29); sensitivity: 88%, specificity: 83%, accuracy: 85%			133
	miR-200b-3p, miR-21-5p, miR-22-3p and miR-26a-5p	N/A	HCC	UC	N/A			134
	58 genes	N/A	BC	SCA	Specificity: 99 ± 1%			135

---

3.2.1 Polymerase-dependent microfluidic-based biosensors. In nucleic acid amplification technology, polymerase chain reaction (PCR) significantly enhances signal detection in biological assays and is widely employed in current research and applications.^136,137 In general, EVs isolated by microchips are lysed for extraction and analysis of nucleic acids. For example, Shao et al. introduced an innovative microfluidic platform called iMER, which effectively combines immunomagnetic enrichment, RNA extraction, and real-time PCR for the analysis of EVs.¹²⁸ In their methodology, EVs are collected and lysed in a lysis buffer. The lysate then undergoes RNA extraction within a specific chamber where RNA binds to a densely packed glass-bead filter through electrostatic interactions. Subsequently, the RNA is eluted and reverse-transcribed for on-chip qPCR analysis (Fig. 4A). The iMER platform enables comparison of the mRNA profiles of two crucial enzymes and observation of dynamic changes occurring between GBM EVs and their parent cells during treatment. The study revealed significant variations in mRNA expression levels.

Fig. 4 Schematic illustrations of representative polymerase-dependent microfluidic platforms for multiplexed profiling of EV nucleic acids. (A) Schematic of iMER platform for qPCR detection of multiplex RNAs. Reproduced with permission.¹²⁸ Copyright 2015, Springer Nature Publishing. (B) Illustration of the EV Click Chip for specific isolation of EV and profiling of EV RNAs by using reverse-transcription droplet digital PCR (RT-ddPCR). Reproduced with permission.¹³¹ Copyright 2023, Wiley-VCH Verlag. (C) Schematic of dual-colour multiplexed photothermal dPCR technique for profiling EV miRNAs. Reproduced with permission.¹³⁴ Copyright 2020, Springer Nature Publishing. (D) Workflow of DMF-assisted EV isolation and profiling of EV miRNAs by using RT-qPCR. Reproduced with permission.¹³⁰ Copyright 2024, Elsevier.

Sun et al. improved the sensitivity of EV nucleic acid detection by developing a covalent chemistry-based purification system tailored for hepatocellular carcinoma (HCC) (Fig. 4B).¹³¹ They integrated this system with RT-ddPCR analysis, which led to an enhanced early diagnosis rate for HCC. In another advancement, Parvin et al. introduced a novel approach combining duplex photothermal digital polymerase chain reaction (dPCR) with a lipid nanoparticle-based EV capture technique (Fig. 4C).¹³⁴ This method enables the profiling and detection of EV-miRNAs in HCC. Beyond the contribution of microfluidic-based biosensors to PCR technology, nanomaterials have also shown significant potential in creating synergistic effects when combined with other technologies. However, traditional channel-based microfluidic systems often require intricate adjustments of micro-scale fluid flows, which can hinder their ability to establish automated and standardized protocols. In contrast, digital microfluidic (DMF) technology offers a significant advantage by enabling precise control of individual droplets on a two-dimensional surface. An array of electrodes serves as actuators, modifying the local wettability of a hydrophobic surface by applying voltage.^138,139 This capability makes DMF particularly suitable for achieving large-scale automation, a less feasible feature with channel-based systems. Moreover, the inherent flexibility and programmability of DMF allows it to execute complex sample processing procedures, including intricate sample pretreatment steps. These attributes render DMF highly advantageous for applications in point-of-care testing, where precise and reliable handling of samples is crucial. Mao et al. proposed a DMF platform to automate the traditional process of EV miRNA detection (Fig. 4D).¹³⁰ This innovative approach enhances the time efficiency of EV isolation, reducing the total duration to 20–30 minutes. Furthermore, this DMF platform integrates the analysis of EV-associated microRNAs (EV-miRNAs), demonstrating its potential for early disease screening. By combining automated EV isolation and EV-miRNA analysis, this DMF-based approach shows promise for rapid and efficient diagnostic applications.

Although polymerase-dependent microfluidic-based biosensors can achieve single-molecule detection sensitivity, its specificity can be inadequate, particularly in complex nucleic acid backgrounds, low template copy numbers, or suboptimal primer design. These issues are exacerbated in multiple rounds of multiplexed PCR, where non-specific amplification can occur due to primer–template mismatches and primer dimer formation. Despite various measures to mitigate these issues, the results can still be unsatisfactory in practical applications.

3.2.2 Polymerase-independent microfluidic-based biosensors. With the development of nucleic acid signal amplification strategies, microfluidic biosensors utilising a variety of polymerase-independent signal amplification and detection methods have emerged as an alternative for the detection of EV-associated nucleic acids, achieving high sensitivity and specificity.

To achieve highly sensitive detection of EV nucleic acids, enzyme-assisted target recycling amplification was utilized to enhance the signal of assistant DNA, thereby increasing the number of captured nanoprobes to ensure both sensitivity and specificity. Liu et al. developed the electrochemical microfluidic chip sensing (EMS) platform to enable multiplexed quantification of miR-21, miR-1246, and miR-155 (Fig. 5A).¹²⁵ This quantification was achieved across concentrations ranging from 0.5 to 1000 pM in a single 18 μL sample, with detection limits as low as 0.17 pM for miR-21, 0.24 pM for miR-1246, and 0.11 pM for miR-155.

Fig. 5 Schematic illustrations of representative polymerase-independent microfluidic platforms for multiplexed profiling of EV nucleic acids. (A) Workflow of EMS platform for multiple miRNAs assay. Reproduced with permission.¹²⁵ Copyright 2024, Elsevier. (B) Illustrations of r EZ-READ platform for reliable profiling of circulating RNAs. Reproduced with permission.¹²⁶ Copyright 2023, Springer Nature Publishing. (C) Illustration of the multiplex EV miRNAs detection using SPRi microfluidic biosensor. Reproduced with permission.¹²⁴ Copyright 2021, Elsevier.

As research advances, significant challenges emerge in achieving reliable measurements and multiplexed detection of various RNA targets, particularly those differing in length and sequence. To address these challenges, Zhang et al. proposed a technology, named enzyme ZIF-8 complexes for regenerative and catalytic digital detection of RNA (EZ-READ), employing an RNA-responsive transducer for direct activation and catalytic digital quantification (Fig. 5B).¹²⁶ This technology allows programmable and reliable detection of RNA subtypes (miRNA and mRNA) directly in minimally processed sample lysates. It establishes a low limit of detection (<10 RNA copies) and completes the process within 30 minutes.

Nanomaterial-based signal amplification technology is crucial for enhancing the efficacy of microfluidic biosensors.¹⁴⁰ It harnesses the distinctive physicochemical properties of nanomaterials, utilizing them either as catalysts or signal markers to enhance sensor sensitivity and accelerate reaction times between markers and the target analytes, thereby achieving more precise detection outcomes. Nanomaterials often exhibit unique physicochemical properties that accelerate reaction rates and prolong the activity of biorecognition elements, offering substantial potential for advancing microfluidic biosensor capabilities. Wu et al. devised a SPRi-based biosensor to concurrently detect multiple EV miRNAs in clinical samples (Fig. 5C).¹²⁴ Their method utilizes an Au-on-Ag heterostructure combined with a DNA tetrahedral framework (DTF). DTF probes, affixed to a gold array chip, captured EV miRNAs. Subsequently, single-stranded DNA (ssDNA)-functionalized silver nanocubes (AgNC) hybridized with the captured EV miRNAs. This was followed by the assembly of ssDNA-coated Au nanoparticles on the AgNC surface, creating Au-on-Ag heterostructures that acted as crucial labels to enhance SPR response. The DNA-programmed Au-on-Ag heterostructure and DTF enabled the biosensor to achieve a broad detection range from 2 fM to 20 nM, an exceptionally low limit of detection of 1.68 fM, with increased capture efficiency and enhanced antifouling properties.

In addition to in situ analysis of RNA extracted from EV lysates, microfluidic platforms enhanced with sophisticated DNA probes and advanced signal amplification and detection strategies have been utilized for the detection of EV-associated RNAs. This approach circumvents the need for EV lysis and reduces interference from non-vesicular RNAs. Zhao et al. have developed a thermophoretic sensor utilizing nanoflares for the in situ detection of EV miRNAs. This innovative approach eliminates the need for RNA extraction or target amplification.¹³³ The method relies on the thermophoretic accumulation of nanoflare-treated exosomes, which enhances the fluorescence signal upon binding with EV miRNAs. This allows for direct and quantitative measurement of EV miRNAs, achieving a detection limit of 0.36 fM in just 0.5 μL of serum samples. This technology represents a significant advance in the sensitive detection of biomolecules, particularly in complex biological samples like serum.

Continuous advances in polymerase-independent microfluidic biosensing technologies, including nanomaterial signal amplification, have facilitated the development of a novel composite signal amplification approach employing multiple nanomaterials. This innovation shows great promise in improving the performance of microfluidic biosensors. Furthermore, the emergence of new nanomaterial types is expected to address current limitations, such as complex preparation procedures, short storage durations, and limited functionalities. Integrating nanomaterials with nucleic acids or enzymes for signal amplification also holds potential for advancing microfluidic biosensors towards greater analytical capabilities and detection efficiencies.

3.3 Profiling of EV glycans

EVs hold substantial promise as diagnostic, prognostic, and therapeutic agents, primarily due to the molecular patterns on their surfaces that indicate their cell of origin and can facilitate targeted delivery to specific cells.¹⁴¹ Cancer often alters both cellular and EV glycosylation patterns, resulting in EV surfaces enriched with glycan moieties. These glycoconjugates of EVs play diverse roles in cancer, including modulation of immune responses, influencing tumour cell behaviour and metastatic site selection, thereby fostering the development of innovative diagnostic tools and therapeutic strategies.^142–145 Despite the ubiquity and versatility of glycan moieties in biological systems, their study has lagged behind the sequencing, structural elucidation, and comprehensive analysis of proteins and nucleic acids. This delayed progression can be attributed to the inherent challenges associated with the biochemical analysis of glycans.^146–149 Microfluidics indeed provides a promising approach for detecting EV-associated glycans. The various microfluidic-based biosensors are summarized in Table 4. Its advantages include high integration, minimal sample usage, and adaptable sensing strategies. These features make microfluidics well-suited for sensitive and efficient detection of EV-associated glycans, offering potential advancements in biomedical research and diagnostics.

Table 4 Summary of the microfluidics-based biosensors for profiling of EV glycans

Sensor type	Glycan type	Cancer type	Detection of limit	EVs separation	Diagnostic performance	Merits	Demerits	Ref.
Note: N/A: not applicable; NeuAc: N-acetylneuraminic acid; polylactosamine: poly-N-acetyllactosamine; Gal: galactose; GlcNAc: N-acetylglucosamine; Fuc: fucose; Man: mannose; GalNAc: N-acetylgalactosamine; Glc: glucose; Sia: sialic acid; LacNAc: N-acetyl-D-lactosamine; GlcA: glucuronic acid; PC: prostatic cancer; TNBC: triple-negative breast cancer; CRC: colorectal cancer; SAC: surface component affinity.
Fluorescence	NeuAc, polylactosamine, Gal, GlcNAc, Fuc, Man	N/A	31–2000 ng mL^−1	Hydrodynamic flows	N/A	High scalability due to lectin array; moderate to high reproducibility	Low throughput; high cost	150
	Gal, GalNAc, GlcNAc, Fuc, Glc, Sia, LacNAc, Man	N/A	Tim4-rBC2LCN: 0.97 ng mL^−1, Tim4-anti-CD63: 19.4 ng mL^−1, Tim4-anti-R-10G: 11.7 ng mL^−1	SAC	N/A			151
	Sia, Man, Fuc	PC	N/A	SAC	N/A			152
	Sia, Fuc, truncated O-glycans	N/A	HeLa exosome: 5.4 × 10^6 particles per mL	SAC	N/A			153
			PANC-1 exosomes: 1.3 × 10^6 particles per mL
	Sia, Man, Fuc	PC	N/A	SAC	N/A			154
	Man, Gal, GalNAc, Glc, GlcNAc, Sia, Fuc, GlcA, lactose	TNBC	4.1 × 10^5 particles per mL	Filtration	Accuracy: 91% (n = 64)			155
Magnetoresistance (GMR) sensor	Fuc, Man, Gal, GalNAc, Glc, GlcNAc	CRC	10^4 particles per mL	SAC	p < 0.0001 (n = 11)	High throughput; low cost; low biological background	Requires specialized equipment and further technical optimization, potentially limiting its accessibility and scalability	156

---

Common methods for detection of glycans are in the form of a microfluidic optical biosensor, usually in a label format and using different lectins. Kuno et al. proposed a high-throughput method for analysing glycans using a lectin microarray coupled with EFF detection.¹⁵⁷ This method allows the detection of glycans that bind to lectins without requiring washing steps. Despite the typically weak interactions between glycans and lectins, this system can accurately analyse glycans, showcasing its sensitivity and efficiency in glycan analysis.

Shimoda et al. also utilized an EFF-assisted lectin microarray to study surface EV glycan patterns (Fig. 6A).^150,158 Their research demonstrated the utility of this approach even with a very small amount of EV sample (<500 ng). Importantly, they observed that glycan patterns on small EVs are influenced by their cells of origin. By comparing EVs derived from 20 different types of cells, they found distinct glycan patterns that reflect the parent cells, suggesting that EV surface glycan profiling can provide insights into the cellular origin and potentially functional characteristics of EVs.

Fig. 6 Schematic illustrations of representative microfluidic platforms for multiplexed profiling of EV surface glycans. (A) Illustrations of high-throughput profiling of EV surface glycans using a lectin microarray with evanescent field fluorescence (EFF) detection. Reproduced with permission.¹⁵⁸ Copyright 2023, Springer Nature Publishing. (B) Illustration of lectin-mediated in situ rolling circle amplification (RCA) on an EV array chip for detection of cancer-related EV glycan pattern. Reproduced with permission.¹⁵³ Copyright 2018, Elsevier. (C) Illustration of dielectrophoretic (DEP) system assisted isolation of EV-bound lectin-Janus nanoparticles and multiplexed profiling of EV glycans. Reproduced with permission.¹⁵² Copyright 2024, Wiley-VCH Verlag. (D) Schematic of the EVLET system for sensitive and specific detection of EV glycans. Reproduced with permission.¹⁵⁵ Copyright 2024, Springer Nature Publishing.

Lectin array-based techniques leverage the specific recognition of glycans by lectins, enabling the direct elucidation of EV glycan profiles on their surface.^21,159–162 However, the quantification of glycan expression relies indirectly on the number of captured EVs, which can be influenced by factors such as the binding affinity between lectins and glycans. Additionally, the sensitivity and stability of these methods are compromised by challenges, such as protein denaturation and limited access to active sites, due to the immobilization of lectins on surfaces.

Feng et al. developed an EV array chip designed for straightforward, sensitive, and multiplexed analysis of cancer-related EV glycan signatures (Fig. 6B).¹⁵³ To improve the sensitivity of detection of EV glycans, this approach utilizes lectin recognition-mediated in situ rolling circle assembly of fluorophore-labelled DNA. Unlike conventional lectin arrays, this innovative method directly converts glycan recognition signals into amplified fluorescence detection signals.

Biosensing predominantly hinges on targets interacting with surface-immobilized probes for affinity capture. In these interfacial processes, the transfer of targets to the surface and the equilibrium and kinetics of binding reactions are pivotal determinants of sensing efficacy.¹⁶³ To overcome these challenges, microfluidics and nanoengineering approaches have been extensively investigated. Choi et al. demonstrated the applicability of glycan-mediated EV capture by employing lectin conjugated Janus nanoparticles (lectin-JNPs) and a dielectrophoretic (DEP) technique for cancer EV detection and characterization (Fig. 6C).¹⁵² Selective EV detection was achieved using lectin-JNPs, followed by the collection of the captured EVs on a DEP-driven electrode system. The lectin-JNPs and EVs–lectin-JNP complexes exhibited distinct DEP behaviours, with the latter being trapped on the electrode upon application of an AC electric field. Furthermore, the integration of a microfluidic chip facilitated the analysis of selectively bound EVs via fluorescence intensity evaluation.

As clinical demands continue to escalate, the direct detection of EV glycans from natural biological samples has become increasingly important. However, the presence of interfering components in natural biological samples, such as glycoproteins and lipoproteins, poses a greater challenge for accurate EV glycan analysis. To address the existing challenges, Li et al. developed a lectin-based thermophoretic assay termed EVLET, enabling rapid, sensitive, and selective analysis of EV glycan profiles using a small volume of serum or plasma sample (Fig. 6D).¹⁵⁵ The method employed a vacuum membrane filtration (VMF) approach to obtain high-purity EVs by effectively removing over 99% of lipoproteins and unbound lectins within 10 minutes. The sensitivity of the thermophoretic assay was two orders of magnitude higher than conventional lectin-based ELISA. The EVLET system allowed for the quantification of EV glycans in cancer patient plasma samples in less than 100 minutes, with a cost of only $15 per patient sample. While the EVLET system has demonstrated competence in the detection of EV glycans, there exists the potential for enhancement, particularly in terms of the assay's capacity for multiplexing and its overall throughput.

Lectin-based detection of EV glycans, despite its high sensitivity and specificity, is hindered by operational complexity and the high cost of antibodies. Notably, the low affinity of lectins, with dissociation constants typically ranging from millimolar to micromolar, coupled with their insufficient specificity, restricts their broad application. Furthermore, while lectin based-techniques facilitate the analysis of the N-glycan component of EVs, their analytical capacity is constrained by the diversity and affinity of lectins. This limitation may result in the inability to encompass all glycan structural variants, thereby impeding comprehensive analysis.

3.4 Simultaneous profiling of multi-components of EVs

Given the inherent heterogeneity of EVs from different individuals or sources, relying on a single type of biomarker for EV detection is insufficient for achieving the desired accuracy in cancer diagnosis and monitoring. To enhance cancer diagnosis accuracy, Zhou et al. developed a high-throughput nano-biochip integrated system (nano-bio chip integrated system for liquid biopsy, HNCIB) for the simultaneous and sensitive analysis of EV derived proteins and miRNAs.¹⁶⁴ As illustrated in Fig. 7A, EV proteins are quantified using fluorescently labelled antibodies, while miRNAs are detected directly without RNA extraction through a novel approach involving liposomes encapsulated with specially designed molecular beacons. The versatility of this integrated chip not only allows for the simultaneous analysis of EV proteins and miRNAs but also enables precise differentiation between healthy individuals and cancer patients, achieving an excellent signal-to-noise ratio. Based on a similar principle, Zhang et al. demonstrated the recognition and co-localization of specific RNAs and proteins within EVs (Fig. 7B).¹⁶⁵ While these methods have enabled detection of various biomolecules, the sensitivity for analysing the same type of biomolecule remains low. To further improve the detection sensitivity, Nguyen et al. combined tyramine signal amplification with plasmon resonance during the detection process.¹⁶⁶ This technique was experimentally shown to achieve high diagnostic accuracy for non-small cell lung cancer. As mentioned in section 2.2, it is highlighted that the inherently low nucleic acid concentrations within EVs necessitate the application of signal amplification methods to augment assay sensitivity. Techniques such as strand displacement reaction (SDR), hybridization chain reaction (HCR), rolling circle amplification (RCA), CRISPR/Cas, and catalytic hairpin assembly (CHA) are frequently employed for this purpose.^167–173 A recent example of a microfluidic-based method for concurrently detecting surface proteins and miRNAs in EVs is illustrated in the study by Jiang et al. (Fig. 7C).¹⁷⁴ This approach involves the utilization of microfluidic channels modified with CD63 aptamers for the capture and analysis of EVs. By integrating a CHA-mediated signal amplification strategy, this technique has accomplished a detection threshold of 83 EVs per microliter. In another example as shown in Fig. 7D, Li et al. further reduced the limit of detection to 4 EVs per microlitre using RCA-based signal amplification.¹⁷⁵

Fig. 7 Schematic illustrations of representative microfluidic platforms for simultaneous multiplexed profiling of EV cargos. (A) Illustration of the HNCIB system for simultaneous detection of PD-L1 membrane protein and mRNA in EV. Reproduced with permission.¹⁶⁴ Copyright 2020, The American Association for the Advancement of Science. (B) Illustration of digital-micromirror device (DMD) UV projection system assisted simultaneous detection of CD63 membrane protein and miRNA in EV. Reproduced with permission.¹⁶⁵ Copyright 2023, Wiley-VCH Verlag. (C) Schematic of EV isolation on the microfluidic chip and simultaneous multiplexed in situ detection of EV proteins and miRNAs by using CHA signal amplification strategy. Reproduced with permission.¹⁷⁴ Copyright 2023, Elsevier. (D) Silica-coated magnetic nanorod integrated chip for multiplexed profiling of EV proteins and miRNAs. Reproduced with permission.¹⁷⁵ Copyright 2024, The American Association for the Advancement of Science.

Simultaneous multi-component detection, despite its advantages, also presents certain drawbacks. Firstly, the majority of multi-component detection systems are predicated on the differentiation of fluorescent signals, which inherently suffer from the low throughput commonly associated with fluorescence-based technique. Secondly, the scope of existing multi-component analyses is rather limited; they predominantly involve a combination of protein and nucleic acid analyses. Given the high heterogeneity of cancer, such a limited focus is often insufficient to address the complexities required for accurate cancer diagnostics. Lastly, there is a potential for cross-reactivity or competitive interactions between the components, which can compromise the specificity and sensitivity of the detection process.

Multiplex profiling of EVs bulk-level can elucidate their inherent heterogeneity. The aforementioned microfluidic-based techniques for EV multiplex component analysis demonstrate significant promise for cancer diagnostics, offering commendable performance in discerning the compositional heterogeneity of EVs.

4. Microfluidic profiling of EVs at single EV or single cell level

The heterogeneity of EV secretion can be categorized in two primary ways: the characteristics of the EVs themselves or the specific cell types from which they originate. When bulk measurements are taken, the mixed phenotypes of EVs derived from different populations of cells can obscure the unique properties of individual vesicles or specific cell types.¹⁷⁶ Hence, analysing EVs at single-cell or single-EV resolution is indeed pivotal for understanding cellular heterogeneity, as these vesicles carry specific molecular signatures indicative of their originating cells. In this regard, microfluidic devices designed specifically for single-cell or single-EV analysis serve as powerful tools, due to their notable advantages, including heightened throughput, minimal sample consumption, accelerated sample processing, and enhanced analytical sensitivity.^177,178 Additionally, microfluidics not only reduces the requisite sample volume but also enhances both throughput and sensitivity. A variety of microfluidic devices have been developed for the analysis of extracellular vesicles at the single-cell level and single-EV level, including microwell/chamber-based microfluidics, imaging-based microfluidics, and sequencing-based microfluidics.

4.1 Microwell/chamber-based microfluidic biosensors

Microwell and microchamber technology represents the most extensively utilized microfluidic platform, readily accommodating the trapping of individual cells for a variety of downstream assays.^179–182 In this respect, Son et al. encapsulated single cells along with antibody-modified sensing magnetic beads in picolitre microchambers, enabling the capture of secreted EVs (Fig. 8A).¹⁸³ The subsequent binding of a secondary antibody to the beads resulted in temporal changes in fluorescence intensity, thereby providing dynamic insights into the secretory activity of individual cells. In addition, Li et al. utilized a spatially patterned antibody barcode system alongside a high-throughput microchamber array to investigate the secretion of EVs from oral squamous cell carcinoma (OSCC) cell lines and primary cells at single-cell resolution (Fig. 8B).¹⁸⁴ The researchers engineered elongated microchambers to capture individual cells and concentrate detection targets. They employed a microchip featuring a highly parallel microchannel array to apply spatially resolved antibody barcodes onto a poly-L-lysine glass slide, facilitating the simultaneous profiling of up to nine distinct EV phenotypes without necessitating EV purification. This method enabled the concurrent detection of EVs and cytokines secreted by the same single cells, thereby providing a multidimensional perspective on cellular communication. Overall, this study highlights the efficacy of the spatially patterned antibody barcode system and high-throughput microchamber array in advancing our understanding of EV secretion and cellular communication at the single-cell level, with promising applications in Cancer Res. and diagnosis.

Fig. 8 Schematic illustrations of representative microfluidic platforms for multiplexed profiling of EVs at individual or single-cell level. (A) Illustration of a microwell-based microfluidics biosensor for dynamic monitoring of single cell-secreted EVs. Reproduced with permission.¹⁸³ Copyright 2016, the Royal Society of Chemistry. (B) Illustration of multiplexed profiling of single-cell EV secretion with microchamber and antibody barcode. Reproduced with permission.¹⁸⁴ Copyright 2023, Wiley-VCH Verlag. (C) Illustration of imaging analysis of individual EV proteins using DNA-PAINT. Reproduced with permission.¹⁸⁵ Copyright 2019, Wiley-VCH Verlag. (D) Steps for EV imaging and the microfluidic chip used for EV capture. Reproduced with permission.¹⁸⁶ Copyright 2023, Springer Nature Publishing. (E) Workflow of sequencing-based microfluidic biosensors for multiplexed profiling of individual EV proteins. Reproduced with permission.¹⁸⁷ Copyright 2021, American Chemical Society.

4.2 Imaging-based microfluidic biosensors

Imaging represents a direct method for observing and analysing single cells or single EVs. Researchers have leveraged super-resolution microscopy to break the diffraction limit and reveal the fine structural details of EVs. Chen et al. achieved continuous quantitative analysis of multiple EV surface biomarkers at the single-EV level by immobilizing EVs on polyethyleneimine (PEI)-coated glass slides, labelling them with DNA-coupled antibodies, and then combining DNA-PAINT (point accumulation for imaging in nanoscale topography) and TIRFM (total internal reflection fluorescence microscopy) techniques (Fig. 8C).¹⁸⁵ Through this approach, they were able to simultaneously analyse four EV surface biomarkers (HER2, GPC-1, EpCAM, and EGFR) to identify EVs of cancerous origin in blood samples. Furthermore, the researchers successfully applied this technology to detect pancreatic and breast cancers in unknown samples with 100% accuracy.

To improve the throughput of EV detection, Spitzberg et al. developed a novel technology called multiplexed analysis of EVs (MASEV) (Fig. 8D).¹⁸⁶ In this approach, EVs were immobilized on the surface of a microfluidic chamber and stained with a fluorescent antibody mixture. Subsequent rounds of staining were facilitated by cleaving the fluorescence on the antibodies using functionalized tetrazine scissors after each imaging acquisition. Through this iterative staining and imaging process, they successfully analysed 15 surface proteins on individual EVs, performing 3 analyses per vesicle over 70-minute cycles. Visualization of the EV subpopulations using t-distributed stochastic neighbourhood embedding (t-SNE) based on the 15 protein markers revealed the potential heterogeneity of these vesicles. Single EV analysis technology has potential application in accurately identifying multiple biomarkers on the surface of single EVs.

4.3 Sequencing-based microfluidic biosensors

The abundance of specific surface proteins on individual EVs is relatively low, necessitating the use of signal amplification methods for their analysis. Unlike PCR, which can exponentially amplify target genes, there is currently no practical method for protein amplification. However, promising solutions include signal amplification techniques that utilize an affinity probe, such as a protein, in conjunction with an amplifiable oligonucleotide. This approach, exemplified by DNA-assisted ligation or elongation assays, facilitates the conversion of protein identities into DNA sequences followed by sequencing methods, thereby enabling the detection of proteins at the single-molecule or molecular-complex level. Ko et al. developed a microfluidic droplet platform, named seiSEQ, for multiplexed profiling of surface proteins on EVs using an antibody immune-sequencing method combined with single-cell RNA sequencing (scRNA-seq).¹⁸⁷ The procedure involves several steps (Fig. 8E). First, EVs are labelled with DNA-barcoded antibodies (Ab-DNABC) specific to various proteins, forming distinct complexes. Next, these labelled EVs and barcoded beads are encapsulated into droplets, with each bead-DNABC corresponding to different EVs. By optimizing bead and channel size, flow rates, EV input concentration, and drop volume, the authors achieved ∼8.1% of droplets containing both a single EV and a single bead. Subsequently, amplicons are synthesized through multiple extension and amplification steps, and the protein composition of individual EVs is determined by sequencing these amplicons. As a proof of concept, the authors profiled eight surface proteins in 1100 EVs, revealing co-expression levels obscured in bulk analyses. SeiSEQ effectively translates protein analysis into DNA sequencing, enabling high-throughput multiplexed analysis of EV surface proteins, with potential applications in vesicle biology and clinical diagnostics.

In summary, microfluidic-based techniques offer insights into the surface and intracellular composition of individual EVs, thereby elucidating the potential heterogeneity inherent within EV populations. Imaging based-techniques are capable of delivering single-molecule resolution on EVs. Nevertheless, the issue of spectral overlap constrains the multiplexed analysis of individual EVs. Sequencing-based approaches effectively circumvent this limitation by transcribing EV surface protein expression into DNA sequences, facilitating high-throughput and multiplexed analysis. However, these methods are accompanied by increased complexity and elevated costs, which are significant considerations for their applications.

5. Conclusion and perspectives

In the era of precision medicine, there is a burgeoning demand for liquid biopsy biomarkers. The capacity to harness reliable, robust, and stable biomarkers for therapeutic applications is gaining significant momentum. The lipid bilayer membranes of EVs shield their cargo from degradation, potentially rendering them the most dependable cell-free biomarkers. Consequently, the prospect of utilizing EVs as diagnostic, prognostic, and therapeutic agents in clinical settings is expanding, as ongoing research elucidates their capabilities through indirect studies. Therefore, we have reviewed the application of microfluidics in the analysis of heterogeneity of EVs, highlighting the potential of microfluidics to improve the efficiency of EV separation and detection. Efficient separation and enrichment of EVs is achieved by a variety of microfluidic techniques, and the microfluidic biosensor combines different multiplexing strategies to achieve multiplexed profiling of proteins, nucleic acids, and glycan on the EVs at the bulk, individual or single cell level (refer to Table 5). These technologies provide new perspectives and tools for a deeper understanding of the biological functions and clinical applications of EVs. Despite the considerable promise of EVs in clinical diagnostics and therapeutics, several challenges impede their application. Future research may need to focus on the following areas:

1) Standardization in multiplexed profiling of EV biomarkers. The diverse array of assay technologies discussed in the previous sections highlight the pressing need for uniformity and standardization across laboratories, a persisting challenge in the field. The selection of EV markers, whether surface or internal, often varies significantly among laboratories. Establishing a comprehensive consensus on these markers is essential for the precise interpretation of disease-associated indicators. This consensus is particularly critical for the validation of EV biomarkers, as it promotes consistency in validation efforts across different laboratories, regardless of the assay platforms employed. Additionally, delays in transport and storage, as well as the effects of multiple freeze–thaw cycles, may adversely affect sample quality. It is, therefore, crucial for studies to provide detailed descriptions of sample handling procedures and the quality control measures implemented. Currently, most benchtop assay platforms are limited to proof-of-concept demonstrations. These platforms rely heavily on complex manufacturing processes for sensor production and a series of optimizations that can only be executed in well-equipped, centralized facilities.¹⁸⁸ Consequently, there is an urgent necessity to develop standardized approaches for both design and manufacturing.

2) More multiplexed profiling strategy. Recent microfluidic sensors have demonstrated some multiplexing capabilities for EV multi-target frontal analysis. However, the number of multiplexed targets analysed is generally fewer than ten. The primary limitation is likely the lack of an efficient signal output encoding mechanism. The genome editing capabilities of the CRISPR/Cas system have been explored as a new multiplexing platform.^189,190 Due to the inherent programmability of Cas proteins, the side-branching activity on multiple 20-mer barcodes can be expanded to report potentially hundreds of orthogonal codes.^191–193 A proof-of-concept study utilizing CRISPR-Cas12a-mediated barcoding has been applied to urine biomarker detection.¹⁹⁴ Moreover, a recent multiplex platform combining mass barcoding and DNA nanotechnology has been successfully applied to a variety of DNA assays,^195,196 demonstrating the flexibility and codability of these configurations to support complex EV analysis.

3) Machine learning assisted multiplexed profiling of EVs. It is becoming increasingly evident that combinations of candidate biomarkers are more likely to yield accurate readings in complex diseases, as single markers have failed to achieve this goal in most studies. Composite biomarkers are expected to better reflect disease stratification or progression and provide accurate diagnosis at the prodromal stage. The development of robust algorithms for selecting, combining, and analysing multiple classes of biomarkers from extensive patient cohorts is a critical factor in the success of EV-based multimodal liquid biopsy panels. Although numerous algorithms claim to accurately and specifically identify EVs, their ability to be utilized across a wide range of applications remains largely unexplored.¹⁹⁷ Standardization of these algorithms is critical, as it will significantly accelerate the development of machine learning-enhanced assays for clinical implementation. This, in turn, will enable the full realization of EVs' potential as reliable tumour diagnostic biomarkers.

4) Integration of high-throughput technologies. Recent advancements in microfluidics seek to characterize the quantity and cargo of EVs. However, these methods often fall short in terms of throughput and necessitate isolation and purification prior to analysis, which can result in the loss of crucial information regarding the original tissue microenvironment.¹⁹⁸ Consequently, there is a pressing need to develop high-throughput techniques that can comprehensively characterize the intricate heterogeneity of EVs while elucidating the underlying cellular behaviours associated with their secretory activities without imposing additional constraints. In recent years, spatial omics (SP) has gained considerable traction within the research community. This technology has been described as having “revolutionized” the field, as it enables not only qualitative analysis but also detailed spatial distribution assessments of analysed substances, all at high throughput and resolution.^199–201 Therefore, the integration of spatial genomics with microfluidic-based EV analysis is anticipated to facilitate in-depth investigations into the spatial heterogeneity of EVs, ultimately providing a robust, high-throughput tool for elucidating the complexities of EV biology.

Table 5 Summary of the different EV multiplexing strategies

EV multiplexing strategies	Principle	Merits	Demerits
Optical strategy	Labelling of different EV analytes with different optical (e.g. fluorescent or Raman) labels	Commercially available optical labels; low cost; high scalability	Spectra overlapping; limited choices; low throughout
Electrochemical strategy	Labelling of different EV analytes using different electrochemical (e.g. metal nanoparticle) labels	Simple synthesis of nanoparticles; low cost	Redox peaks overlapping; limited choices; low throughout; poor reproducibility
Sequencing strategy	Labelling and sequencing of different EV analytes using different DNA-conjugated recognition elements (e.g. antibodies, aptamers, etc.)	High throughout; prone to signal amplification	High cost; complex syntheses of DNA-conjugated complexes
Spatial segmentation strategy	Immobilization of different recognition elements on different spaces (e.g. solid supports or liquid droplets)	High throughout; low interference	Lack of instruments for simultaneous detection

---

In conclusion, EVs, as integral components of liquid biopsy, harbour immense potential for the precise diagnosis and treatment of various diseases. As technological advancements and innovations persist, the role of EVs in biomedical research and clinical applications is set to expand significantly. They are anticipated to contribute substantially to uncovering the intricacies of disease mechanisms, to the development of innovative diagnostic and therapeutic modalities, and ultimately, to propelling the evolution of precision medicine. The capacity to harness EVs for these purposes could revolutionize the approach to early detection, monitoring, and personalized treatment strategies, thereby enhancing patient outcomes and transforming the landscape of modern healthcare.

Data availability

No primary research results, software or code have been included and no new data were generated or analysed as part of this review.

Author contributions

Guihua Zhang wrote the original draft and the review. Xiaodan Huang, Sinong Liu, Yiling Xu, and Nan Wang edited the draft. Chaoyong Yang supervised and wrote the review. Zhi Zhu supervised, wrote the original draft and wrote the review. All the authors provided critical feedback and approved the manuscript.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Key R&D Program of China (2021YFA0909400, 2019YFA0905800), the National Natural Science Foundation of China (22325404, 21974113, 21927806), and the Fundamental Research Funds for the Central Universities (20720210001, 20720220005).

References

1. F. Bray, M. Laversanne, H. Sung, J. Ferlay, R. L. Siegel, I. Soerjomataram and A. Jemal, CA Cancer J Clin, 2024, 74, 229–263.
2. Y. Zhou, L. Tao, J. Qiu, J. Xu, X. Yang, Y. Zhang, X. Tian, X. Guan, X. Cen and Y. Zhao, Signal Transduction Targeted Ther., 2024, 9, 132.
3. Q. Zhu, M. Heon, Z. Zhao and M. He, Lab Chip, 2018, 18, 1690–1703.
4. R. B. Corcoran, Nat. Med., 2020, 26, 1815–1816.
5. F. Cocozza, E. Grisard, L. Martin-Jaular, M. Mathieu and C. Théry, Cell, 2020, 182, 262–262 , e261.
6. C. Alix-Panabières and K. Pantel, Nat. Biomed. Eng., 2017, 1, 0065.
7. G. Siravegna, S. Marsoni, S. Siena and A. Bardelli, Nat. Rev. Clin. Oncol., 2017, 14, 531–548.
8. R. Kalluri and V. S. LeBleu, Science, 2020, 367, eaau6977.
9. G. Raposo and P. D. Stahl, Nat. Rev. Mol. Cell Biol., 2019, 20, 509–510.
10. L. Wu, Y. Wang, X. Xu, Y. Liu, B. Lin, M. Zhang, J. Zhang, S. Wan, C. Yang and W. Tan, Chem. Rev., 2021, 121, 12035–12105.
11. L. A. Mulcahy, R. C. Pink and D. R. F. Carter, J. Extracell. Vesicles, 2014, 3, 24641.
12. A. Becker, B. K. Thakur, J. M. Weiss, H. S. Kim, H. Peinado and D. Lyden, Cancer Cell, 2016, 30, 836–848.
13. M. Mathieu, L. Martin-Jaular, G. Lavieu and C. Théry, Nat. Cell Biol., 2019, 21, 9–17.
14. G. van Niel, G. D'Angelo and G. Raposo, Nat. Rev. Mol. Cell Biol., 2018, 19, 213–228.
15. S. Li, A. Wong, H. Sun, V. Bhatia, G. Javier, S. Jana, Q. Wu, R. B. Montgomery, J. L. Wright, H.-M. Lam, A. C. Hsieh, B. M. Faltas, M. C. Haffner and J. K. Lee, Nat. Genet., 2024, 56, 371–376.
16. B. Marte, Nature, 2013, 501, 327–327.
17. A. Zomer and J. van Rheenen, Cancer Res., 2016, 76, 2071–2075.
18. C. Fedele, R. W. Tothill and G. A. McArthur, Cancer Discovery, 2014, 4, 146–148.
19. H. L. Williams, A. Dias Costa, J. Zhang, S. Raghavan, P. S. Winter, K. S. Kapner, S. P. Ginebaugh, S. A. Väyrynen, J. P. Väyrynen, C. Yuan, A. W. Navia, J. Wang, A. Yang, T. L. Bosse, R. L. Kalekar, K. E. Lowder, M. C. Lau, D. Elganainy, V. Morales-Oyarvide, D. A. Rubinson, H. Singh, K. Perez, J. M. Cleary, T. E. Clancy, J. Wang, J. D. Mancias, L. K. Brais, E. R. Hill, M. M. Kozak, D. C. Linehan, R. F. Dunne, D. T. Chang, A. C. Koong, A. F. Hezel, W. C. Hahn, A. K. Shalek, A. J. Aguirre, J. A. Nowak and B. M. Wolpin, Cancer Res., 2023, 83, 441–455.
20. M. L. Liu, K. J. Williams and V. P. Werth, in Advances in Clinical Chemistry, ed. G. S. Makowski, Elsevier, 2016, vol. 77, pp. 125–175.
21. A. Shimoda, R. Miura, H. Tateno, N. Seo, H. Shiku, S.-I. Sawada, Y. Sasaki and K. Akiyoshi, Small Methods, 2022, 6, 2100785.
22. S. Mathiasen, S. M. Christensen, J. J. Fung, S. G. F. Rasmussen, J. F. Fay, S. K. Jorgensen, S. Veshaguri, D. L. Farrens, M. Kiskowski, B. Kobilka and D. Stamou, Nat. Methods, 2014, 11, 931–934.
23. D. Freitas, M. Balmaña, J. Poças, D. Campos, H. Osório, A. Konstantinidi, S. Y. Vakhrushev, A. Magalhães and C. A. Reis, J. Extracell. Vesicles, 2019, 8, 1621131.
24. M. Zhang, K. Jin, L. Gao, Z. Zhang, F. Li, F. Zhou and L. Zhang, Small Methods, 2018, 2, 1800021.
25. Z. Zheng, W. Harshani, K. G. Andrew and A. S. Steven, Extracellular Vesicles and Circulating Nucleic Acids, 2021, 2, 80–103.
26. T. N. T. Dao, M. G. Kim, B. Koo, H. Liu, Y. O. Jang, H. J. Lee, Y. Kim, Y.-Y. Park, H. S. Kim, C.-S. Kim and Y. Shin, J. Extracell. Vesicles, 2022, 11, e12195.
27. S. Cheng, Y. Li, H. Yan, Y. Wen, X. Zhou, L. Friedman and Y. Zeng, Lab Chip, 2021, 21, 3219–3243.
28. J. S. Brzozowski, H. Jankowski, D. R. Bond, S. B. McCague, B. R. Munro, M. J. Predebon, C. J. Scarlett, K. A. Skelding and J. Weidenhofer, Lipids Health Dis., 2018, 17, 211.
29. C. M. Abreu, B. Costa-Silva, R. L. Reis, S. C. Kundu and D. Caballero, Lab Chip, 2022, 22, 1093–1125.
30. L. Gardner, K. Kostarelos, P. Mallick, C. Dive and M. Hadjidemetriou, Nat. Rev. Clin. Oncol., 2022, 19, 551–561.
31. J. M. Jackson, M. A. Witek, J. W. Kamande and S. A. Soper, Chem. Soc. Rev., 2017, 46, 4245–4280.
32. J. Chen, D. Chen, Y. Xie, T. Yuan and X. Chen, Nano-Micro Lett., 2013, 5, 66–80.
33. E. K. Sackmann, A. L. Fulton and D. J. Beebe, Nature, 2014, 507, 181–189.
34. G. M. Whitesides, Nature, 2006, 442, 368–373.
35. Y.-T. Kang, T. Hadlock, T.-W. Lo, E. Purcell, A. Mutukuri, S. Fouladdel, M. D. S. Raguera, H. Fairbairn, V. Murlidhar, A. Durham, S. A. McLean and S. Nagrath, Adv. Sci., 2020, 7, 2001581.
36. S. Zhou, T. Hu, G. Han, Y. Wu, X. Hua, J. Su, W. Jin, Y. Mou, X. Mou, Q. Li and S. Liu, Small, 2020, 16, 2004492.
37. Z. Zhou, Y. Chen and X. Qian, Biosensor, 2022, 12, 257.
38. L. Yang, X. Tong, Y. Zhang, Y. Li, J. Liu, R. Yin, D. Zeng, Y. Yuan and K. Deng, Adv. Mater Technol., 2022, 7, 2101115.
39. B. Mun, H. Jeong, R. Kim, B. Gu, J. Kim, H. Y. Son, H. W. Rho, E.-K. Lim and S. Haam, Chem. Eng. J., 2024, 482, 148851.
40. B. Mun, R. Kim, H. Jeong, B. Kang, J. Kim, H. Y. Son, J. Lim, H. W. Rho, E.-K. Lim and S. Haam, Biosens. Bioelectron., 2023, 239, 115592.
41. M. C. DeRosa, A. Lin, P. Mallikaratchy, E. M. McConnell, M. McKeague, R. Patel and S. Shigdar, Nat. Rev. Methods Primers, 2023, 3, 54.
42. Q. Niu, J. Gao, K. Zhao, X. Chen, X. Lin, C. Huang, Y. An, X. Xiao, Q. Wu, L. Cui, P. Zhang, L. Wu and C. Yang, Proc. Natl. Acad. Sci. U. S. A., 2022, 119, e2213236119.
43. L. Zheng, H. Wang, P. Zuo, Y. Liu, H. Xu and B.-C. Ye, Anal. Chem., 2022, 94, 7703–7712.
44. J. Zhang, C. Chen, R. Becker, J. Rufo, S. Yang, J. Mai, P. Zhang, Y. Gu, Z. Wang, Z. Ma, J. Xia, N. Hao, Z. Tian, D. T. W. Wong, Y. Sadovsky, L. P. Lee and T. J. Huang, Sci. Adv., 2022, 8, eade0640.
45. Y. Fan, X. Wang, J. Ren, F. Lin and J. Wu, Microsyst. Nanoeng., 2022, 8, 94.
46. K. Lee, H. Shao, R. Weissleder and H. Lee, ACS Nano, 2015, 9, 2321–2327.
47. M. Wu, C. Chen, Z. Wang, H. Bachman, Y. Ouyang, P.-H. Huang, Y. Sadovsky and T. J. Huang, Lab Chip, 2019, 19, 1174–1182.
48. T. D. Naquin, A. J. Canning, Y. Gu, J. Chen, C. M. Naquin, J. Xia, B. Lu, S. Yang, A. Koroza, K. Lin, H.-N. Wang, W. R. Jeck, L. P. Lee, T. Vo-Dinh and T. J. Huang, Sci. Adv., 2024, 10, eadm8597.
49. F. Liu, O. Vermesh, V. Mani, T. J. Ge, S. J. Madsen, A. Sabour, E.-C. Hsu, G. Gowrishankar, M. Kanada, J. V. Jokerst, R. G. Sierra, E. Chang, K. Lau, K. Sridhar, A. Bermudez, S. J. Pitteri, T. Stoyanova, R. Sinclair, V. S. Nair, S. S. Gambhir and U. Demirci, ACS Nano, 2017, 11, 10712–10723.
50. V. Sunkara, C.-J. Kim, J. Park, H.-K. Woo, D. Kim, H. K. Ha, M.-H. Kim, Y. Son, J.-R. Kim and Y.-K. Cho, Theranostics, 2019, 9, 1851–1863.
51. K. Kim, J. Park, J.-H. Jung, R. Lee, J.-H. Park, J. M. Yuk, H. Hwang and J. H. Yeon, APL Bioeng, 2021, 5, 016103.
52. N. Ramnauth, E. Neubarth, A. Makler-Disatham, M. Sher, S. Soini, V. Merk and W. Asghar, Sensors, 2023, 23, 8292.
53. Z. Li, C. Liu, Y. Cheng, Y. Li, J. Deng, L. Bai, L. Qin, H. Mei, M. Zeng, F. Tian, S. Zhang and J. Sun, Sci. Adv., 2023, 9, eade2819.
54. H. M. Tay, S. Y. Leong, X. Xu, F. Kong, M. Upadya, R. Dalan, C. Y. Tay, M. Dao, S. Suresh and H. W. Hou, Lab Chip, 2021, 21, 2511–2523.
55. L. R. Huang, E. C. Cox, R. H. Austin and J. C. Sturm, Science, 2004, 304, 987–990.
56. B. H. Wunsch, J. T. Smith, S. M. Gifford, C. Wang, M. Brink, R. L. Bruce, R. H. Austin, G. Stolovitzky and Y. Astier, Nat. Nanotechnol., 2016, 11, 936–940.
57. J. T. Smith, B. H. Wunsch, N. Dogra, M. E. Ahsen, K. Lee, K. K. Yadav, R. Weil, M. A. Pereira, J. V. Patel, E. A. Duch, J. M. Papalia, M. F. Lofaro, M. Gupta, A. K. Tewari, C. Cordon-Cardo, G. Stolovitzky and S. M. Gifford, Lab Chip, 2018, 18, 3913–3925.
58. Y. Meng, Y. Zhang, M. Bühler, S. Wang, M. Asghari, A. Stürchler, B. Mateescu, T. Weiss, S. Stavrakis and A. J. deMello, Sci. Adv., 2023, 9, eadi5296.
59. Y. Zhou, Z. Ma, M. Tayebi and Y. Ai, Anal. Chem., 2019, 91, 4577–4584.
60. C. Liu, J. Guo, F. Tian, N. Yang, F. Yan, Y. Ding, J. Wei, G. Hu, G. Nie and J. Sun, ACS Nano, 2017, 11, 6968–6976.
61. K. G. Wahlund and J. C. Giddings, Anal. Chem., 1987, 59, 1332–1339.
62. C. Contado, Anal. Bioanal. Chem., 2017, 409, 2501–2518.
63. H. Zhang and D. Lyden, Nat. Protoc., 2019, 14, 1027–1053.
64. J. Wang, P. Ma, D. H. Kim, B.-F. Liu and U. Demirci, Nano Today, 2021, 37, 101066.
65. Y. Chen, Q. Zhu, L. Cheng, Y. Wang, M. Li, Q. Yang, L. Hu, D. Lou, J. Li, X. Dong, L. P. Lee and F. Liu, Nat. Methods, 2021, 18, 212–218.
66. L. Casadei, A. Choudhury, P. Sarchet, P. Mohana Sundaram, G. Lopez, D. Braggio, G. Balakirsky, R. Pollock and S. Prakash, J. Extracell. Vesicles, 2021, 10, e12062.
67. S. Marczak, K. Richards, Z. Ramshani, E. Smith, S. Senapati, R. Hill, D. B. Go and H.-C. Chang, Electrophoresis, 2018, 39, 2029–2038.
68. Z. Zhao, Y. Yang, Y. Zeng and M. He, Lab Chip, 2016, 16, 489–496.
69. R. Chinnappan, Q. Ramadan and M. Zourob, Biosens. Bioelectron., 2023, 220, 114856.
70. Y.-T. Kang, E. Purcell, C. Palacios-Rolston, T.-W. Lo, N. Ramnath, S. Jolly and S. Nagrath, Small, 2019, 15, 1903600.
71. Y. Wan, M. Maurer, H.-Z. He, Y.-Q. Xia, S.-J. Hao, W.-L. Zhang, N. S. Yee and S.-Y. Zheng, Lab Chip, 2019, 19, 2346–2355.
72. M. Wu, Y. Ouyang, Z. Wang, R. Zhang, P.-H. Huang, C. Chen, H. Li, P. Li, D. Quinn, M. Dao, S. Suresh, Y. Sadovsky and T. J. Huang, Proc. Natl. Acad. Sci. U. S. A., 2017, 114, 10584–10589.
73. J. Rufo, P. Zhang, Z. Wang, Y. Gu, K. Yang, J. Rich, C. Chen, R. Zhong, K. Jin, Y. He, J. Xia, K. Li, J. Wu, Y. Ouyang, Y. Sadovsky, L. P. Lee and T. J. Huang, Microsyst. Nanoeng., 2024, 10, 23.
74. Z. Han, C. Peng, J. Yi, D. Zhang, X. Xiang, X. Peng, B. Su, B. Liu, Y. Shen and L. Qiao, Sens. Actuators, B, 2021, 333, 129563.
75. X. Hua, Q. Zhu, Y. Liu, S. Zhou, P. Huang, Q. Li and S. Liu, Anal. Chim. Acta, 2023, 1258, 341160.
76. C. Liu, J. Zhao, F. Tian, L. Cai, W. Zhang, Q. Feng, J. Chang, F. Wan, Y. Yang, B. Dai, Y. Cong, B. Ding, J. Sun and W. Tan, Nat. Biomed. Eng., 2019, 3, 183–193.
77. F. Tian, S. Zhang, C. Liu, Z. Han, Y. Liu, J. Deng, Y. Li, X. Wu, L. Cai, L. Qin, Q. Chen, Y. Yuan, Y. Liu, Y. Cong, B. Ding, Z. Jiang and J. Sun, Nat. Commun., 2021, 12, 2536.
78. Y.-G. Zhou, R. M. Mohamadi, M. Poudineh, L. Kermanshah, S. Ahmed, T. S. Safaei, J. Stojcic, R. K. Nam, E. H. Sargent and S. O. Kelley, Small, 2016, 12, 727–732.
79. S. Jeong, J. Park, D. Pathania, C. M. Castro, R. Weissleder and H. Lee, ACS Nano, 2016, 10, 1802–1809.
80. Y. An, R. Li, F. Zhang and P. He, Anal. Chem., 2020, 92, 5404–5410.
81. J. Park, J. S. Park, C.-H. Huang, A. Jo, K. Cook, R. Wang, H.-Y. Lin, J. Van Deun, H. Li, J. Min, L. Wang, G. Yoon, B. S. Carter, L. Balaj, G.-S. Choi, C. M. Castro, R. Weissleder and H. Lee, Nat. Biomed. Eng., 2021, 5, 678–689.
82. F. Wang, Y. Gui, W. Liu, C. Li and Y. Yang, Anal. Chem., 2022, 94, 875–883.
83. T. Kilic, Y. K. Cho, N. Jeong, I.-S. Shin, B. S. Carter, L. Balaj, R. Weissleder and H. Lee, ACS Cent. Sci., 2022, 8, 110–117.
84. P. Zhang, X. Zhou and Y. Zeng, Chem. Sci., 2019, 10, 5495–5504.
85. J. Ko, N. Bhagwat, S. S. Yee, N. Ortiz, A. Sahmoud, T. Black, N. M. Aiello, L. McKenzie, M. O'Hara, C. Redlinger, J. Romeo, E. L. Carpenter, B. Z. Stanger and D. Issadore, ACS Nano, 2017, 11, 11182–11193.
86. P. Zhang, X. Wu, G. Gardashova, Y. Yang, Y. Zhang, L. Xu and Y. Zeng, Sci. Transl. Med., 2020, 12, eaaz2878.
87. Q. Li, Y. Wang, Y. Xue, L. Qiao, G. Yu, Y. Liu and S. Yu, ACS Appl. Mater. Interfaces, 2022, 14, 14693–14702.
88. J. Wei, K. Zhu, T. Wang, T. Qi, Z. Wang, J. Li, S. Zong and Y. Cui, ACS Nano, 2024, 18, 10206–10215.
89. Y. Wang, S. Wang, A. Chen, R. Wang, L. Li and X. Fang, Analyst, 2022, 147, 4237–4248.
90. L. Zhu, K. Wang, J. Cui, H. Liu, X. Bu, H. Ma, W. Wang, H. Gong, C. Lausted, L. Hood, G. Yang and Z. Hu, Anal. Chem., 2014, 86, 8857–8864.
91. J. Park, H. Im, S. Hong, C. M. Castro, R. Weissleder and H. Lee, ACS Photonics, 2018, 5, 487–494.
92. X. Wu, H. Zhao, A. Natalia, C. Z. J. Lim, N. R. Y. Ho, C.-A. J. Ong, M. C. C. Teo, J. B. Y. So and H. Shao, Sci. Adv., 2020, 6, eaba2556.
93. S. Pan, Y. Zhang, A. Natalia, C. Z. J. Lim, N. R. Y. Ho, B. Chowbay, T. P. Loh, J. K. C. Tam and H. Shao, Nat. Nanotechnol., 2021, 16, 734–742.
94. C. Z. J. Lim, Y. Zhang, Y. Chen, H. Zhao, M. C. Stephenson, N. R. Y. Ho, Y. Chen, J. Chung, A. Reilhac, T. P. Loh, C. L. H. Chen and H. Shao, Nat. Commun., 2019, 10, 1144.
95. H. Im, H. Shao, Y. I. Park, V. M. Peterson, C. M. Castro, R. Weissleder and H. Lee, Nat. Biotechnol., 2014, 32, 490–495.
96. J. Wang, A. Wuethrich, A. A. I. Sina, R. E. Lane, L. L. Lin, Y. Wang, J. Cebon, A. Behren and M. Trau, Sci. Adv., 2020, 6, eaax3223.
97. J. Wang, A. Wuethrich, R. J. Lobb, F. Antaw, A. A. I. Sina, R. E. Lane, Q. Zhou, C. Zieschank, C. Bell, V. F. Bonazzi, L. G. Aoude, S. Everitt, B. Yeo, A. P. Barbour, A. Möller and M. Trau, ACS Sens., 2021, 6, 3182–3194.
98. J. Wang, Y.-C. Kao, Q. Zhou, A. Wuethrich, M. S. Stark, H. Schaider, H. P. Soyer, L. L. Lin and M. Trau, Adv. Funct. Mater., 2022, 32, 2010296.
99. Z. Han, X. Peng, Y. Yang, J. Yi, D. Zhao, Q. Bao, S. Long, S.-X. Yu, X.-X. Xu, B. Liu, Y.-J. Liu, Y. Shen and L. Qiao, Biosens. Bioelectron., 2022, 217, 114709.
100. G. Zhang, Y. Wang, W. Zhou, Y. Lei, J. Lu, W. Yin, Z. Zhu, C. Yang and P. Zhang, Angew. Chem., Int. Ed., 2023, 62, e202315113.
101. W. Lin, L. Yuan, Z. Gao, G. Cai, C. Liang, M. Fan, X. Xiu, Z. Huang, S. Feng and J. Wang, Sens. Actuators, B, 2023, 394, 134355.
102. J. Li, M. Li, A. Wuethrich, R. Guan, L. Zhao, C. Hu, M. Trau and Y. Sun, Anal. Chem., 2024, 96, 7651–7660.
103. Q. Zhou, J. Wang, Z. Zhang, A. Wuethrich, R. J. Lobb and M. Trau, Biosens. Bioelectron., 2024, 244, 115819.
104. J. Li, Y. Li, Q. Li, L. Sun, Q. Tan, L. Zheng, Y. Lu, J. Zhu, F. Qu and W. Tan, Angew. Chem., Int. Ed., 2024, 63, e202314262.
105. M. Lin, P. Song, G. Zhou, X. Zuo, A. Aldalbahi, X. Lou, J. Shi and C. Fan, Nat. Protoc., 2016, 11, 1244–1263.
106. F. Haghayegh, R. Salahandish, A. Zare, M. Khalghollah and A. Sanati-Nezhad, Lab Chip, 2022, 22, 108–120.
107. Y. Jahani, E. R. Arvelo, F. Yesilkoy, K. Koshelev, C. Cianciaruso, M. De Palma, Y. Kivshar and H. Altug, Nat. Commun., 2021, 12, 3246.
108. T. Takeuchi, K. Mori, H. Sunayama, E. Takano, Y. Kitayama, T. Shimizu, Y. Hirose, S. Inubushi, R. Sasaki and H. Tanino, J. Am. Chem. Soc., 2020, 142, 6617–6624.
109. Y.-T. Kang, E. Purcell, T. Hadlock, T.-W. Lo, A. Mutukuri, S. Jolly and S. Nagrath, Analyst, 2019, 144, 5785–5793.
110. C. Jiang, Y. Fu, G. Liu, B. Shu, J. Davis and G. K. Tofaris, Nano-Micro Lett., 2021, 14, 3.
111. Y. Bai, Y. Lu, K. Wang, Z. Cheng, Y. Qu, S. Qiu, L. Zhou, Z. Wu, H. Liu, J. Zhao and H. Mao, Nano-Micro Lett., 2019, 11, 59.
112. P. Singh, Sens. Actuators B Chem., 2016, 229, 110–130.
113. A. S. Ghrera, C. M. Pandey and B. D. Malhotra, Sens. Actuators B Chem., 2018, 266, 329–336.
114. X. Su, X. Liu, Y. Xie, M. Chen, C. Zheng, H. Zhong and M. Li, ACS Nano, 2023, 17, 4077–4088.
115. Y. Tian, L. Ma, M. Gong, G. Su, S. Zhu, W. Zhang, S. Wang, Z. Li, C. Chen, L. Li, L. Wu and X. Yan, ACS Nano, 2018, 12, 671–680.
116. A. Möller and R. J. Lobb, Nat. Rev. Cancer, 2020, 20, 697–709.
117. W. Stoorvogel, Blood, 2012, 119, 646–648.
118. X. Chen, H. Liang, J. Zhang, K. Zen and C.-Y. Zhang, Trends Cell Biol., 2012, 22, 125–132.
119. Y. Zhang, D. Liu, X. Chen, J. Li, L. Li, Z. Bian, F. Sun, J. Lu, Y. Yin, X. Cai, Q. Sun, K. Wang, Y. Ba, Q. Wang, D. Wang, J. Yang, P. Liu, T. Xu, Q. Yan, J. Zhang, K. Zen and C.-Y. Zhang, Mol. Cell, 2010, 39, 133–144.
120. J. R. Chevillet, Q. Kang, I. K. Ruf, H. A. Briggs, L. N. Vojtech, S. M. Hughes, H. H. Cheng, J. D. Arroyo, E. K. Meredith, E. N. Gallichotte, E. L. Pogosova-Agadjanyan, C. Morrissey, D. L. Stirewalt, F. Hladik, E. Y. Yu, C. S. Higano and M. Tewari, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 14888–14893.
121. M. Zhou, C. Li, R. Wei, H. Wang, H. Jia, C. Yan and L. Huang, Anal. Chem., 2024, 96, 4322–4329.
122. J. Tang, Q. Li, C. Yao and D. Yang, ChemPlusChem, 2023, 88, e202200345.
123. X. Wang, S. Shepherd, N. Li, C. Che, T. Song, Y. Xiong, I. R. Palm, B. Zhao, M. Kohli, U. Demirci, Y. Lu and B. T. Cunningham, Angew. Chem., Int. Ed., 2023, 62, e202217932.
124. W. Wu, X. Yu, J. Wu, T. Wu, Y. Fan, W. Chen, M. Zhao, H. Wu, X. Li and S. Ding, Biosens. Bioelectron., 2021, 175, 112835.
125. X. Liu, Y. Wang, Y. Du, J. Zhang, Y. Wang, Y. Xue, J. Zhao, L. Ge, L. Yang and F. Li, Chem. Eng. J., 2024, 486, 150233.
126. Y. Zhang, C. Y. Wong, C. Z. J. Lim, Q. Chen, Z. Yu, A. Natalia, Z. Wang, Q. Y. Pang, S. W. Lim, T. P. Loh, B. T. Ang, C. Tang and H. Shao, Nat. Commun., 2023, 14, 4278.
127. J. Qian, Q. Zhang, M. Liu, Y. Wang and M. Lu, Biosens. Bioelectron., 2022, 196, 113707.
128. H. Shao, J. Chung, K. Lee, L. Balaj, C. Min, B. S. Carter, F. H. Hochberg, X. O. Breakefield, H. Lee and R. Weissleder, Nat. Commun., 2015, 6, 6999.
129. Y. Bai, Y. Qu, Z. Wu, Y. Ren, Z. Cheng, Y. Lu, J. Hu, J. Lou, J. Zhao, C. Chen and H. Mao, Biosens. Bioelectron., 2019, 142, 111523.
130. Z. Tong, D. Yang, C. Shen, C. Li, X. Xu, Q. Li, Z. Wu, H. Ma, F. Chen and H. Mao, Anal. Chim. Acta, 2024, 1296, 342337.
131. N. Sun, Y.-T. Lee, R. Y. Zhang, R. Kao, P.-C. Teng, Y. Yang, P. Yang, J. J. Wang, M. Smalley, P.-J. Chen, M. Kim, S.-J. Chou, L. Bao, J. Wang, X. Zhang, D. Qi, J. Palomique, N. Nissen, S.-H. B. Han, S. Sadeghi, R. S. Finn, S. Saab, R. W. Busuttil, D. Markovic, D. Elashoff, H.-H. Yu, H. Li, A. P. Heaney, E. Posadas, S. You, J. D. Yang, R. Pei, V. G. Agopian, H.-R. Tseng and Y. Zhu, Nat. Commun., 2020, 11, 4489.
132. W. Chen, R. Cao, W. Su, X. Zhang, Y. Xu, P. Wang, Z. Gan, Y. Xie, H. Li and J. Qin, Lab Chip, 2021, 21, 1759–1770.
133. J. Zhao, C. Liu, Y. Li, Y. Ma, J. Deng, L. Li and J. Sun, J. Am. Chem. Soc., 2020, 142, 4996–5001.
134. R. Parvin, L. Zhang, Y. Zu and F. Ye, Small, 2023, 19, 2207672.
135. M. Hu, V. Brown, J. M. Jackson, H. Wijerathne, H. Pathak, D. C. Koestler, E. Nissen, M. L. Hupert, R. Muller, A. K. Godwin, M. A. Witek and S. A. Soper, Anal. Chem., 2023, 95, 7665–7675.
136. C. D. Ahrberg, A. Manz and B. G. Chung, Lab Chip, 2016, 16, 3866–3884.
137. A. M. Wang, M. V. Doyle and D. F. Mark, Proc. Natl. Acad. Sci. U. S. A., 1989, 86, 9717–9721.
138. J. Li, N. S. Ha, T. L. Liu, R. M. van Dam and C.-J. Kim, Nature, 2019, 572, 507–510.
139. X. Xu, L. Cai, S. Liang, Q. Zhang, S. Lin, M. Li, Q. Yang, C. Li, Z. Han and C. Yang, Lab Chip, 2023, 23, 1169–1191.
140. T. Chen, S. Yin and J. Wu, TrAC, Trends Anal. Chem., 2021, 142, 116309.
141. S. Das, C. J. Lyon and T. Hu, ACS Nano, 2024, 18, 9784–9797.
142. V. Vrablova, N. Kosutova, A. Blsakova, A. Bertokova, P. Kasak, T. Bertok and J. Tkac, Biotechnol. Adv., 2023, 67, 108196.
143. C. Williams, F. Royo, O. Aizpurua-Olaizola, R. Pazos, G.-J. Boons, N.-C. Reichardt and J. M. Falcon-Perez, J. Extracell. Vesicles, 2018, 7, 1442985.
144. Á. M. Martins, C. C. Ramos, D. Freitas and C. A. Reis, Cells, 2021, 10, 109.
145. P. Li, Q. Chang, M. Liu, K. Lei, S. Ping, J. Wang, Y. Gu, H. Ren and Y. Ma, Anal. Chem., 2023, 95, 17467–17476.
146. K. T. Schjoldager, Y. Narimatsu, H. J. Joshi and H. Clausen, Nat. Rev. Mol. Cell Biol., 2020, 21, 729–749.
147. S. S. Pinho and C. A. Reis, Nat. Rev. Cancer, 2015, 15, 540–555.
148. J. Pan, Y. Hu, S. Sun, L. Chen, M. Schnaubelt, D. Clark, M. Ao, Z. Zhang, D. Chan, J. Qian and H. Zhang, Nat. Commun., 2020, 11, 6139.
149. B. A. H. Smith and C. R. Bertozzi, Nat. Rev. Drug Discovery, 2021, 20, 217–243.
150. A. Shimoda, Y. Tahara, S.-I. Sawada, Y. Sasaki and K. Akiyoshi, Biochem. Biophys. Res. Commun., 2017, 491, 701–707.
151. S. Saito, K. Hiemori, K. Kiyoi and H. Tateno, Sci. Rep., 2018, 8, 3997.
152. Y. Choi, K. Akyildiz, J. Seong, Y. Lee, E. Jeong, J.-S. Park, D. H. Lee, K. Kim, H.-J. Koo and J. Choi, Adv. Healthcare Mater., 2024, 13, 2302313.
153. Y. Feng, Y. Guo, Y. Li, J. Tao, L. Ding, J. Wu and H. Ju, Anal. Chim. Acta, 2018, 1039, 108–115.
154. Y. Choi, U. Park, H.-J. Koo, J.-S. Park, D. H. Lee, K. Kim and J. Choi, Biosens. Bioelectron., 2021, 177, 112980.
155. Y. Li, S. Zhang, C. Liu, J. Deng, F. Tian, Q. Feng, L. Qin, L. Bai, T. Fu, L. Zhang, Y. Wang and J. Sun, Nat. Commun., 2024, 15, 2292.
156. Z. Wang, X. Sun, A. Natalia, C. S. L. Tang, C. B. T. Ang, C.-A. J. Ong, M. C. C. Teo, J. B. Y. So and H. Shao, Matter, 2020, 2, 150–166.
157. A. Kuno, N. Uchiyama, S. Koseki-Kuno, Y. Ebe, S. Takashima, M. Yamada and J. Hirabayashi, Nat. Methods, 2005, 2, 851–856.
158. A. Shimoda and K. Akiyoshi, Pharm. Res., 2023, 40, 795–800.
159. H. Du, H. Yu, F. Yang and Z. Li, Nat. Protoc., 2021, 16, 3470–3491.
160. J. Hirabayashi, M. Yamada, A. Kuno and H. Tateno, Chem. Soc. Rev., 2013, 42, 4443–4458.
161. K.-L. Hsu, K. T. Pilobello and L. K. Mahal, Nat. Chem. Biol., 2006, 2, 153–157.
162. R. Tian, H. Zhang, H. Chen, G. Liu and Z. Wang, Adv. Sci., 2018, 5, 1800214.
163. T. M. Squires, R. J. Messinger and S. R. Manalis, Nat. Biotechnol., 2008, 26, 417–426.
164. J. Zhou, Z. Wu, J. Hu, D. Yang, X. Chen, Q. Wang, J. Liu, M. Dou, W. Peng, Y. Wu, W. Wang, C. Xie, M. Wang, Y. Song, H. Zeng and C. Bai, Sci. Adv., 2020, 6, eabc1204.
165. J. Zhang, X. Y. Rima, X. Wang, L. T. H. Nguyen, K. Huntoon, Y. Ma, P. L. Palacio, K. T. Nguyen, K. Albert, M.-D. Duong-Thi, N. Walters, K. J. Kwak, M. J. Yoon, H. Li, J. Doon-Ralls, C. L. Hisey, D. Lee, Y. Wang, J. Ha, K. Scherler, S. Fallen, I. Lee, A. F. Palmer, W. Jiang, S. M. Magaña, K. Wang, B. Y. S. Kim, L. J. Lee and E. Reátegui, J. Extracell. Vesicles, 2023, 12, 12369.
166. L. T. H. Nguyen, J. Zhang, X. Y. Rima, X. Wang, K. J. Kwak, T. Okimoto, J. Amann, M. J. Yoon, T. Shukuya, C.-L. Chiang, N. Walters, Y. Ma, D. Belcher, H. Li, A. F. Palmer, D. P. Carbone, L. J. Lee and E. Reátegui, J. Extracell. Vesicles, 2022, 11, e12258.
167. W. Tang, W. Zhong, Y. Tan, G. A. Wang, F. Li and Y. Liu, Top. Curr. Chem., 2020, 378, 10.
168. S. Guo, W. N. Lin, Y. Hu, G. Sun, D.-T. Phan and C.-H. Chen, Lab Chip, 2018, 18, 1914–1920.
169. Y. Guo, W. Li, R. Zhang, S. Cao, X. Zhu, G. Chen and C. Feng, Lab Chip, 2023, 23, 2601–2610.
170. W. Yin, K. Hu, B. Yu, T. Zhang, H. Mei, B. Zhang, Z. Zou, L. Xia, Y. Gui, J. Yin, W. Jin and Y. Mu, Lab Chip, 2024, 24, 4659–4668.
171. R. Bruch, G. A. Urban and C. Dincer, Trends Biotechnol., 2019, 37, 791–792.
172. J. Liu, Y. Zhang, H. Xie, L. Zhao, L. Zheng and H. Ye, Small, 2019, 15, 1902989.
173. Q. Liu, Y. Huang, Z. Li, L. Li, Y. Zhao and M. Li, Angew. Chem., Int. Ed., 2022, 61, e202214230.
174. Y. Chen, D. Gao, Q. Zhu, B. Chu, J. Peng, J. Wang, L. Liu and Y. Jiang, Analyst, 2023, 148, 2387–2394.
175. Z. Li, K. Guo, Z. Gao, J. Chen, Z. Ye, M. Cao, S. E. Wang, Y. Yin and W. Zhong, Sci. Adv., 2024, 10, eadh8689.
176. G. Bordanaba-Florit, F. Royo, S. G. Kruglik and J. M. Falcón-Pérez, Nat. Protoc., 2021, 16, 3163–3185.
177. F. Zhu, Y. Ji, J. Deng, L. Li, X. Bai, X. Liu, B. Lin and Y. Lu, Chin. Chem. Lett., 2022, 33, 2893–2900.
178. G. Wu, Y. Zhao, X. Li, M. M. Ali, S. Jia, Y. Ren and L. Hu, TrAC, Trends Anal. Chem., 2023, 159, 116930.
179. J. Zhang, J. Xue, N. Luo, F. Chen, B. Chen and Y. Zhao, Lab Chip, 2023, 23, 1066–1079.
180. R. Kida, M. Tsugane and H. Suzuki, Lab Chip, 2024, 24, 572–583.
181. S. Li, Y. Zhu, R. Haghniaz, S. Kawakita, S. Guan, J. Chen, Z. Li, K. Mandal, J. Bahari, S. Shah, J. Guo, H. Kang, W. Sun, H.-J. Kim, V. Jucaud, M. R. Dokmeci, P. Kollbaum, C. H. Lee and A. Khademhosseini, Adv. Funct. Mater., 2022, 32, 2206620.
182. S. Ansaryan, Y.-C. Liu, X. Li, A. M. Economou, C. S. Eberhardt, C. Jandus and H. Altug, Nat. Biomed. Eng., 2023, 7, 943–958.
183. K. J. Son, A. Rahimian, D.-S. Shin, C. Siltanen, T. Patel and A. Revzin, Analyst, 2016, 141, 679–688.
184. L. Li, H. Su, Y. Ji, F. Zhu, J. Deng, X. Bai, H. Li, X. Liu, Y. Luo, B. Lin, T. Liu and Y. Lu, Adv. Sci., 2023, 10, 2301018.
185. C. Chen, S. Zong, Y. Liu, Z. Wang, Y. Zhang, B. Chen and Y. Cui, Small, 2019, 15, 1901014.
186. J. D. Spitzberg, S. Ferguson, K. S. Yang, H. M. Peterson, J. C. T. Carlson and R. Weissleder, Nat. Commun., 2023, 14, 1239.
187. J. Ko, Y. Wang, K. Sheng, D. A. Weitz and R. Weissleder, ACS Nano, 2021, 15, 5631–5638.
188. N. Soda, B. H. A. Rehm, P. Sonar, N.-T. Nguyen and M. J. A. Shiddiky, J. Mater. Chem. B, 2019, 7, 6670–6704.
189. C. Zhang, Z. Huang, Z. Li, J. Hu, R. Liu and Y. Lv, TrAC, Trends Anal. Chem., 2024, 170, 117431.
190. D. Zhang, L. Cai, X. Wei, Y. Wang, L. Shang, L. Sun and Y. Zhao, Nano Today, 2021, 40, 101268.
191. D. Chen, W. Huang, Y. Zhang, B. Chen, J. Tan, Q. Yuan and Y. Yang, Angew. Chem., Int. Ed., 2023, 62, e202304298.
192. Z. Xu, D. Chen, T. Li, J. Yan, J. Zhu, T. He, R. Hu, Y. Li, Y. Yang and M. Liu, Nat. Commun., 2022, 13, 6480.
193. N. L. Welch, M. Zhu, C. Hua, J. Weller, M. E. Mirhashemi, T. G. Nguyen, S. Mantena, M. R. Bauer, B. M. Shaw, C. M. Ackerman, S. G. Thakku, M. W. Tse, J. Kehe, M.-M. Uwera, J. S. Eversley, D. A. Bielwaski, G. McGrath, J. Braidt, J. Johnson, F. Cerrato, G. K. Moreno, L. A. Krasilnikova, B. A. Petros, G. L. Gionet, E. King, R. C. Huard, S. K. Jalbert, M. L. Cleary, N. A. Fitzgerald, S. B. Gabriel, G. R. Gallagher, S. C. Smole, L. C. Madoff, C. M. Brown, M. W. Keller, M. M. Wilson, M. K. Kirby, J. R. Barnes, D. J. Park, K. J. Siddle, C. T. Happi, D. T. Hung, M. Springer, B. L. MacInnis, J. E. Lemieux, E. Rosenberg, J. A. Branda, P. C. Blainey, P. C. Sabeti and C. Myhrvold, Nat. Med., 2022, 28, 1083–1094.
194. L. Hao, R. T. Zhao, N. L. Welch, E. K. W. Tan, Q. Zhong, N. S. Harzallah, C. Ngambenjawong, H. Ko, H. E. Fleming, P. C. Sabeti and S. N. Bhatia, Nat. Nanotechnol., 2023, 18, 798–807.
195. Y. Ji, D. Sun, Y. Zhao, J. Tang, J. Tang, J. Song, J. Zhang, X. Wang, W. Shao, D. Chen, J. Sheng and T. Liang, Nano Today, 2023, 52, 101940.
196. X. Zhang, Y. Dong, Y. Wang, Z. Zhang, X. Zhang, J.-J. Zhu, Y. Tian and Q. Min, Angew. Chem., Int. Ed., 2024, 63, e202313446.
197. Q. Zhang, T. Ren, K. Cao and Z. Xu, Biosens. Bioelectron., 2024, 251, 116076.
198. R. He, J. Zhu, P. Ji and F. Zhao, Nat. Methods, 2024, 21, 259–266.
199. J. L. Carstens, S. N. Krishnan, A. Rao, A. G. Sorace, E. H. Seeley, S. Ferri-Borgogno and J. K. Burks, Nat. Rev. Methods Primers, 2024, 4, 54.
200. J. Chen, L. Larsson, A. Swarbrick and J. Lundeberg, Nat. Rev. Clin. Oncol., 2024, 21, 660–674.
201. Y. Long, K. S. Ang, R. Sethi, S. Liao, Y. Heng, L. van Olst, S. Ye, C. Zhong, H. Xu, D. Zhang, I. Kwok, N. Husna, M. Jian, L. G. Ng, A. Chen, N. R. J. Gascoigne, D. Gate, R. Fan, X. Xu and J. Chen, Nat. Methods, 2024, 21, 1658–1667.

---