​

​

​

​

Microfluidic biosensing is transforming diagnostics by enabling faster, more sensitive, and highly accessible detection of clinically relevant biomarkers. This presentation details our strategies that integrate nanostructured materials, novel synthetic biorecognition elements, and engineered 3D-printed platforms to develop robust and scalable lab-on-chip systems.

We will begin by discussing our ultra-sensitive localized surface plasmon resonance (LSPR) biosensors based on sharp gold nanospikes. This compact system enables label-free, real-time detection of high-impact targets, such as anti-SARS-CoV-2 antibodies and prostate cancer markers, using only microliters of sample. These devices are readily adaptable for point-of-care diagnostics, achieving rapid assay times of 7–30 minutes.

Next, we introduce an electrochemical cortisol biosensor designed for real-time hormone monitoring. This system uniquely combines highly conductive porous laser-induced graphene (LIG) electrodes with in situ electropolymerized molecularly imprinted polymer nanoparticles (nanoMIPs). Integrated within a microfluidic module, the sensor achieves sub-nanomolar affinity and >95% recovery in serum under continuous flow.

Finally, we present our universal surface-functionalization strategy for 3D-printed microfluidics . This method embeds reactive functional monomers directly during photopolymerization, which permits stable protein immobilization and specific nanoMIP-based recognition across complex microfluidic architectures.

These integrated technologies collectively illustrate a powerful, unified approach toward building next-generation microfluidic biosensing platforms that are both scalable and poised to accelerate rapid, decentralized diagnostics.

Liquid metal-based sensors are transforming wearable technology by offering high conductivity, flexibility, and self-healing properties. Recent advances include microtubular sensors, which use gallium-indium-filled polymer tubes for sensitive, skin-conforming physiological monitoring, and bilayer liquid-solid conductors (b-LSC), which combine robust, self-healing interfaces with ultra-stretchable, low-resistance connections. These innovations enable precise, long-term monitoring and advanced applications in human-interfaced electronics, smart wearables and intelligent healthcare.

Label-free and pretreatment-free quantification of small extracellular vesicles (sEVs) and viruses would enable precision point-of-care diagnostics. We present a novel 30-minute Immuno-Janus Particle (IJPs) assay for several microliters of untreated plasma, urine, serum or cell-culture media. It is based on smart-phone detectable Brownian blinking of Janus microparticles that yields an LOD of 10^3 sEV or virus per ml. For virus, this is comparable to the detection limit of PCR in the original sample prior to RNA/DNA extraction.

This unprecedented sensitivity for untreated plasma is because rotational diffusivity of the microon-sized IJP scales as a sensitive -3 power of its size and the docking of a single sEV or virus can significantly change its blinking frequency. To enhance robustness and selectivity, we developed a particle tracking algorithm that deconvolves fluctuations due to translational diffusivity and a multi-valent immunocapture design to allow sEVs or viruses to outcompete fouling proteins for the antibodies on the IJP. False positives that plague labeling assays are eliminated because of the disparity in size between target particles and dispersed proteins.

In a small pilot study involving 87 subjects, including individuals with colorectal cancer, pancreatic ductal adenocarcinoma, glioblastoma, Alzheimer's disease, and healthy controls, our method accurately identified the type of disease with a high 0.90-0.99 AUC in a blind setting. Compared with an orthogonal ultracentrifugation plus surface plasmon resonance (UC+SPR) method that requires about 24 hours of pretreatment, the sensitivity and dynamic range of IJP are better by 2 logs. We will discuss our current work to extrapolate this preliminary work to barcoded IJPs that can quantify more than 10 different sEVs in the same sample.

​

​

Mesenchymal stromal cell-derived exosomes (MSC-EVs) are gaining recognition as powerful modulators of disease, not by targeting individual symptoms, but by restoring immune balance at the source. This talk explores the emerging view that immune modulation is the central mechanism driving the broad therapeutic potential of MSC exosomes across diverse conditions—from inflammation and fibrosis to tissue injury and autoimmunity. I will highlight key immunologic pathways influenced by MSC-EVs, including macrophage polarization, T cell regulation, and inhibition of neutrophil and complement activity. By reframing MSC exosomes as immune reprogramming agents, this session provides a unifying lens to understand their pleiotropic effects and positions them as next-generation, cell-free biologics with cross-disease relevance.

As a person’s physiological regulation of biomarker molecules is challenged by acute illness, exposure to toxins or even surgery, the concentration these molecules can give important information about their health. For premature infants their regulation systems have yet to mature, so they can also suffer rapid changes in biomarker concentrations that can have devastating health consequences for the baby. Our view is that to monitor such biomarker changes effectively ideally requires moment-by-moment measurement of blood or tissue concentrations. The person acts as their own control allowing acute deterioration to be noticed quickly. We have been developing a range of sensing and biosensing solutions for the invasive, minimally invasive, and non-invasive monitoring of people in healthcare situations. Microfluidics coupled to novel biosensors provide a valuable means of clinical sampling and robust quantification of measured biomarkers. I will describe the key challenges in the development of such integrated microfluidic sensing devices and present our recent data from the neonatal intensive care unit.

​

​

Microfluidics has demonstrated numerous advantages for isolation and characterization of liquid biopsy biomarkers in oncology, enabling their implementation in clinical routine. Microfluidics is also very relevant to build biomimetic and dynamic 3D models to better understand the process of metastasis or to test the efficacy and safety of therapeutic formulations. However, the implementation of these systems for the development of translation models to study disease evolution is still elusive.

In this talk, I present our most recent work in the development of tumor-on-chip models for translational studies in oncology.

This study presents a novel microneedle patch designed for the rapid analysis of interstitial fluid and continuous monitoring of skin wound pH. The microneedle array, fabricated from biocompatible materials with 3D-printing, enables minimally invasive sampling of interstitial fluid, facilitating real-time biochemical analysis. The patch incorporates pH-sensitive sensors that provide immediate feedback on wound healing conditions, allowing for timely interventions. We demonstrate the efficacy of the microneedle patch in both in vitro and in vivo models, highlighting its potential to enhance wound management by providing critical information on the biochemical environment of the wound. This innovative approach not only improves patient comfort but also offers a promising tool for personalized wound care and monitoring, paving the way for advanced therapeutic strategies in dermatology.

Photolithography is a key step in fabricating high-resolution microfluidic devices. This talk compares the two main approaches, mask-based exposure and direct laser writing, within the specific context of microfluidics, highlighting how the field’s unique requirements influence the choice of fabrication method.

​

​

Extracellular vesicles, specifically exosomes, play a crucial role in various biological processes, including cell-to-cell communication and intercellular delivery. Additionally, exosomes are regarded as liquid biopsy biomarkers for diagnostic applications. However, their submicron size presents challenges in isolating exosomes from other cell-derived extracellular vesicles (EVs). In this talk, we will explore microfluidic devices designed to separate EVs using electrokinetic methods.

A challenge with exosome-based diagnostics and therapeutics that poses a barrier to their translation into clinical practice is the low number of exosomes that can be isolated. Various methods have therefore been developed to increase exosome yield, although they are not without considerable limitations, not least the heterogeneity in the exosomal population that is obtained. We demonstrate that exposing to just several minutes of low-level high frequency (10 MHz) insults in the form of nanometer-amplitude acoustic waves is capable of effectively stimulating exosomal production and secretion. Given the very high post-stimulation retention in cell viabilities (>95%), it is possible to subject the same population of cells to repeated stimulation cycles to maintain lipodome and proteome homogeneity. In particular, we show the possibility of achieving eight- to ten-fold increases in homogeneous exosome production over 7 stimulation cycles, equivalent to approximately 2-fold/hr. In addition, we also explicate the calcium-dependent mechanistic pathway that governs the cells’ response to the high frequency mechano-stimulation, and show that such an intricate ability to modulate second messenger signalling can facilitate the possibility of directing downstream fates across different cell types.

Introducing EXODUS, our innovative automatic exosome isolation system, which employs nanofiltration techniques combining the double-coupled harmonic oscillation and periodic negative pressure oscillation. This unique integration ensures the high yield and purity of label-free exosome isolation. Due to its capbility of processing diverse range of samples, EXODUS has been widely embraced in research for disease biomarker discovery. Besides, EXODUS also features a large-scale model, EXODUS-T, specifically tailored to meet the demands of industrial-scale production.

​

​

​

​

Spatial metabolomics is driving innovation in mass spectrometry, metabolomics and spatial omics. With exciting discoveries, complementary capabilities and increasing accessibility, it has secured its place in the spatial biology toolbox, showing promise in biology, environmental sciences, medicine and pharmacology. Spatial metabolomics can be routinely applied to tissue sections for untargeted detection of metabolites and lipids. The most accessible protocols and instruments enable analyses with a near-single-cell resolution that is well-suited for interrogating human biopsies, resection samples, parts of human organs, and organs from animal models. Analyses at the subcellular, organismal and supra-organismal scale have been demonstrated but require innovation of developments of key components of mass spectrometers such as sampling systems, ionization sources and mass analyzers etc. This development has been largely driven by several technical challenges that are specific to spatial metabolomics, including the need for increased sensitivity, speed and accurate metabolite identification as well as increased computational demands for analysing and visualizing large datasets.

Biosensors offer a powerful, label-free technique allowing us to perform analysis of molecular interactions in real-time. SPR spectroscopy can address questions such as specificity of an interaction, dissociation and association rate constants; binding kinetics, binding affinity, and concentrations of selected molecules present in a sample of interest.

In this work, we report the novel cell/organoid integrated sensing platforms that allow us to real-time monitor cell and 3D tissue activities upon various of stimulations. Using the novel set up, we measured and compared the binding affinity of vascular endothelial growth factor (VEGF) to vascular endothelial growth factor receptor (VEGFR) and VEGF to bevacizumab. We also identified neruon chemical activities of neuron sepheres The microelectrode array based sensing device of the present invention comprises at least one cell culture module for culturing living cells, wherein the cell culture module is configured so that analytes secreted from the living cells can be released onto the electrode sensing surface.

Microdroplet dispensing refers to the controlled deposition of picoliter- to nanoliter-scale liquid volumes onto solid substrates, and is widely used in multiplex immunoassays, protein and antibody microarrays, and high-throughput screening. However, most existing commercial systems are expensive and operationally complex, limiting routine access to microdroplet-based workflows for many academic and translational laboratories. We have developed a pneumatic microvalve–based microdroplet dispensing platform capable of reliably generating droplets ranging from approximately 400 pL to 50 nL on solid surfaces. The platform includes, to our knowledge, world’s first commercially developed handheld microdroplet dispenser tailored for routine laboratory use. This device enables researchers to rapidly prototype assays and explore microdroplet-based experiments directly on standard substrates such as slides and microplates. For industrial-scale applications, our workstation configurations enable automated, high-speed, and high-throughput production through multi-channel synchronous dispensing. Our patented rapid solution-switching strategy allows more than 200 distinct printing solutions to be processed within approximately one hour per dispensing channel, supporting dense condition matrices and scalable manufacturing workflows. Compared with prevailing dispensing technologies, the pneumatic actuation strategy emphasizes low dead volume, minimal carryover during frequent solution switching, and gentle handling of bioactive reagents. These features are critical for workflows involving expensive proteins, antibodies, and enzymes, and support stable long-duration operation with practical maintenance in research environments. Beyond instrumentation, we provide an integrated end-to-end workflow encompassing droplet deposition, optical readout, software-based analysis, and application-level technical guidance. This allows researchers to enter microdroplet-based assay development and screening workflows within a short timeframe, without prolonged trial-and-error. The platform has already been deployed across multiple academic institutions and industrial partners, supporting applications in biological research, drug screening, and IVD assay development and production.

​

Three-dimensional (3D) multicellular spheroids represent a sophisticated approach to modeling complex physiological systems, offering significant advantages over traditional two-dimensional cell cultures. Conventional analytical techniques often rely on destructive methodologies that compromise real-time cellular monitoring and comprehensive physiological assessment. Our study introduces integrated microfluidic platforms that enable multimodal analysis of liver and tumor cellular aggregates through a non-destructive, label-free approach, with compatibility with conventional biological assay. The electrochemical modality provides real-time, label-free monitoring of cellular metabolic processes, drug response, and cellular differentiation, while the integrated MS analysis enables precise characterization of drug metabolism and cellular interactions. Our approach addresses critical limitations in current organ-on-chip technologies by facilitating continuous, non-invasive cellular assessment across multiple organ models.

Drug delivery to the central nervous system (CNS) is still challenging in treating Alzheimer’s disease (AD). Here, we developed combinatorial prodrug conjugates of galantamine and memantine and incorporated them with different neurotransmitter-lipids to formulate lipid nanoparticle (LNP), enhancing CNS delivery and therapeutic efficacy. Two drug molecules were individually linked with enzymatically cleavable ester or amide bonds, by simply reacting with fatty acids consisting of different hydrocarbon chains in flow, resulting in two types of single-prodrugs. In addition, a dual co-prodrug, covalently linking two drugs with two cleavable bonds, was newly synthesized by a 5-stepwise flow process in high purity and yield. The LNP formulations were screened in the brain organoid-based Transwells with the endothelial cell layer. Subsequently, dual co-prodrug LNP most effectively preserved neuronal integrity and mitigated degeneration and AD-related pathological markers, including amyloid-beta and phosphorylated Tau within brain organoids. These findings highlight the potential of dual co-prodrug-derived combinational drug delivery systems as a promising CNS therapy, validated through our integrated human drug screening platform.

​

​

With the advancement of microfluidic and organoid technologies, neuron-on-a-chip platforms have demonstrated tremendous potential in precisely replicating the three-dimensional structure of the brain, simulating intercellular interactions, and reconstructing in vitro physiological models. On one hand, compared to traditional two-dimensional cell cultures, neuron-on-a-chip systems can not only construct more complex physiological architectures but also provide a dynamically controllable developmental environment for neural tissues. On the other hand, compared to animal models, brain-on-a-chip technology not only effectively addresses the limitations in the number of available animal models, but also alleviates, to some extent, the ethical concerns associated with animal experiments. Through induced pluripotent stem cell (iPSC) technology, we can directly model human brain pathological phenotypes on chips, overcoming challenges posed by interspecies differences in animal models, and enabling the investigation of key events in the progression of human neurological disorders. In addition, neuron-on-a-chip systems can be integrated with other organoid tissues on microfluidic chips to create multi-organ-on-a-chip platforms. Compared to single-organ chips, multi-organ systems can reconstruct multiple human tissue organoids in vitro, enabling precise simulation of inter-organ interactions.

Most bioparticles, such as red blood cells and bacteria, are non-spherical in shape. However, conventional microfluidic separation devices are designed for spherical particles. This poses a challenge in designing separation devices for non-spherical bioparticles, as the smallest dimension of the bioparticle must be considered, which increases the difficulty of manufacturing and reduces the throughput. To address this challenge, we developed a technique to separate non-spherical bioparticles such as red blood cells using a series of pillars with different shapes. This technique takes into account the shape of the bioparticle and induces a rotational motion that allow us to leverage on the largest dimension and thereby increase the separation size, making it effective for separating micro- and nano-sized bioparticles. In addition, the detection of biomolecules such as proteins is usually performed using colorimetric assays or fluorescently labeled secondary antibodies, which requires complex optical detection equipment such as fluorescence microscopy or spectrophotometry. Based on this device, we developed a label-free method to quickly and accurately detect and quantify nano-bioparticles using only a standard laboratory microscope or a mobile phone without the use of any fluorescent labels.

Epilepsy affects over 70 million people globally, with sudden and unpredictable seizures that can cause irreversible neurological damage or death if not treated promptly. Current monitoring methods like scalp EEG lack stability and continuous real-time capability, while systemic drug therapies suffer from slow onset and side effects. To overcome these limitations, we developed an integrated closed-loop system for real-time monitoring and automated intervention. We first established a cross-scale research pathway from cellular to animal levels. A microelectrode with platinum resistor array (MEPRA) chip was designed to simultaneously monitor electrophysiological signals (10 μV resolution) and temperature (0.05 °C resolution), providing a dual-mode basis for seizure detection. An upgraded IDT-MEPRA chip incorporating ultrasound emission revealed the synergistic mechanism of low-intensity pulsed ultrasound (LIPUS) with valproate-loaded nanobubbles (VPA@NBs), confirming that ultrasound cavitation enhances mechanical energy and therapeutic efficacy. Building on these insights, we created a miniaturized implantable epilepsy therapy patch (ETP). It combines a 3D-printed conductive microneedle array for cortical recording and local drug delivery with a LIPUS-VPA@NBs intervention strategy. Using an adaptive algorithm based on real-time ECoG and temperature feedback, the ETP triggers immediate treatment upon detecting abnormalities, suppressing seizures while dynamically preventing overstimulation.

This study explores, for the first time, the efficacy of a continuous-flow magnetophoretic chip for the ultrasensitive digital immunoassay of p-tau217 biomarker. By utilizing ultrabright quantum dot nanospheres (QDNs) for direct single-molecule immunoassay rather than enzyme-substrate amplification, we developed an enzyme-free, oil-free platform that eliminates chamber compartmentalization, simplifying the assay system and reducing costs. However, conventional magnetic separation proved insufficient for eliminating background interference arising from excess free quantum dot nanospheres and nonspecific adsorption. To overcome this, a device is proposed with tunable magnetic and hydrodynamic forces to ensure high-signal to noise ratio. Furthermore, an AI algorithm is developed fusing a bright-field magnetic bead recognition mask with adaptive attention-based dark-field fluorescence detection, enabling high-precision single-molecule recognition. Notably, this AI-driven approach achieves the unification of digital and analog modes, thereby significantly expanding the dynamic range. The integrated platform enabled the detection of p-tau217 down to 149.23 fg/mL in serum samples. In clinical assays, the system successfully differentiated AD patients from healthy controls, showing a 7.12-fold difference in serum concentration (2.144 pg/mL vs. 0.301 pg/mL). With an AUC of 0.9889, these results validate the diagnostic utility of p-tau217 and demonstrate the potential of our proposed method for precise AD diagnosis.

​

​

​