Stimulated Raman scattering (SRS) microscopy has emerged as a powerful chemical imaging platform that translates molecular bond vibrations into rapid visualization of biomolecules, metabolites, and drugs. As SRS moves from proof-of-concept imaging toward live cell analysis, thick tissue interrogation, biochemical readout, and computational histopathology, its further development is increasingly limited by coupled trade-offs among sensitivity, acquisition speed, spectral bandwidth, penetration depth, spatial resolution, and interpretability. This review summarizes recent advances in enhancement strategies for SRS microscopy through four interconnected dimensions: signal, acquisition, light-field, and information engineering. Signal enhancement improves the detectability of weak Raman responses, acquisition enhancement increases imaging throughput and spectral accessibility, light-field enhancement extends SRS toward deep, volumetric, and super-resolution imaging, and information enhancement expands SRS from chemical mapping to multidimensional molecular analysis. Through this framework, we highlight how these enhancement strategies are enabling SRS microscopy to evolve into a more sensitive, faster, deeper, and more interpretable molecular imaging platform for biology, pathology, and precision medicine.
Photobiomodulation (PBM) is a promising therapeutic intervention for age-related neurodegenerative diseases. It has recently been discovered that the meningeal lymphatic vessels (MLVs) are targets for PBM, which mediates a PBM-related increase in brain drainage and clearance. However, with age, the MLV function significantly declines, which may impact the effectiveness of PBM for treating the aging brain. Age-related differences in sensitivity to PBM remain poorly understood, which motivated our research. The studies were conducted on C57BL/6 male mice at key age stages, including young (3 months), middle-aged (6 months), transitional (12 months), aging (18 months), and old (24 months) age, with a focus on examining the 10-day course of PBM (LED 1050[Formula: see text]nm in pulse mode, 30[Formula: see text]J/cm 2 for one PBM exposure) effects on the lymphatic clearance of the toxic metabolite amyloid beta (A[Formula: see text]) from the brain assessed by ex vivo confocal analysis of intensity of fluorescent signal from A[Formula: see text] in the brain, its meninges and in the deep cervical lymph nodes as well as by immunohistochemical assay of A[Formula: see text]1–42 levels in the tested tissues. The results clearly demonstrate age-related changes in the MLV network and the effects of PBM on A[Formula: see text] clearance. Beginning at 12 months, a gradual proliferation of the MLV lymphatic endothelium was observed, with a dramatic decrease in A[Formula: see text] clearance in 18- and especially 24-month-old mice, suggesting that these changes lead to a decline in the MLV function of A[Formula: see text] removal from the brain with age. Indeed, we observed an increase in A[Formula: see text] levels in the brain and meninges in 18- and, most significantly, 24-month-old mice. PBM effectively reduces high A[Formula: see text] levels in the brain and meninges in 18-month-old mice to the level of 12-month-old mice, but no PBM effects are observed in 24-month-old mice. The obtained results highlight parallel processes that may be interrelated, i.e., aging of MLVs and the effectiveness of PBM stimulation of their function in the form of A[Formula: see text] clearance. These findings provide an important informative basis for improving the clinical guidance for PBM use for the treatment of brain diseases associated with amyloidosis and contribute to a better understanding of the limitations of PBM in elderly patients.
Surface-Enhanced Raman Spectroscopy (SERS) is an ultrasensitive molecular detection technique whose signal amplification arises from two principal mechanisms: electromagnetic enhancement (EM) and chemical enhancement (CM). In recent years, two-dimensional (2D) materials, including graphene, transition metal dichalcogenides (TMDs), hexagonal boron nitride (h-BN), and MXenes, have emerged as highly promising candidates for SERS substrate fabrication, owing to their distinctive atomic-scale thickness, large specific surface area, tunable electronic structures, and exceptional chemical stability. In this review, we systematically survey the latest advances in SERS research involving both standalone 2D material substrates and their hybrid systems with metallic nanostructures, with particular emphasis on the physical underpinnings of EM and CM enhancement mechanisms. Special attention is devoted to several frontier topics, including the coupling physics between plasmonic near-fields and the electronic systems of 2D materials, hot-carrier injection dynamics, exciton–plasmon strong coupling, and phonon-assisted enhancement. Furthermore, the application of theoretical computational approaches, such as finite-difference time-domain (FDTD) simulations and density functional theory (DFT) calculations in mechanistic elucidation is discussed. Finally, key challenges confronting the field and prospective future directions are outlined.
Chronic wounds represent a significant burden on global healthcare systems due to their prolonged healing duration, high rate of complications, and considerable economic cost. Accurately assessing the spatiotemporal changes in tissue structure, microcirculation, and collagen remodeling during the wound healing process is critical for guiding effective clinical interventions. Optical Coherence Tomography (OCT), an imaging technique based on low-coherence interferometry, enables high-resolution, noninvasive evaluation of structural and functional changes in soft tissue, thereby providing a precise method for daily monitoring of chronic wounds. OCT technology has evolved from early time-domain OCT (TD-OCT) to more advanced frequency-domain OCT (FD-OCT). Specialized implementations such as ultrahigh-resolution OCT (UHR-OCT), OCT angiography (OCTA), dynamic OCT (D-OCT), and polarization-sensitive OCT (PS-OCT) have further broadened its functional applications. Although OCT has been applied in wound assessment, the relative advantages and specific applicability of different OCT variants across various wound healing phases have not been systematically summarized. Therefore, this review presents a comprehensive overview of OCT technology and its applications in chronic wound healing, detailing the technical characteristics of various OCT variants and offering guidance for researchers in applying OCT technology. Future research should focus on optimizing the trade-off between imaging resolution and penetration depth. Integration with complementary imaging technologies and deep learning algorithms is expected to enable more comprehensive, automated wound evaluation and accurate healing trajectory prediction.
Accurate differentiation between optic disc edema (ODE) and other optic disc anomalies, such as pseudopapilledema (PPE), is vital yet challenging in neuro-ophthalmology due to their highly overlapping visual features on 2D fundus images. This paper proposes the D 3 -Mamba (Dual-branch, Dual-difference, Dual-domain Mamba), a novel hybrid architecture designed to capture fine-grained discriminative features for robust classification. First, we introduce the Hierarchical Mamba Serialization Strategies (HMSS) to tailor scanning mechanisms to different feature resolutions. Featuring an adaptive spiral scan mechanism, this strategy is designed to explicitly model the concentric structure of the optic disc, mimicking the focused visual path of clinicians. Additionally, a History-aware Dual-Difference Interaction (HDDI) module is incorporated to facilitate feature interaction between the CNN and Mamba branches via a recursive history update mechanism, ensuring that global context and local details are effectively fused. Finally, the Dual-Domain Aggregation (DDA) network integrates spatial features with frequency-domain analysis to compensate for the information loss inherent in spatial downsampling, thereby enriching the representation of complex pathological patterns. The model was evaluated on the Papilledema2018 and the Eye Disease Image Dataset. Experimental results demonstrate that D 3 -Mamba achieves robust diagnostic performance compared to mainstream deep learning architectures. Notably, the model exhibited high sensitivity for detecting ODE across both datasets, highlighting its potential utility in reducing clinical misdiagnosis risks. This research presents a promising approach for automated primary care screening, offering a fair balance between diagnostic reliability and computational efficiency.
Rheumatoid arthritis (RA) is a chronic autoimmune disease pathologically defined by the formation of pannus, a highly vascularized tissue with pronounced proliferative, invasive, and destructive properties. With the molecular mechanisms of angiogenesis becoming increasingly elucidated, targeted interventions are now being developed to disrupt the pro-angiogenic cellular network within the pannus, which is orchestrated by activated endothelial cells, invasive fibroblasts, and infiltrating immune cells. From the perspective of early detection, sensitive noninvasive imaging techniques, including photoacoustic, ultrasound, and nuclear imaging, enable visualization of angiogenesis in RA joints, providing a critical window for timely diagnosis and antiangiogenic treatment. This review synthesizes recent advances in RA angiogenesis research, particularly focusing on molecular mechanisms, novel targeted therapeutic strategies, and advanced imaging techniques for in vivo assessment. Collectively, these developments highlight the paradigm of image-guided antiangiogenic theranostics, which integrates real-time visualization of neovascularization with nanomaterial-based targeted therapeutic intervention to offer a refined approach for early RA management.
Early and ultra-sensitive detection of cancer biomarkers is essential for effective diagnosis and timely clinical intervention. In this study, we propose a quantum-enhanced graphene-plasmonic metasurface biosensor designed for multi-cancer biomarker detection with femtomolar-level sensitivity. The hybrid architecture integrates a tunable graphene monolayer with a metasurface composed of square-ring and nanowire resonators, enabling strong electromagnetic field confinement and quantum-corrected plasmonic enhancement in nanoscale gaps. Graphene functionalized with specific antibodies and aptamer probes facilitates selective detection of clinically relevant biomarkers, including prostate specific antigen (PSA), carcinoembryonic antigen (CEA), alpha-fetoprotein (AFP), and circulating tumor DNA (ctDNA). Finite-difference time-domain simulations and experimental validation using microfluidic delivery establish the device's high sensitivity, narrow resonance linewidths, and dynamic spectral tunability. Real-time resonance shifts were analyzed using an AI-assisted classification framework incorporating machine learning models and ROC-AUC validation to achieve robust detection accuracy. Experimental characterization using SEM, AFM, and Raman spectroscopy confirms fabrication quality and graphene integrity. Comparative analysis demonstrates that the quantum-corrected metasurface outperforms classical plasmonic designs in sensitivity and figure-of-merit. These results position the proposed platform as a promising foundation for universal multi-cancer screening systems with high precision and clinical translational potential. Even though the investigated biomarkers are neither adrenal-cancer-specific, they are commonly utilized in clinical differential diagnosis of adrenal malignancies, where biomarker expression is usually low and heterogeneous. The innovative platform is thus placed as a multi-purpose ultra-sensitive detecting architecture that can equally be used in the screening of adrenal cancer and a larger multi-cancer diagnostic.
Multiphoton microscopy (MPM) is a rapidly advancing, noninvasive optical imaging technique that enables subcellular-resolution imaging of the skin. It has shown strong potential for in vivo detection of skin cancers and other dermatologic diseases. In this paper, we present an up-to-date review of recent progress in MPM for in vivo skin imaging and dermatologic interventions, with emphasis on its capabilities and current challenges. Topics include: principles and advances in MPM instrumentation; three-dimensional (3D) volumetric imaging with large field of view (FOV) and subcellular resolution; MPM-based diagnosis of skin cancer and skin diseases; multiphoton-absorption-based photothermolysis; longitudinal imaging of cellular dynamics during dermatologic interventions; and recent developments in artificial intelligence for MPM image processing, including image restoration, segmentation, virtual histology, and skin disease classification.
Micro- and nanoplastics (MNPs) pose increasing risks to both human health and ecological systems, driving an urgent need for sensitive and reliable analytical methods. Among existing detection techniques, Raman-based approaches are particularly attractive due to their label-free, chemically specific, and in situ analysis with high spatial resolution. Despite numerous Raman-based MNP studies, recent advances have not been systematically reviewed, especially the applications of surface-enhanced Raman scattering (SERS) spectroscopy and coherent Raman scattering (CRS) microscopy in environmental analysis. Here, we show that the primary goal of SERS studies has shifted from maximizing detection sensitivity toward developing robust, matrix-tolerant methodologies for real environmental samples. Additionally, new analytical strategies have emerged that integrate the ultrasensitive identification power of SERS with the rapid quantification and three-dimensional imaging offered by CRS microscopy. Representative experiments have demonstrated the feasibility of this multimodal framework, providing a promising route for investigating the fate and environmental impact of MNP pollution.
3D Structured Illumination Microscopy (3D-SIM) enables high-resolution volumetric imaging of subcellular structures, but low signal-to-noise ratio (SNR) conditions can cause noise amplification, artifacts, and structural loss during reconstruction. Existing learning-based SIM methods often process axial sections independently, which limits their ability to use volumetric continuity. To address this limitation, we propose a High-Frequency Enhanced Dense-Residual-Connected Transformer (HF-DRCT) for low-SNR 3D-SIM reconstruction. HF-DRCT combines a multi-slice input strategy, a High-Frequency Enhancement Block, and a structure-aware composite loss to improve adjacent-slice context modeling, frequency-aware feature modulation, and structural detail preservation. High-SNR Wiener reconstructions are used as supervised reference labels rather than independently measured optical ground truth, so PSNR and SSIM indicate similarity to these references. Simulated experiments on multiple biological structures show that HF-DRCT achieves higher reference-based quantitative performance than Wiener deconvolution and existing deep learning baselines. Real low-SNR experiments further suggest that HF-DRCT suppresses background artifacts and preserves more coherent biological morphology in practical 3D-SIM reconstruction.
Cisplatin is a commonly used chemotherapeutic agent for bladder cancer, but drug resistance remains a major challenge. Previous studies have suggested that lipid metabolic reprogramming is closely associated with tumor resistance. However, the relationship between cisplatin resistance and lipid metabolism in bladder cancer remains unclear. To address this, our study performed quantitative lipid analysis leveraging stimulated Raman scattering (SRS) microscopy integrated with Raman probes, which offers high chemical specificity and spatial resolution. This profiling was conducted across cisplatin-sensitive (BIU-87) and resistant (BIU-87-CisR) bladder cancer cell lines, as well as in patient-derived primary cancer cells. Results showed that lipid accumulation, represented by lipid droplets (LDs) area fraction per cell, was significantly higher in BIU-87-CisR cells than in BIU-87 cells, and lipids within the LDs originated primarily from exogenous fatty acid uptake rather than de novo synthesis. Additionally, the lipid unsaturation within the LDs of BIU-87-CisR cells was significantly lower than that of BIU-87 cells. Upon cisplatin stimulation, the LD area fraction per cell increased markedly in BIU-87 cells but remained unchanged in BIU-87-CisR cells. These findings suggest that resistant bladder cancer cells possess higher levels of lipid accumulation, and LD-enriched cells are likely more resistant to chemotherapy. Finally, SRS imaging of patient-derived primary cells showed that the LD area fraction per cell increased with higher tumor stages and grades, and lipid accumulation primarily relied on exogenous fatty acid uptake as well. In conclusion, our findings demonstrate aberrant lipid accumulation in cisplatin-resistant bladder cancer cells, suggesting that LD content could serve as a potential indicator for assessing chemoresistance. This imaging-based approach may offer new avenues for drug sensitivity testing and therapeutic strategy development, although further validation is needed.
Circulating tumor cells (CTCs) are informative biomarkers of hematogenous dissemination and metastatic risk, but conventional assays still capture only a limited and discontinuous fraction of the circulation. In vivo photoacoustic flow cytometry (PAFC) offers a label-free route for monitoring melanoma CTCs, yet most analyses still treat PAFC primarily as a thresholding and counting tool. Here we used an end-to-end simulation framework to determine how vascular position, cell size, cell-vessel coupling, cluster morphology, and sampling conditions reshape PAFC waveforms across both the single-pulse and event domains. Position mainly controlled pulse visibility and defined a robust central operating window of about -4.2 to 4.2 mu m, cluster morphology broadened events from about 2.8ms in the compact state to 5.6ms in the single-file state, cell size most clearly altered pulse width and N-wave structure, and contrast-optimal coupled states did not coincide with peak-optimal states. Flow speed together with beam width then determined whether the event sequence was sampled densely enough for reliable decoding. These results convert PAFC waveforms from simple count signals into interpretable signatures of melanoma CTC state and provide a mechanistic basis for longitudinal monitoring and diagnostic support.
Diffuse correlation spectroscopy (DCS) is a critical noninvasive technique for cerebral blood flow monitoring; its accuracy is frequently impaired by extra-cerebral or superficial layer interference. In this study, we propose a residual-corrected deep-learning framework specifically designed to stabilize cerebral blood flow estimation. By incorporating a residual learning architecture, the model effectively captures the subtle deviations between theoretical analytical solutions and experimental measurements, thereby improving the robustness of feature extraction from autocorrelation signals. We validated the proposed framework using a two-layer blood flow phantom dataset across three source-detector separations (SDS) of 1, 2 and 3 cm. Ablation experiments were further performed to verify the effectiveness of the key modules, including dual-branch architecture, attention mechanism, residual correction module, and three-stage training strategy. The experimental results demonstrate that the proposed method significantly reduces estimation errors and enhances the reliability of deep-layer blood flow quantification.
Imaging flow cytometry is rapidly developing toward higher spatial resolution and enhanced capability for resolving subcellular structures, which makes such systems increasingly dependent on high-numerical-aperture objectives for higher-resolution imaging, thereby imposing higher requirements on the stability of three-dimensional hydrodynamic focusing, the quality of optical interfaces, and the structural compatibility of microfluidic chips. In this work, we propose a low-cost fabrication method for glass microfluidic chips for imaging flow cytometry, based on laser processing and a two-step thermocompression bonding strategy. By separating microchannel pre-bonding from overall encapsulation, this method preserves the structural integrity of the microchannels while constructing an optical observation window compatible with high-numerical-aperture objectives, thereby meeting the adaptation requirements of mainstream microscope objectives in terms of working distance and mechanical interfacing. Experimental results show that the optical detection region of the fabricated chip exhibits nanoscale surface quality and good optical transmission performance. Flow imaging experiments further demonstrate that the chip achieved stable three-dimensional hydrodynamic focusing and high-quality fluorescence imaging under high-speed flow conditions, and that the resulting signals exhibited a narrow distribution with low fluctuation, indicating good detection stability and imaging reliability. The study shows that the proposed two-step thermocompression bonding process achieves a balance among the optical performance, structural integrity, and manufacturability of glass microfluidic chips, providing a feasible solution for the fabrication of high-throughput, low-cost, and biosafe chips for imaging flow cytometry.
Polycystic ovary syndrome (PCOS) is a common endocrine disorder characterized by ovarian dysfunction and hyperandrogenism. Research into its dynamic pathological mechanisms is limited by the invasive nature of conventional histological methods, which preclude monitoring in live models. The zebrafish has emerged as a valuable model for PCOS due to its physiological similarities to humans. This study explores the application of spectral-domain optical coherence tomography (SD-OCT) as a high-resolution imaging tool for the assessment of PCOS progression and treatment in zebrafish. A testosterone-induced PCOS model was established in adult female zebrafish. The therapeutic efficacy of two doses of Metformin was evaluated. The core methodology involved in vivo SD-OCT imaging of ovarian follicles at multiple time points over 72h. These morphological changes observed via SD-OCT were consistently corroborated by histology. Metformin treatment, particularly at the high dose (100mg/L), demonstrated a pronounced therapeutic effect, effectively reversing the testosterone-induced follicular arrest and normalizing ovarian testosterone levels and gonadosomatic index (GSI). This study validates SD-OCT as a powerful and reliable modality for non-invasive monitoring of PCOS in a live zebrafish model.
Folic acid (FA) tablets are widely used by pregnant women during the first few weeks of pregnancy. The qualitative analysis and classification of FA tablets from different companies of China and Pakistan are reported through a comparative laser-induced breakdown spectroscopy (LIBS) machine-learning framework. LIBS examination recognized many elements like calcium, magnesium, sodium, and potassium present in these tablets. Thirteen machine learning (ML) algorithms from the classification learner app were used for FA tablet classification. This research focused on Pakistani and Chinese tablet samples within the terms of classification. The usage of supervised ML classification models gives significant results with 91.7% and 91.6% accuracy by two models, efficient logistic regression and a wide neural network, respectively. It shows that ML algorithms are efficient at classifying FA tablets.
Optical-resolution photoacoustic microscopy is an emerging imaging technique that enables high-resolution visualization of biological structures. However, the presence of system noise significantly degrades image quality. As a key post-processing step in photoacoustic imaging, denoising based on deep learning is limited by the small size of available training datasets, which restricts its application in high-precision imaging of in vivo tissues under complex noise conditions. To address the issue, we built a photoacoustic microscopy and proposed a deconvolution-enhanced self-supervised learning framework for denoising. The system integrates the network, enabling training on noisy measurements and operation with just a single photoacoustic image. The method generates photoacoustic microscopy training pairs using a directional neighborhood random down-sampling strategy and employs deconvolution to preserve vascular continuity and noise independence. A self-constrained loss function is designed to improve data utilization and improve the effectiveness of noise suppression. Experimental results on phantom, mouse ear, and mouse brain demonstrate that the proposed system and method effectively suppress intrinsic noise and reveal clearer structural details. Compared with the raw data, the average signal-to-noise ratio, contrast-to-noise ratio, and Brenner gradient index across these datasets increase by approximately 20%, 40%, and 30%, respectively. The proposed method can achieve denoising using only noisy photoacoustic data. The method and system demonstrate strong performance in high-fidelity photoacoustic imaging, showing great potential for biomedical applications.
Different academic disciplines involve different types of knowledge and learning processes, which may be reflected in variation in prefrontal hemodynamic responses during learning. To simulate real learning scenarios, we used multidisciplinary instructional videos and functional near-infrared spectroscopy(fNIRS) to examine changes in prefrontal hemodynamics. Thirty-five participants completed video learning tasks classified according to the soft-hard dimension of Biglan's framework. A General Linear Model (GLM) was applied to estimate beta coefficients for eight prefrontal channels. Results revealed significant differences between soft and hard disciplines in beta values in channels corresponding primarily to Brodmann areas BA10(frontopolar cortex), BA11(orbitofrontal cortex), and BA46(dorsolateral prefrontal cortex). Specifically, soft disciplines (social sciences and humanities) showed significantly larger beta estimates in channels 5 and 6(BA10/11) compared with hard disciplines (natural sciences and applied sciences). Within the same disciplinary context, participants with post-test accuracy below 70% exhibited significantly higher beta values in multiple BA10 channels than participants with higher accuracy, indicating greater prefrontal engagement under equivalent task demands. These findings describe differences across disciplines and across performance levels in GLM prefrontal hemodynamic responses during naturalistic video learning, providing quantitative evidence that disciplinary context and learner proficiency are associated with variation in prefrontal engagement.
Diffuse correlation spectroscopy (DCS) is a powerful optical technique for noninvasive quantification of deep-tissue microvascular blood flow quantification. Conventional DCS systems are bulky and rely on external computing devices, limiting their point-of-care applicability. In this study, we implemented the proposed DCS system on a Xilinx Zynq-7000 system-on-chip (SoC) via hardware-software co-design approach. The proposed approach employed the programmable logic (PL) for a multi-tau digital autocorrelator and the processing system (PS) for real-time blood flow index (BFI) extraction via the Nelder-Mead algorithm. Arterial occlusion and breath-holding in vivo tests demonstrated strong consistency and high correlation with a conventional PC-based DCS system. Additionally, the proposed system significantly reduces instrumentation size and cost, enabling portable assessment of tissue microcirculation.
Nonalcoholic fatty liver disease (NAFLD) is one of the most prevalent chronic liver diseases globally. Hepatic steatosis, characterized by the accumulation of lipid droplets (LDs), is a hallmark of NAFLD. Traditional pathology and specific fluorescent dye imaging have difficulty in observing LDs in their natural state within hepatocytes rapidly. In this study, a customized dynamic full-field optical coherence tomography (D-FFOCT) system was applied for label-free visualization of LDs and rapid ex vivo assessment of NAFLD. Two groups of mice fed the methionine choline-deficient (MCD) diet for one week and two weeks, respectively, underwent dual-mode FF-OCT imaging and comparative quantitative analysis. The subcellular structures - including nuclei and small LDs - were clearly observed. Quantitative analysis of D-FFOCT results revealed significant quantitative profiles in LD characteristics across various NAFLD stages. Furthermore, imaging of a nonalcoholic steatohepatitis (NASH) mouse model fed a high-fat diet for 14 weeks provided insights into the identification of inflammatory cells via D-FFOCT. These results indicate that D-FFOCT can effectively resolve the subcellular structure and quantify LDs in NAFLD models across varying severities, thereby demonstrating its potential for rapid NAFLD assessment and pharmaceutical development applications.
