This PDF file contains the front matter associated with SPIE Proceedings Volume 13507, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

Analyzing the vascular network structure of the brain is crucial for understanding brain function and preventing and treating cerebrovascular diseases. By studying indicators such as the length density and branching point density of the brain’s microvascular network, we can better understand the brain’s vascular network and provide a scientific basis for cerebrovascular disease research. Traditional methods rely on statistical analysis of isolated brain regions, which fail to capture the spatial distribution characteristics of micro-vessels across the entire brain and overlook spatial information from different brain regions. To address this, we propose a computational pipeline for the spatial distribution of vascular density in the brain. First, we use a deep learning model to segment brain vascular images from HD-fMOST imaging of mouse brains. Then, we perform skeletonization on the 3D vascular segmentation images and obtain vessel diameter information through distance transformation. Finally, we filter capillaries based on their diameter and calculate the feature parameters of these capillaries to acquire vascular parameter distributions across the entire brain. The entire process is enhanced by MPI parallel processing and multithreading techniques to improve image processing efficiency. Results indicate that our proposed method can capture the spatial distribution characteristics of microvascular density across the whole brain by integrating multiple datasets. This method provides a powerful tool for comprehensive research on brain vasculature.

Minimal Hepatic Encephalopathy (MHE) is a potential neurological complication in patients with liver cirrhosis. Here, we propose the use of Photoacoustic Imaging (PAI) for non-invasive assessment of the structure and function of the liver in MHE. Our PAI results showed that the hepatic vascular structure in the acute MHE mice model did not undergo significant changes, but the hemoglobin oxygen saturation content was markedly decreased. In contrast, in the chronic MHE model, the hepatic vasculature of the mice was disorganized, and the architecture of the hepatic lobule was not intact, and the liver function of hepatic metabolism was also significantly impaired. This study has demonstrated that PAI holds immense potential to assess liver structure and function in MHE disease.

With fast developments of computational power and algorithms, deep learning has made breakthroughs and been applied in many fields. However, generalization remains to be a major challenge, and the limited generalization capability severely constrains applications of deep learning in practice. The hallucinations issue is another unresolved conundrum faced by deep learning and large models. By leveraging a physical model of imaging through scattering media, we studied the lack of generalization to datasets and system response functions in deep learning respectively, identified their causes, and proposed universal solutions. The research also provides an explanation for hallucinations. In general, it enhances the interpretability of deep learning from a physics-based perspective. It will pave a way for direct interaction between deep learning and the physical world, facilitating the transition of deep learning from a demo model to a practical tool.

Elasticity is one of the fundamental properties of materials. It is an important diagnostic parameter to investigate physiological dysfunctions in biological tissues. This paper presents a novel method to accurately measure the elastic by Rayleigh wave tracing on the surface of phantoms using laser profilometer with airpuff excitation. We developed a laser profilometer elastography system, including the optical system, timing sequence, and control software. The elastics of agarose phantoms with 0.6%, 0.8%, 1.2% and 1.5% concentration were measured. The method was validated by comparing it with the results obtained by using a laser speckle elastography. To the best of our knowledge, this is the first elastic modality based on a laser profilometer. This method can be used for online elasticity measurement of biological.

Raman spectroscopy is an important analytical tool, but the weak intensity of Raman signals caused by Stokes shifts limits its practical applications. To enhance Raman signals, techniques such as Surface-Enhanced Raman Spectroscopy (SERS) and Stimulated Raman Scattering (SRS) are commonly used, but these methods rely on external enhancement agents and complex equipment. In this study, we propose a novel deep learning-based method that can directly enhance Raman signals without dependence on external agents or complicated setups. Experimental validation using Rhodamine 6G and Crystal Violet demonstrated significant signal enhancement. To verify the accuracy of the enhancement, we calculated the Root Mean Square Error (RMSE) and Normalized Mean Square Error (NMSE) for Rhodamine 6G, which were 534.2 and 0.21%, respectively, achieving a 209.1-fold signal enhancement. This approach demonstrates potential for directly amplifying weak Raman signals, offering new insights for advancing Raman spectroscopy applications across various fields.

A multifunctional fluorescence imaging system based on LabVIEW has been developed, which can perform fluorescence intensity scanning imaging, and fluorescence lifetime imaging (FLIM). The system utilizes software based on LabVIEW to control a Data Acquisition (DAQ) card, Time-Correlated Single Photon Counting (TCSPC) card, and galvanometer to achieve point-scan imaging. The FLIM module integrates Dynamic Link Libraries (DLLs) for galvanometer control, allowing simultaneous management of the galvanometer while outputting clock synchronization signals from the DAQ card as trigger signals for FLIM. This feature facilitates TCSPC FLIM. In addition, by integrating related DLLs and introducing a 3D sample stage, the system can manipulate the 3D sample stage to perform fluorescence imaging at different depths, thereby generating 3D images of the samples and achieving 3D scanning imaging. Ultimately, the system seamlessly integrates LabVIEW software, confocal scanning system, DAQ card, 3D sample stage, and TCSPC card into a single platform for multifunctional fluorescence imaging. All imaging operations are performed within a single software interface, which outputs image data for further processing. Compared to commercial FLIM systems that run on multiple independent software platforms, our system minimizes unexpected errors and delays caused by inter-software communication and file transfer, successfully achieving multifunctional imaging.

This study explores a new method for identifying prohibited drugs using X-ray Absorption Spectroscopy (XAS) detection and machine learning algorithms. A laboratory-developed scientific equipment was employed to obtain x-ray absorption spectra of ten different chemical substances, including various isomers of prohibited drugs. Principal Component Analysis (PCA) was then applied to extract the spectral features, minimizing data redundancy. Subsequently, the Extreme Learning Machine (ELM) combined with the Sparrow Search Algorithm (SSA) was utilized for analysis. The results demonstrate that this method can accurately identify prohibited drugs, even excelling in automatically identifying isomers. This research offers a novel and promising technique for quick non-destructive drug detection.

Polarization imaging is widely applied in the biomedical field. Mueller matrix imaging is an effective way to extract optical properties of biomedical samples. The mainstream technology is to use Division of Focus Plane (DoFP) system to obtain Mueller matrix images, because Mueller matrix microscope based on dual Division of Focal Plane (DOFP) polarimeters can realize fast full polarization imaging. However, DoFP system has a trade-off between resolution and imaging speed. Although using high numerical aperture objectives can improve resolution, it leads to a decrease of the field of view. To address this issue, we propose a deep learning network named MMRE-GAN (Mueller Matrix Resolution Enhancement GAN), aimed at improving the resolution of Mueller matrix images while also keeping the big field of view of low numerical aperture objectives. Our method does not require paired images and effectively improves the resolution of Mueller matrix images through Generative Adversarial Networks (GANs). In detail, we used a dual Division of Focal Plane (DoFP) polarimeters-based full Mueller Matrix Microscope (DoFPs-MMM) to quickly get the Mueller matrix images of thyroid slices. Then we used the dataset to validate the effectiveness of MMRE-GAN. The results show that our method successfully improves the resolution of Mueller matrix images while keeping the field of view in 4x objective, meaning a shortened scanning time for the whole slides. And our method also enhances the details and signal-to-noise ratio of imaging. It provides support for the application of DoFPs-MMM in biomedical imaging.

As plastic pollution escalates, microplastics, particles less than 5 mm, have attracted significant attention due to their origins in everyday plastic waste, microbeads in personal care products, and bottled water packaging, posing threats to ecosystems and health. Polystyrene (PS), one type of widely used plastics, affect human health via absorption, ingestion, and dermal contact. In this study, we used microscopy methods to monitor the invasion progress of PS microplastics into KYSE-150 cells and to study the following influence of this invasion on the cellular status. Firstly, the mitochondrial probe Rhodamine 123 was used to label KYSE-150 cells, which indicated the mitochondrial status during the exposure of PS. Then the KYSE-150 cells were treated with fluorescent PS microspheres (approximately 100 nm in diameter) for one day. During this process, cells were observed by using both a confocal microscopy system and a Fluorescence Lifetime Imaging Microscopy (FLIM) system. Results showed that the microplastics appeared in the cell cytoplasm within one hour. The amount of microplastics within the cells kept increased till 24 hours. Images of incubated cells at each time point (1 h, 3 h, 6 h or 24 h) were collected by using the FLIM system and then analyzed with the software SPCImage. The data showed that the fluorescence lifetime values of R123 in the mitochondria of microplastic treated cells increased according to the incubation time. By contrast, the lifetime values in the control group (cells without PS) kept constant during the same period. These results showed that changes of lifetime values might be due to the accumulation of microplastics within the cells. Briefly, this study suggested that FLIM could be a rapid, non-destructive, and sensitive method suitable to monitor the invasion of microplastics into live cells, providing appropriate data to the toxicity assessment of microplastics.

This study focuses on the synergistic effect of the development of oil glands in kumquats and the synthesis and accumulation of secondary metabolites. Through the distribution of oil glands in kumquat peels, it guides the cultivation of oily varieties, removes the bitter taste of peels, and improves the sustainable development of the kumquat industry. This study used Optical Coherence Tomography (OCT) technology to perform non-destructive scanning of kumquat peel, obtaining high-resolution images of its internal oil glands. Using the U-Net model of Convolutional Neural Network (CNN) to accurately segment OCT images of oil glands, extract the boundary and morphological features of oil glands, and perform 3D reconstruction on the segmented oil glands. Experiments have shown that the cross-sectional image of oil gland cells approximates an elliptical model, which can collect geometric parameters such as the long axis, short axis, and area of the oil gland. Three-dimensional reconstruction of the segmented oil gland reveals its cylindrical structure. The results indicate that the morphological parameters and characteristic distribution of oil glands will provide comprehensive, accurate, and efficient data support for the quality evaluation and grading of citrus, which will help optimize the quality control and scientific grading of kumquats, and provide reference for the sustainable development of kumquat industry.

Caffeine is an alkaloidal stimulant prevalently in food additives. The widespread presence of caffeine in a wide range of beverages has led to increased consumption, commonly exceeding safe levels. In this study, about 30 beverages available in the Chinese market, were analyzed using a 785 nm Raman spectrometer. The raw spectral data were preprocessed by noise reduction and baseline correction. The qualitative model was established using Principal Component Analysis (PCA) and Partial Least Squares Discriminant Analysis (PLS-DA), and the classification discrimination was carried out using K-Nearest Neighbor algorithm (KNN) and Support Vector Machine algorithms (SVM). Especially, the two models could identify “Spirte” (without caffeine) from other beverages. It was demonstrated that Raman spectroscopy combined with machine learning can provide non-destructive, fast and efficient qualitative detection of caffeine in liquid beverages.

Photoacoustic Microscopy (PAM) has emerged as a rapidly advancing non-invasive medical imaging method in recent years. However, slow imaging speed has impeded its widespread adoption in clinical applications. In certain scenarios, sparse spatial sampling is essential for PAM so there is a trade-off between spatial resolution and imaging speed. To address this limitation, we propose a frequency domain index based on Cumulative Power Difference (CPD) to determine rapidly the optimal down-sampling factors. In this study, the structural images of mice ears were acquired by the PAM system. Subsequently, the optimal down-sampling factor was determined through CPD analysis of these images via interpolation. Finally, the correlation between cumulative power difference and the image quality loss curve were analyzed. The results indicate that the quality of the reconstructed images decreases with the increasing down-sampling factor. Moreover, the cumulative power difference is an effective tool for rapid assessment of reconstruction image quality degradation due to varying down-sampling factors.

Surface Enhanced Raman Spectroscopy (SERS) has been widely used as a sensitive sensing technology in the medical field because of its unique molecular fingerprint information. In this paper, silver nanospheres (Ag NPs) and silver nano-cube (Ag NC) nanoparticles with different morphologies were constructed based on silver nanomaterials for the purpose of early screening of kidney cancer. By analyzing the SERS characteristics of nanoparticles, it was found that the enhancement effect of Ag NC was greater than that of Ag NPs. SERS detection was performed on the urine of 30 patients with kidney cancer and 30 normal subjects. By analyzing the spectral differences between cancer patients and normal people, the cancer group and normal people were preliminarily distinguished. We analyzed the measured spectral data. The analysis methods mainly included Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) diagnostic algorithms, as well as recursive weighted Partial Least Squares (PLS) and Support Vector Machine (SVM) algorithms. The comparison of the two classification algorithms shows that the classification accuracy of PLS-SVM is 99.13%, sensitivity was 98.67%, specificity was 100%, and AUC value was 1. The classification effect was much higher than PCA-LDA. The results of this Exploratory research show that the combination of Ag NC substrate and PLS-SVM algorithm has greater potential in the pre diagnosis and screening of Kidney cancer.

In the previous research on medical diagnosis of SERS, the most common machine learning method was Principal Component Analysis-Linear Discriminant Analysis (PCA-LDA), but this method had certain limitations and low classification accuracy. Therefore, this study compared surface-enhanced Raman spectroscopy (SERS) of plasma from Prostate Cancer (PC) patients and healthy controls using a deep learning method known as multilayer perceptron (MLP). First of all, the average spectrum and difference spectrum of the two are made to roughly compare the difference between them. Secondly, several Raman positions with obvious characteristics are selected to conduct histogram comparative analysis. Finally, PCA-LDA algorithm and MLP algorithm are respectively used to diagnose the obtained experimental results, and visualization analysis is carried out in combination with figure. The final results showed that although the classification accuracy of PCA-LDA was also good for PC, the diagnostic results of MLP were better than PCA-LDA, and the sensitivity and specificity were also higher than PCA-LDA. This indicated that MLP algorithm combined with SERS improved the accuracy of early detection of prostate cancer and was expected to be expanded in the early screening of other cancers, which had great potential in clinical application.

The biological behavior of tumor cells is closely related to tumor invasion, but the relationship between the spatial distribution of tumor cells at the invasion boundary and prognosis is still unknown. In this study, we based Multiphoton Microscopy (MPM) to image the Tumor-Associated Collagen Signatures (TACS), a biomarker known to be associated with Breast Cancer (BC) prognosis and captured 180μm × 180μm regions of interest (ROI) on its colocalization of hematoxylin and eosin (H&E) images. The deep neural network was used to extract the 51 microscopic features of the spatial distribution of tumor cells from H&E images. The least absolute shrinkage and selection operator (LASSO) regression was used to select the most robust features to build a prognostic score. Our results showed that the spatial distribution of tumor cells was a poor prognostic factor in univariate analysis, and was proved to be an additional histological variable with an independent influence on disease-free survival (DFS) in the multivariate analysis in patients with BC. The area under the receiver operating characteristic curve (AUC) of the prognostic score was 0.784 in the training cohort and 0.717 in the validation cohort. The study suggested that the spatial distribution of tumor cells was an independent prognostic marker for DFS in BC and can provide prognostic information for BC.

In endoscopic OCT systems, the imbalance in torque load during rotational scanning will lead to non-uniform rotation of the distal probe, inevitably resulting in Non-Uniform Rotational Distortion (NURD) in the images. NURD can cause image translation and stretching or compression issues at arbitrary positions in OCT images, leading to misalignment of image information and impeding the implementation of endoscopic OCTA. In high-speed distal scanning OCT system, the instantaneous rotational speed of the micro-motor's metal struts was measured for OCT data resampling, enabling NURD correction, and OCTA was successfully realized. In proximal scanning, NURD is a more serious problem due to torque transmission over longer distances, resulting in asymmetric friction at different positions. In recent years, local block matching (LBM) and improved Features from Accelerated Segment Test (FAST) algorithms were used to solve NURD in B-scan images in proximal controlling OCT system. Cross-sectional OCTA was successfully implemented. In this paper, Global registration and A-line registration were used to correct NURD in continuous rotation and retraction of proximally controlled OCT imaging. Global registration was used to correct extensive NURD in B-scan images and A-line registration was applied for fine correction of minor NURD. Results from microfluidic data collected under the same position and retraction conditions demonstrate the effectiveness of NURD correction, and en face OCTA imaging was realized for the first time in a proximally controlled endoscopic OCT.

Extracellular Matrix (ECM) play a key role in the growth and biomechanical properties of fibrous meningiomas. In particular, tumor cells' growth and invasive behavior may be impacted by the distribution and accumulation of collagen. In this study, we used Multiphoton Microscopy (MPM) for label-free imaging of fibrous meningiomas tissues. MPM based on two-photon fluorescence and second harmonic visualization of fibrous meningioma tumor cell morphological features and collagen fiber distribution around tumor cells. The results demonstrate that the MPM imaging technique enables the visualization of the intricate fibrous network structure formed by collagen surrounding tumor cells. This finding provides new insights into the role of collagen in fibrous meningiomas and further lays a scientific foundation for the development of personalized treatment strategies and novel diagnostic and therapeutic targets.

Pressure ulcer is a common condition for patients who are bedridden or have limited mobility. In this study, two models of pressure ulcer on mouse ears were established, and the blood flow after pressure ulcers was monitored using a Laser Speckle Contrast Imaging (LSCI) system. Maps of blood flow revealed that after a single 1.5-hour pressure ulcer on the mouse ear, the blood flow perfusion at the compressed area gradually recovered over time and essentially returned to normal by the fifth day. Conversely, with continuous daily pressure ulcers for 1.5 hours, the compressed area showed almost no perfusion by the fourth day, and the ulceration was fully formed by the 5th day. Compared to white light imaging, LSCI offers more precise monitoring of blood flow at pressure ulcer sites. The research solves the problem of dynamic functional monitoring of the blood flow velocity changes at spatial resolution. It provides a new technology for dynamic assessment of pressure ulcers and offers a new method for the care and treatment of clinical patients.

Blood can obscure the surgeon’s view during surgery, increasing the difficulty of the procedure. To mitigate its influence, methods such as hemostatic forceps, balloon occlusion, or repeated saline flushing are commonly employed. However, these methods may lead to tissue damage and increase surgical risks. Safely and effectively minimizing the impact of blood remains a significant challenge. This work presents the possibility of directly imaging the target tissue without removing blood. The three-dimensional layered vascular model, comprising blood and the vessel wall, is established, and the Monte Carlo method is employed to simulate the photon propagation within the model. The Gaussian beam is vertically incident on the surface of model with its focus located on the interface of blood and vessel wall. To maximally improve the penetration depth, the wavelength of 1270nm is chosen, and the horizontal distance between the fiber detector and the light source is set to 0.35 cm. First, the accuracy of the model is verified by the exponential decrease of the signal detected with the thickness of the blood layer. Then the possibility of directly imaging is revealed that the light (2.5μW) reflected from the vessel wall 5mm below the blood could be successfully detected when the intensity of the incident light is 50mW. The result might pioneer a new way for clinical diagnosis.

Hypertrophic scars are a type of pathological scar that can cause discomfort, pain, and itching. The assessment of hypertrophic scars depends largely on the clinician’s long-term experience. Computer-aided assessment can greatly improve the efficiency and accuracy of scar assessment. In this paper, we propose a deep neural network-based assessment method for the degree of scar pigmentation and vascularity using real scar images. Firstly, a large amount of hypertrophic scar images were collected to produce a dataset including patients of all ages and the distribution of lesions in different parts of the body. These images were then scored on pigmentation and vascularity using the Vancouver Scar Scale (VSS). Then, the residual network model was proposed to evaluate the degree of scar pigmentation and vascularity according to the hypertrophic scar dataset. The degree of the pigmentation and vascularity were assessed respectively by two residual network models. The experimental results demonstrated that our method provides a reliable tool for the application of a scar diagnosis system.

As strong optical absorption of DNA/RNA to Ultraviolet (UV) light, UV-Photoacoustic Microscopy (UV-PAM) can highlight the cell nuclei of tissues at outstanding imaging contrast without introduction of exogenous contrast agents. However, a short Depth of Focus (DOF) in the traditional UV-PAM results in out-of-focus imaging artifacts when imaging unprocessed tissue specimens that are usually characterized by uneven surface topology, which possibly degrades the histopathological examinations. We propose the wavefront engineering of UV photoacoustic illumination using a liquid crystal-based diffractive optical element (LC-DOE) to precisely manipulate the phases of the incident UV beam, enabling UV-PAM to produce an enlarged DOF (~250 μm) while maintaining subcellular lateral resolution (~1.02 μm) within the extended depth range. The system demonstrates high-resolution volumetric image of a phantoms made from many tungsten filaments. We anticipate that, with the liquid crystal-based diffractive optical element, the UV-PAM accommodates to image the unprocessed tissue samples, and thus potentially assists for non-destructive and label-free histopathology assessments of various disorders.

In view of the problem that traditional fingerprint devices are prone to incomplete fingerprint information due to surface noise pollution (such as scratches and peeling) when collecting fingerprints on the tip of the finger, a method using Optical Coherence Tomography (OCT) to collect subcutaneous internal fingerprints to compensate the external fingerprint information is proposed. In the experiment, a traditional fingerprint device is used to obtain the external fingerprint of the fingertip, and defective fingerprint information was obtained by erasing some information of the fingerprint with image processing technology. Then, OCT is used to collect the subcutaneous internal fingerprint, and the features are compared with the external fingerprints. The subcutaneous internal fingerprint is rotated to align with the external fingerprint. Finally, image fusion is used to complete the information of the external fingerprint on the fingertip. The experimental results show that the missing external fingerprint information can be completely replaced by the subcutaneous internal fingerprint information at the same location.

Optical Coherence Tomography (OCT) enables the non-invasive acquisition of high resolution three-dimensional cross-sectional images at a micrometer scale for (epi-)dermis layer and subcutaneous fat layer of mouse skin. In this paper, we present a method to predict individual weight loss based on a deep-learning algorithm for the automatic segmentation of several layers of mouse skin in OCT image using Convolutional Neural Network (CNN). Our results demonstrate that individual weight loss is related to the subcutaneious fat layer of mice. Then Diacylglycerol, composition containing Diacylglycerol and Xylooligosaccharide, and composition containing Diacylglycerol, medium chain triglycerides, Medium- and long-chain triacylglycerol oil, and Xylooligosaccharide have a small positive effect on weight loss, since that the subcutaneious fat layer become thinning. The potential suggests the subcutaneous fat layer of mice based on OCT combined with CNN could be used as a method to evaluate the weight loss. It was shown that composition containing Diacylglycerol, medium chain triglycerides, Medium- and long-chain triacylglycerol oil, and Xylooligosaccharide had the greatest effect on the adipose layer, which resulted in the greatest degree of adiposity thinning.

The heart is the first organ to develop in zebrafish, and its developmental process is similar to that of the human heart. Zebrafish are an important model for studying human heart disease, and the accurate quantification of zebrafish heartbeats is instrumental in elucidating the intricacies of cardiac function and the pathogenesis of heart diseases, thereby facilitating a comprehensive understanding of potential therapeutic avenues. However, traditional methods such as electrocardiograph make it difficult to measure the heart rate of zebrafish in water. In this study, Optical Coherence Tomography (OCT) is employed to obtain high-precision images of the zebrafish heart area. These images are processed with an optical flow estimation algorithm to capture the motion information of the heart structure. Subsequently, a linear motion trajectory analysis algorithm is applied to fit waveforms to the extracted motion trajectories, enabling accurate computation of the zebrafish's heartbeat. The study demonstrates that this integrated approach enables high-precision, non-invasive measurement of zebrafish heartbeats, which can be used to monitor the changes in the zebrafish heartbeat induced by different anesthetic formulations.

The kidney is an important organ for excreting metabolic wastes and maintaining the stability of the internal environment in the body. The renal tubule is an essential structure for the nephron with reabsorption and excretion. Structural changes in the renal tubules can lead to kidney dysfunction and thus cause renal diseases. With the development of imaging technology, mesoscopic optical imaging can obtain kidney images at cell resolution. It is significant to reconstruct the three-dimensional (3D) morphology of renal tubules from the image to understand renal function and explore the pathogenesis of renal diseases. However, the large volume of high-resolution image data and the extensive spatial distribution of the renal tubule throughout the kidney present significant challenges for 3D reconstruction. To address this, we propose a deep learning-based method for renal tubule reconstruction. First, we propose a deep learning-based method for renal tubule reconstruction. First, we imaged mouse kidneys using High-Definition Fluorescent Micro-Optical Sectioning Tomography (HD-fMOST) to obtain kidney images at cellular resolution. We then employed a U-Net model to segment the renal tubules in two-dimensional (2D) kidney images, producing binary segmentation results. Finally, we performed the connected domain analysis of the segmentation results in 3D space and reconstructed the 3D morphology of all renal tubules. Our method demonstrates efficient and accurate reconstruction of renal tubules in mesoscopic kidney images.

Photodynamic Therapy (PDT) represents a crucial strategy in cancer treatment, exploiting the unique properties of photosensitizers. Silicon phthalocyanine, a second-generation photosensitizer, is known for its outstanding photophysical properties and biocompatibility, while betulinic acid is recognized for its potent anticancer activity. In this study, betulinic acid was conjugated at the axial position of silicon phthalocyanine to synthesize Betulinic Acid-Axially Substituted Silicon Phthalocyanine (Bai-SiPc). The compound was then encapsulated with TPA-mPEG (triphenylamine polyethylene glycol) to form TPA-mPEG@Bai-SiPc nanoparticles. Structural characterization of the synthesized complex was conducted using 1H NMR, FT-IR, and MALDI-TOF-MS. The photodynamic efficacy of TPA-mPEG@Bai-SiPc was assessed in vitro with the MCF-7 breast cancer cell line. Results indicate that TPA-mPEG@Bai-SiPc exhibits excellent biocompatibility and pronounced phototoxicity, positioning it as a highly effective photosensitizer with superior PDT performance against breast cancer.

Rotational Cherenkov-Excited Luminescence Scanned Tomography (RCELST) is an emerging optical imaging technology that visualizes the distribution of luminescent quantum yield within a treated subject. This technology involves collecting luminescence signals resulting from the excitation of luminescence probes by Cherenkov emissions induced by the rotational scanning of MV X-rays. These signals are then mapped into a sinogram for reconstructing the distribution of luminescent quantum yield by neural networks. Vision Transformers (ViTs), an effective deep learning algorithm known for capturing long-distance dependencies, have been applied to medical image reconstruction tasks. However, the large scale of medical images, combined with the quadratic complexity of ViTs, leads to erratic and time-consuming reconstruction performance. Therefore, a more efficient algorithm is essential for reducing reconstruction time while maintaining accuracy. In this study, we propose the Symmetry Vision Mamba (S-VM) to address this challenge, reducing computational time while maintaining high reconstruction accuracy. The S-VM builds on the Vision Mamba, which leverages State Space Models (SSMs) to extract global information from 2D sinogram signals. With linear computational complexity, S-VM significantly accelerates the learning process compared to the Transformer algorithms. Additionally, S-VM utilizes a symmetrical encoder architecture, incorporating convolutional stems to extract local features and enable multi-scale feature fusion by sharing parameters between the two encoder branches. Training on 10,000 sinogram signals, the S-VM algorithm achieves a peak signal-to-noise ratio (PSNR) of up to 38.67 dB and a Structural Similarity Index Measure (SSIM) of 0.97. Remarkably, these results are achieved in a Floating Point Operations (FLOPs) of 4.7G.

Acne vulgaris, a common dermatological condition affecting the pilosebaceous follicular unit, arises from a complex interplay of factors including hyperseborrhea, Propionibacterium acnes colonization, hyperkeratosis, and inflammation. Existing pharmacodynamic evaluation techniques do not facilitate real-time and non-invasive assessment of acne and its severity within living organism. For this study, the aim was to establish a murine model of acne vulgaris, elucidate its pathological processes, and quantitatively assess the severity of the condition using non-invasive imaging techniques. We employed optical coherence tomography (OCT) in conjunction with microscopy to scrutinize the complete life cycle of acne lesions. microscopy provided insights into the micro-anatomical characteristics of murine skin lesions, while OCT allowed for the assessment of skin structural changes. Our findings indicate that the epidermis undergoes hyperplasia concurrent with acne development, and the subsequent reparative phase is marked by scab formation. Integrating deep learning algorithms enabled the precise quantification of epidermal and scab thickness variations. This novel approach provides a unique perspective for acne assessment, potentially guiding more precise clinical drug selection and improving therapeutic outcomes.

Diffuse Optical Tomography (DOT) is a non-invasive, label-free imaging technique widely used in applications such as breast cancer diagnosis and brain imaging. It allows for the quantitative measurement of tissue functional parameters, including the concentrations of oxyhemoglobin, deoxyhemoglobin, and water. However, the quality of reconstructed images is poor due to light scattering. To address this challenge, one effective strategy is to incorporate anatomical information from high-resolution imaging to guide DOT reconstruction. In this study, a new approach (GRI) is developed to leverage MRI images based on graph structure for DOT reconstruction. And its feasibility and effectiveness were evaluated with numerical simulations. The results indicate that GRI significantly improves the Structural Similarity Index (SSIM) of reconstructed total hemoglobin (HbT) images, outperforming conventional Tikhonov regularization and Direct Regularization Imaging (DRI) method by more than 63% and 82%, respectively. These findings are further supported by experiments using real patient data, which demonstrate GRI’s potential in breast cancer diagnosis.

Flow field analysis is vital for biomedical applications such as drug delivery, dosage optimization, and hemodynamic studies. Conventional optical Particle Image Velocimetry (PIV) struggles with dynamic three-dimensional analysis in turbid media. This study introduces a photoacoustic volumetric particle image velocimetry (PAV-PIV) system to address these challenges, enabling rapid three-dimensional flow field reconstruction in optically opaque environments without complex calibration. Vascular phantoms made from silicone, polyurethane, and agarose were tested, with the optimal material selected based on acoustic impedance contrast. Nickel-coated polystyrene particles served as tracers, uniformly dispersed in the fluid. Particle displacement was measured at 20 ms intervals, allowing for flow field analysis, including velocity and vorticity distributions. Reconstructed three-dimensional photoacoustic signals demonstrated superior imaging quality in turbid media, with detailed flow field characteristics derived from the best-quality images. The system's performance and measurement precision were further evaluated. Results confirm that the PAV-PIV system effectively overcomes the limitations of traditional optical methods, offering robust and dynamic flow field analysis in challenging environments. This approach shows great potential for advancing real-time biomedical applications requiring the analysis of complex flow behaviors.

Silicon phthalocyanine complexes, characterized by two additional axial coordination sites, have gained significant attention as second-generation photosensitizers for Photodynamic Therapy (PDT). In this study, di-(5-methylthiazole-1- ethoxy) axially substituted phthalocyanine (SZ-SiPc) was synthesized and further modified through methyl iodide treatment to produce di-(4-N-methyl-5-methylthiazole-1-ethoxy) axially substituted phthalocyanine diiodide (SZ-SiPc-I2). The cellular uptake, phototoxicity, and dark toxicity of SZ-SiPc-I2 were evaluated in BxPC-3 human pancreatic cancer cells using laser scanning confocal microscopy and cytotoxicity assays. Results revealed that SZ-SiPc-I2 was rapidly and effectively internalized by BxPC-3 cells, exhibited excellent biocompatibility, and demonstrated high phototoxicity under light irradiation. These findings suggest that SZ-SiPc-I2 holds great promise as a novel and effective photosensitizer for PDT, offering potential advancements in cancer therapy.

Optical Coherence Tomography Angiography (OCTA) images are susceptible to motion artifacts caused by slightly jitter, which was derived mainly from the heartbeat and respiration of live samples. These motion artifacts lead to a greater degree of decorrelation in static tissues compared to vascular regions, while the unwanted strip artifacts are presented in en-face images of OCTA. In this paper, we propose an additional gradient analysis for surface identification as part of the pre-processing stage using conventional normalized cross-correlation correction algorithm. Subsequently, the upper and lower surfaces of OCT structures were utilized to extract B-scans containing only the static tissue and blood vessel region. The results demonstrated that the normalized cross-correlation of cropped B-scans presents better motion artifact correction performance than the normalized cross-correlation of successive entire B-scans.