This PDF file contains the front matter associated with SPIE Proceedings Volume 13242, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

Optical coherence tomography angiography (OCTA) is a label-free, high-resolution imaging technique for detecting blood flow based on optical coherence tomography (OCT) and time-series signal analysis. In OCTA, the time-series signals at the same position are captured, and the changes in the signals are analyzed to detect the blood flow. In this study, we evaluated different scan protocols for the OCTA regarding image quality and sampling time, including the dense A-scan, dense B-scan, and multiple B-scan protocols. In the dense A-scan or the dense B-scan protocols, the beam continues scanning with a slight change between adjacent positions. Whereas, the scan beam will pause at each slow scan position to repeatedly capture the B-scans in the multiple B-scan protocol. After the time-series signals were captured using different scan protocols and analyzed using an OCTA algorithm, the vasculature of the rat tissue was visualized. The image quality was analyzed to assess the efficiency of the scan protocols. The quantitative evaluation of the scan protocols allows for optimizing the sampling schemes in the OCTA imaging of biological tissues.

Endoscopic optical coherence tomography (OCT) can provide high-resolution, real-time images of luminal tissues. Three-dimensional (3D) airway reconstruction from endoscopic OCT images can visualize the anatomical structure of the airway and assist clinicians in diagnosing and analyzing airway-related diseases. Previous airway reconstruction methods did not differentiate respiratory states during the reconstruction process. By pulling back the endoscopic probe, B-scans at different cross-sections were captured in various respiratory states such as expiration or inspiration. Changes in luminal structure in the images could be attributed to respiratory motion. In this study, an automatic airway 3D reconstruction framework with respiratory motion correction is proposed for endoscopic OCT images. The proposed framework consists of three main steps: automatic airway lumen segmentation by leveraging a convolutional neural network, respiratory state-consistent data extraction according to the contour variation during the respiratory motion, and rotational distortion correction based on the segmented lumen contours. The proposed reconstruction framework was validated on experimental datasets acquired from the rabbit. After acquiring airway B-scans covering the entire respiratory cycle at each spatial position, the airway 3D structure can be reconstructed. The results show that the proposed framework can automatically reconstruct the 3D airway from endoscopic OCT images and keep the respiratory state consistent with accurate performance.

With the advancement of technology and increasing security demands, the exploration and extraction of new internal fingertip features have become a significant trend. In traditional fingerprint recognition systems, enhancing anti-spoofing capabilities is crucial. Conventional fingerprints are typically obtained through surface imaging, making their texture features easily susceptible to theft. Optical coherence tomography (OCT) technology offers non-invasive, high-resolution, and live tissue detection advantages, providing micron-level resolution images of biological tissues within a millimeter depth range. This enables the capture of more secure and stable internal biometric features such as internal fingerprints, sweat pores, and sweat glands. Subcutaneous fingerprints are stable, difficult to alter, and possess strong anti-spoofing characteristics. Consequently, subcutaneous fingerprint recognition promises higher security and reliability, addressing the shortcomings of currently prevalent fingerprint recognition systems. This paper presents a subcutaneous fingerprint recognition scheme based on an embedded system. The scheme utilizes Xilinx's Zynq, an all-programmable System on Chip (SoC), and employs OCT technology for fingerprint capture to meet the reliability demands of fingerprint recognition. It addresses issues in traditional OCT capture systems, such as large size, high power consumption, and poor scalability. By using image processing algorithms such as Gray level Co-occurrence Matrix(GLCM), the system extracts features from subcutaneous fingerprint images, achieving low-cost, real-time subcutaneous fingerprint image capture and recognition.

Optical Coherence Tomography(OCT) system is a non-contact imaging modality based on low-coherence optical interferometry, used for imaging turbid scattering media. They excel in rendering depth-resolved images of internal structures with micrometer-scale resolution. Previous OCT systems have some defects in image reconstruction, which are limited by complex signal processing and mathematical computation, slow image processing speed, difficult to realize real-time imaging, and high equipment cost. This paper proposes an OCT image reconstruction algorithm acceleration scheme based on the combination of FPGA (Field-Programmable Gate Array) and Python, aiming at accelerating and simplifying the image acquisition and processing of the OCT system through Python, so as to enhance the efficiency of medical diagnosis and biological research. Using Python as the upper computer control software, provide user-friendly graphical interface, output spectral waveform and then realize the Fourier transform, de-direct current and autocorrelation terms and other algorithmic steps to generate OCT images, to realize the real-time data transmission and processing. Python not only has a powerful data visualization ability, but also has the advantages of simple operation, easy to develop the program to ensure that the system operates efficiently.

Background: Singlet oxygen (¹O₂) is a key therapeutic molecule in photodynamic therapy (PDT). Quantitation of ¹O₂ luminescence is important for monitoring and optimizing PDT process. Objective: The aim of this study was to evaluate a custom-built superconducting strip single photon detector (SSPD)-based time-resolved photon counting system for ¹O₂ luminescence detection. Materials and Methods: The wavelength responses of optical collection system were verified by a spectrum analyzer. A dual-channel signal generator simulates pulsed signals of different frequencies were used to verify the circuit system. ¹O₂ luminescence generated by the photoexcitation of Rose Bengal solution was examined. Results: The ¹O₂ detection system could transmit photons of 1270 nm and the time-resolved system showed the response down to the nanosecond range and was capable of converting the different time responses into a square wave signal. When 10 μM aqueous and methanolic RB solutions were excited with a 20 mW 532 nm laser the measured ¹O₂ lifetimes were 2.93±0.37 μs and 9.45±0.83 μs, respectively. At the same concentration, when the excitation power increased the number of singlet oxygen produced per unit of time also increased. Conclusions: The custom-built SSPD-based ¹O₂ detection system provides a reliable and sensitive means for the quantitation of ¹O₂ luminescence generated from PDT process.

The viscoelasticity of the retina can reflect its function and state. Accurate assessment of retinal viscoelasticity can assist in early diagnosis of retinal diseases. With high-resolution and non-contact features, optical coherence elastography (OCE) has been used to evaluate the retinal elasticity based on the elastic wave velocity measurement. Nonetheless, the retinal viscosity cannot be assessed. In this study, a shear wave dispersion OCE method was proposed to measure the retinal viscosity and elasticity. After acoustic radiation force (ARF) induces a shear wave, optical coherence tomography (OCT) visualizes shear wave propagation in the retina. The wave velocities at different frequencies are analyzed, and the viscoelasticity is quantified based on the dispersion analysis. The accuracy of the method was verified on phantoms with different glycerol concentrations. The viscosity of the phantom is related to the concentration of glycerin, and its elasticity is adjusted by the agar concentration. The OCE results closely matched the elasticity measured by a mechanical testing system. Furthermore, the retinal shear wave velocity dispersion on the ex-vivo porcine eye was analyzed to determine its viscoelasticity. Our results demonstrate that ARF-OCE can quantitatively evaluate the viscoelasticity of the retina. The shear wave dispersion OCE method has great potential for diagnosing retinal diseases.

Background and objectives: Dermatoscope is an important optical tool for dermatologists. Its illumination system is a key component for high quality visual observation and digital photography. The aim of this study was to evaluate the illumination system of a high-end handhold dermatoscope. Materials and Methods: DermLite DL5 was used for this study. The dermatoscope equipped with sophisticated illumination system for visualization under white, yellow and ultraviolet (UV) light with or without the use of polarization. A fiber optic spectrometer was used to measure the spectra of each lighting mode. A handhold spectrometer was used to measure the color temperature, luminance and chromatic aberration of various lighting conditions under non-polarized, cross-polarized and parallel-polarized modes. The uniformity of each lighting condition was analyzed by pixel analysis of projected digital images. Results: The peak wavelength of the UV LEDs was 377 nm and the full width at half maximum (FWHM) was 21.27 nm. The peak wavelength of the yellow LEDs was 591 nm and the FWHM was 15.5 nm. The color temperature of white LEDs was over 9000 k in several modes, whereas the color temperature of the yellow LEDs varied widely and the color temperature of mixed lighting remained stable at 5000 k. The uniformity of white, yellow and mixed lighting was less than ±10%. Conclusions: The designed higher color temperature of the tested handhold dermatoscope can accommodate the physician's view of skin blood vessels. Good uniformity under different lighting and viewing modes not only satisfies the human eye but also ensure the high quality of dermatoscope digital image of the skin.

Optical coherence tomography (OCT) represents an advanced, real-time, and non-invasive imaging modality that has found widespread application in both biomedical diagnostics and materials inspection. Despite its versatility, the absence of global positional feedback for the OCT probe constrains the system’s ability to conduct large-area scans, thereby limiting its applicability across a wider range of contexts. In this work, we introduce a novel visual tracking algorithm designed for real-time localization of the OCT probe. By integrating spatial position data with captured OCT images, this approach facilitates the potential for three-dimensional (3D) image reconstruction. The proposed method employs an AprilTag-based tracking framework, selected for its robustness in dynamic environments. Fiducial markers (AprilTags) are utilized to enhance the accuracy and reliability of probe localization across six degrees of freedom (6-DOF). Using a high-precision motion stage, we evaluated positional errors in both two-dimensional (2D) and three-dimensional (3D) scanning scenarios. The results demonstrate that our system achieves localization accuracy with an average position error as low as 2.44 × 10⁻⁵ m at the processing speed 83 frames per second. Compared to existing OCT localization techniques, our method offers a precise, cost-effective, and portable solution for large-area probe localization.

Breast cancer (BC) is a significant health concern for women, with its classification into multiple stages contingent upon the dimensions of the tumor, the extent of lymph node involvement, and the presence of distant metastasis. Despite the application of uniform treatment protocols to cases of similar staging, the outcomes are subject to variability due to the inherent heterogeneity of the disease, highlighting an urgent need for further investigation. The tumor microenvironment (TME) plays a pivotal role in tumor progression and metastasis, with collagen fibers emerging as a critical component of the TME that is implicated in these processes. However, the precise interplay between collagen fibers and tumor staging remains to be elucidated. Advancements in multiphoton microscopy (MPM), which capitalizes on nonlinear optical phenomena, have yielded impressive imaging capabilities, facilitating the real-time visualization of tumor histology and the quantification of metabolic activity within tumors. Recent studies have underscored the intricate relationship between collagen fibers and the dynamics of tumor evolution.

In this study, we utilized multiphoton microscopy to image three distinct tumor-associated collagen signatures (TACS) at the invasive front of the tumor. We then used MATLAB to extract the corresponding collagen morphological features and analyzed their correlation with clinical staging. Our results revealed significant changes in the morphological features of collagen fibers in TACS across different stages of BC at the tumor invasion front. Notably, the proportionate area and number of collagen fibers were found to be inversely correlated with the clinical staging risk group of the disease. Our findings offer new perspectives for the clinical staging of BC, providing valuable insights that may enhance the predictive accuracy of disease progression and prognostic outcomes.

As DECT becomes widely accepted in the field of diagnostic radiology, there is growing interest in using dual-energy imaging to improve other scenarios. In this context, a new mobile dual-source dual-energy CBCT is being developed for scenarios such as radiotherapy and interventional radiology. The device performs dual-energy measurements by utilizing two X-ray sources mounted side-by-side in the z-axis direction, causing the problem of a mismatch in the fields of view of high-energy and low-energy sources in the z-axis. To solve this problem, this study proposes a method based on deep learning to generate high-energy and low-energy CT images in the missing fields of view. This method can generate high-energy (or low-energy) images from low-energy (or high-energy) images, and then complete the information in the missing fields of view. Furthermore, to enhance the quality of the generated images, a plug-and-play frequency-domain Mamba module is designed to extract frequency-domain features in the latent space, and then the redundant feature maps are filtered out through the designed frequency channel filtering module so that the model can pay more attention to learn and extract the effective features. Experimental results on the simulated data show that the proposed method can effectively generate the missing low- and high-energy CT images, and the SSIM, PSNR, and MAE are up to 99.3%, 48.1dB, and 6.3HU, respectively. Moreover, the generated images could maintain good continuity in the z-axis, which means that our method can effectively ensure the consistency in the fields of view of dual sources. In addition, our model can be further fine-tuned online using the paired dual-energy data in the overlap fields of view when dealing with data from unseen patients, constructing the patient-specific model to ensure the robustness against different samples.

Computed Tomography (CT) is a high-precision medical imaging technique that utilizes X-rays and computer reconstruction to provide detailed three-dimensional images of human anatomy. It is used for clinical diagnosis and treatment. Non-ideal scanning conditions often occur, including the presence of metal implants in the human body and limited-angle scanning. These non-ideal conditions result in serious metal artifacts and limited-angle artifacts. To address the above challenge, in this paper, we propose a novel deep dual-domain progressive diffusion network, namely DPD-Net, to jointly suppress metal artifact and limited-angle artifact for the first time. DPD-Net leverages the advantage of dual-domain strategy for limited-angle artifact suppression in image-domain and metal trace inpainting in sinogram-domain simultaneously. To sufficiently solve dual-artifact problem, the dual-domain generative diffusion models are designed for data distribution learning. The proposed DPD-Net is trained and evaluated on a publicly available dataset. Extensive experimental results validate that the proposed method outperforms the state-of-the-art competing methods.

Removing ring artifacts presents a significant challenge in x-ray computed tomography (CT) systems, particularly in those utilizing photon-counting detectors. To solve this problem, this study proposes the Inter-slice Complementarity Enhanced Ring Artifact Removal (ICE-RAR) algorithm, which is based on a learning-based approach. The variability and complexity of detector responses make it challenging to acquire enough paired data for training neural networks in real-world scenarios. To address this, the research first introduces a data simulation strategy that incorporates the characteristics of specific systems in accordance with the principles of ring artifact formation. Following this, a dual-branch neural network is designed, consisting of a global artifact removal branch and a central region enhancement branch, aimed at improving artifact removal, especially in the central region of interest where artifacts are more difficult to eliminate. Additionally, considering the independence of different detector element responses, the study proposes leveraging inter-slice complementarity to improve image restoration. The effectiveness of the central region reinforcement and inter-slice complementarity was confirmed through ablation experiments on simulated data. Both simulated and real-world results demonstrated that the ICE-RAR method effectively reduces ring artifacts while preserving image details. More importantly, by incorporating specific system characteristics into the data simulation process, models trained on simulated data can be directly applied to unseen real data, presenting significant potential for addressing ring artifact removal (RAR) issue in practical CT systems.

Inadequate training data and class imbalances will often affect the generalizability of many deep learning models, primarily those meant to detect rare illnesses. We propose that increasing the number of positive samples would address the difficulty of detecting breast lesions in whole slide images. We employ a controllable image synthesis framework for data augmentation inspired by CycleGAN. We use a semantic mask to guide the image-to-image translation between the healthy and pathological domains. We introduce pathology on healthy whole slide images in a location specified on the binary mask and train our model using adversarial learning. The masks provide exact information about the shape and location of the pathological features, resulting in realistic images that can be used alongside real data. We then add the synthetic images to the real data from the publicly available BReAst Carcinoma Subtyping (BRACS) dataset comprising breast histology images and use the augmented data to detect lesions. When enhanced with classical data augmentation, our enriched dataset increases breast lesion detection capabilities, offering a unique opportunity for early cancer diagnosis. The model trained with the combined data had its area under the curve (AUC) closest to one, implying a minimal risk of missing potential positive diagnoses and the chance to identify potential breast cancer cases early. We demonstrate that leveraging synthetic images as an additional augmentation tool potentially solves the challenge of insufficient pathological data in biomedical imaging. Our code is available on GitHub.

Recently, a new paradigm of network interactions of physiologically relevant cortical rhythms has been proposed, revealing different classes of interaction patterns that coexist during a given physiological state and are reorganized during transitions between physiological states. It has been shown that physiological states cannot be fully described by focusing only on individual rhythms, and continuous coordination between brain rhythms as a complex network underlies physiological function. Based on this paradigm, we study age-related differences in cross-communication of brain rhythms during sleep-wake transitions. We show that such differences are more pronounced during wakefulness.

40-Hz visual stimulation (VS) and near-infrared (NIR) light transcranial photobiomodulation (PBM) have been increasingly receiving attention in phototherapy of Alzheimer disease (AD). Both of them can be effectively employed to suppress the progression of AD. The present report includes a comparative study of different light stimulation parameters (40Hz pulsed irradiation from 808 nm laser or visible LED) to assess the therapeutic efficacy. Our research has revealed that PBM irradiation with AD mice model displayed the considerable attenuation of the accumulation of amyloid-β peptide (Aβ) plagues by using fluorescence staining and two-photon excited fluorescence (TPEF) microscopy of Aβ peptide plaques in the brain slices. Meanwhile, we also evaluated the Aβ plagues by in vivo imaging which combined TPEF and coherent anti-Stokes Raman scattering microscopy. The cerebral amyloid angiopathy and multiple Aβ plagues was much less pronounced in AD mice which were treated by PBM.

The response of cells of different origin to photodynamic treatment in vitro was studied using two realizations of quantitative phase imaging (QPI): off-axis digital holographic microscopy and spatial light interference microscopy (SLIM), and using fluorescence lifetime imaging microscopy (FLIM). Holographic techniques were shown to allow noninvasive monitoring and analysis of the response of both individual cells in a sample and their entire population to photodynamic treatment. Dynamics of changes in the phase shift introduced by cells provided information on cell death type and rate. Utilization of a low-coherence radiation source in the SLIM realization ensured reduced measurement error due to an absence of coherent noise. Changes in the fluorescence intensity and decay time of the applied chlorin-based photosensitizer in cells were shown to be due to photobleaching of the photosensitizer, rather than to intracellular processes occurring in the course of cell death. The observed variations in optical and morphological parameters of cells as a function of treatment dose were shown to conform to the specific cell death pathways. The advantages and disadvantages of each technique are discussed.

Photoacoustic Tomography (PAT) combines the advantages of optical imaging and ultrasound imaging, playing an indispensable role in biomedical research and clinical investigations. However, current advanced PAT systems are large in size and expensive, limiting their widespread adoption. Despite attempts to reduce costs by using low-cost light sources, research on sensor driving and data acquisition optimization remains lacking. Therefore, we present a low-cost and high-speed PAT system consisting of a 20Hz pulse frequency PhotoSonus-YAG laser, ultrasound array transducer, a multi-channel high-speed data acquisition system, and computer.This study designs a multi-channel high-speed data acquisition system (DAS) based on FPGA. The system, with FPGA as the core, DDR III SDRAM as the storage device, and a 14-bit high-performance ADC as the core analogto- digital conversion chip, utilizes a USB-based high-speed data acquisition card solution. To meet the demand for synchronous processing of multi-channel signals, the system employs high-end FPGA chips from Xilinx’s ZYNQ7000 series and Texas Instruments’ AFE5816 chips from the ultrasonic AFE series. These components are interconnected via Low Voltage Differential Signaling (LVDS) interfaces to ensure high-speed and highly reliable digital signal transmission. The designed high-speed data acquisition system achieves a collection of 65MSPS × 14 Bit × 16 channels, with a maximum data acquisition speed of 1000 frames per second. This design not only significantly reduces the volume and cost of the PAT system but also ensures the quality of image preservation through real-time data acquisition and processing.

This study investigates the coupling strength between low-frequency peripheral and cerebral hemodynamics among young, healthy volunteers, with concurrent acquisition of peripheral NIRS, brain fMRI, and EEG across wake and NREM sleep. The results document a strong positive coupling between low-frequency peripheral and cerebral hemodynamics during all stages except deep sleep (NREM3). Collectively, our results demonstrate that systemic physiology remains a dominant source of variability in brain hemodynamics both during resting wakefulness and light NREM sleep. However, deep NREM3 sleep may be an exception to this phenomenon implicative of its noteworthy role in optimal restoration of cerebral vasomotion.

Metamaterials, due to their extraordinary physical properties not found in natural materials, can control their electromagnetic properties by adjusting their geometry and structural parameters. Traditional MRl systems are often affected by magnetic field distortion and signal attenuation caused by metal structures. The introduction of metamaterials provides a new way to overcome these problems. However, designing metamaterials in the traditional way can present highly complex optimization challenges. In order to solve this problem, we introduce deep learning technology to study the relationship between the properties of metamaterials and the electromagnetic response under electromagnetic drive by training neural networks, and the experimental results show that compared with traditional hand-designed metamaterials. Metamaterials optimized by deep learning show superior performance in MRl systems. The combination of deep learning and electromagnetic metamaterial design opens up new directions for the development of MRI technology and has the potential to advance the entire field of healthcare.

The paper performs results of the use of machine learning methods to differentiate SERS spectra in patients with and without cardiovascular pathologies. Approaches were applied to processing spectral data arrays consisting of 1266 spectra for various groups of patients: healthy patients, patients with pathology of cardiovascular diseases, healthy patients receiving therapy, and patients with pathology of cardiovascular diseases receiving therapy. The applicability of the random forest algorithm for classification problems were shown. Potential spectral biomarkers of differences between the groups of patients on whom these algorithms were tested were identified. The achieved classification accuracy using the random forest spectra algorithm for the groups of healthy patients without therapy and patients with cardiovascular pathology without therapy was 83.4%. When classifying the presence of therapy in healthy patients (control), the accuracy was 76.26%; in patients with cardiovascular pathology, the accuracy was 70%.

Non-invasive optical glucose detection faces significant challenges due to the need to identify and extract glucose-induced signals amidst continuous human variations and probing disturbances. To ensure stable near-infrared optical signal acquisition in vivo, we enhanced the design of a wearable detector and introduced strategies to mitigate human-induced variations, aiming to minimize unnecessary fluctuations and interferences. Our custom-designed multi-ring InGaAs detector, combined with a differential method, achieved a high signal-to-noise ratio (SNR) during in vivo data acquisition. The proposed posture-aiming method enabled continuous, high-stability data collection for 1-2 hours in vivo, even with slight human motion. These enhancements enable the direct acquisition of near-infrared optical signals modulated by blood glucose levels in vivo. Results from Monte Carlo (MC) simulations and data collected from fasting subjects validated the detection approaches’ capability for stable spectroscopic detection. We conducted 30 oral glucose tolerance tests (OGTT) involving 28 volunteers. At 1550 nm, we successfully extracted optical signals that were continuously synchronized with blood glucose fluctuations, achieving an average coefficient of determination (R²) of 0.82 across the 30 OGTT tests.

Vitamin A, a fat-soluble vitamin essential for various biological processes, exists in multiple metabolic forms that are primarily converted in the liver. This complex and organ-specific metabolism depends on subcellular localization for proper conversion and function. Even within a single cell, the distribution of vitamin A is highly heterogeneous. Leveraging sub-micron spatial resolution and chemical selectivity, hyperspectral stimulated Raman scattering microscopy (hSRS) has been widely used to visualize metabolites in complex biological samples. Here, we employ (hSRS) imaging in fingerprint regions to visualize vitamin A distribution in mammalian cells.

This study demonstrates the possibility of using the SERS method to detect methotrexate (MTX) molecules in the blood plasma of patients in concentrations up to 10^-6 M. The paper performs investigation for samples from patients who took the drug at different doses. Borosilicate glasses coated with silver nanoparticles were used as the surface. We analyzed the differences between the spectra of patients after taking medication and healthy volunteers without taking medication. The characteristic maxima of methotrexate were determined.

Background: Quantitative measurement of photosensitizer during photodynamic therapy (PDT) is critical. One of the practical approaches is to measure photosensitizer’s fluorescence, however, the influence of tissue optic properties needs to be considered in in situ measurement. Objective: To design and validate a combined system for dual detection of fluorescence and tissue diffuse reflectance. Materials and Method: For fluorescence detection, a 405 nm laser was used for excitation, a power meter for monitoring laser power and a fiber optic spectrometer for recording fluorescence. For diffuse reflectance measurement, a halogen lamp was used as light source and a fiber optic spectrometer for recording diffuse reflectance. Hemoporfin (HMME) was used as a model photosensitizer. Mice were used to evaluate the dual detection function of fluorescence and tissue diffuse reflectance after i.v. injection of HMME. Results: Solutions of different concentrations of HMME were used to test the intensity responses of the fluorescence detection system. A linear correlation was observed at HMME concentration lower than 10 μg/ml. Dual-band correction using corresponding diffuse reflectance data was used for the correction of original fluorescence spectrum. Diffuse reflectance corrected fluorescence data might truly reflect HMME concentration in the skin tissue. Conclusion: The preliminary evaluation suggests that the combined system can be used for quantitative measurement of photosensitizer fluorescence.

Owing to the minimal invasiveness, cytology is an indispensable technique in the routine pathology. However, traditional cytology only enables the low sensitivity (50%-60%) and high time-consuming for the diagnosis. Our previous study demonstrated that stimulated Raman molecular cytology (SRMC), which is label-free, faster, and noninvasive, provides additional composition information leading to higher diagnostic accuracy around 85%. However, current AI-assisted SRMC generally involves cell segmentation and feature extraction steps, which may involve issues of the artifacts. Recently, various methods for global feature analysis, such as Transformer and CNN, are capable of preserving both global and local information. Therefore, we propose an end-to-end Transformer hybrid model combining the advantages of both Transformer and CNN to analyze stimulated Raman cytology images for accurate and rapid peritoneal metastasis (PM) diagnosis of gastric cancer (GC). The Transformer hybrid model can enhance the Transformer’s global modeling ability simultaneously with the local guidance from CNN features. To evaluate the performance of this Transformer method, we collected 816 stimulated Raman cytology images from 80 locally advanced gastric cancer patients, with 36 PM positive and 44 PM negative. The Transformer method could reach 88.89% sensitivity, 86.36% specificity, and an AUC of 0.903 with leave-one-out cross-validation for 80 patients. Compared with traditional cytology, the false negative rate of our label-free stimulated Raman cytology reduces by about 30-50%. Together, our Transformer approach demonstrates the potential for accurate and rapid PM diagnosis based on exfoliated cytology.

Optical Coherence Elastography (OCE) is a non-invasive elastography technique, which can deduce the elastic properties of tissue by measuring the displacement or deformation in tissue caused by internal or external excitation. In recent years, it has been well accepted that the biomechanical properties of cornea are associated with various ophthalmic diseases. OCE has exhibited a good potential in measuring corneal elasticity, and however in vivo quantitative measurement remains challenging. Thus, in this paper, we designed an OCE system equipped with a load-measuring stress sensor to measure the deformation and geometric parameters of cornea during compression. Furthermore, a compressive OCE corneal elasticity measurement model was developed based on shell theory, which is dedicated to translating the measured value into Young's modulus. Finally, the method was evaluated on both artificial eye model and porcine cornea ex vivo. The results indicate that the OCE measurement combined with shell model can potentially aid in clinical measurement on cornea elasticity in vivo.

In many surgical procedures, the objective is to restore tissue oxygenation and achieve effective revascularization. However, methods for evaluating tissue perfusion and oxygen metabolism are limited. Current clinical approaches mostly assess the patient's overall oxygen saturation level and lack a non-invasive, real-time method to evaluate local hypoxia during surgery. In this study, New Zealand white rabbits were first anesthetized and endotracheally intubated. The experiment began with oxygen supply being suspended for five minutes to establish a rabbit model of hypoxia. Subsequently, synchronous blood gas analysis and photoacoustic imaging tests were conducted on the carotid and femoral arteries of the rabbits. Blood gas analysis showed that the carotid arterial oxygen saturation of three rabbits was 100% before oxygen deprivation; after a five-minute cessation of oxygen, the carotid arterial oxygen saturation decreased to 3.6% ± 1.5%, with a rapid initial decline followed by a slower rate. Photoacoustic imaging results indicated that the oxygen saturation in the femoral arteries of the three rabbits dropped sharply from 100% to below 10% at the moment of oxygen deprivation, while muscle oxygen saturation fell from around 90% to below 30%. The trend of change was consistent with arterial oxygen saturation, but with a slightly smaller amplitude. This study validates that photoacoustic imaging technology can accurately reflect changes in vascular and tissue oxygen saturation within the body during hypoxic conditions, and due to its capability for localized detection and non-invasive real-time monitoring, it holds promise for future use in assessing anastomotic patency in revascularization surgeries.

Photobiomodulation (PBM) is a regulatory approach that utilizes red or near-infrared (NIR) light to promote tissue repair and blood microcirculation, thereby generating positive therapeutic effects at multiple levels. In recent years, as an emerging non-invasive therapy, PBM has been widely applied in the modulation of neurodegenerative diseases. However, due to the light scattering and absorption characteristics in the skull and the incomplete understanding of the intracranial propagation properties of light at different wavelengths, the application of optical techniques in transcranial modulation is greatly limited. In order to optimize the light therapeutic efficacy of neurological diseases, there is an urgent need to investigate the physical property of light transcranial spread. In this work, the transmittance of NIR light at different power levels and frequencies in the mouse skull at various points were detected. Additionally, the penetrations of light at 808 nm and 660 nm with varying tissue thickness were compared by Monte Carlo simulation analysis, providing data support for establishing the optimal therapeutic strategy of transcranial PBM.

Laser speckle contrast imaging (LSCI) stands as a critical imaging modality capable of real-time acquisition of dynamic blood flow information, finding extensive utility in clinical and experimental settings requiring prompt feedback. While traditional Reflect Laser Speckle Contrast Imaging (R-LSCI) excels in superficial vascular imaging, Transmissive Laser Speckle Contrast Imaging (T-LSCI) offers superior performance in imaging deeper vascular structures. In this study, to harness the complementary strengths of R-LSCI and T-LSCI, we introduce a novel approach: the space domain contrast optimization algorithm. This method integrates laser speckle contrast images captured under both conditions through spatial domain processing and optimized contrast ratio selection. Our results demonstrate that the proposed method enhances vascular visualization and achieves superior vascular imaging outcomes when compared to R-LSCI and T-LSCI under space contrast (sK), a spatial domain method.

The laser speckle contrast imaging (LSCI) technique, based on dynamic light scattering theory, is a non-scanning method for blood flow imaging that offers wide-field coverage. However, under traditional single-exposure conditions, static scattering may interfere with imaging, leading to reduced contrast and diminished image clarity, thereby affecting the accuracy of blood flow monitoring in biological tissues. This study aims to address the challenge of static scattering under single-exposure conditions by employing the adaptive window space direction contrast (awsdk) imaging method proposed in the laboratory. Through validation in reflective systems using phantoms and rabbit ear regions, this research combines the awsdk imaging method with optimized single-exposure techniques, effectively correcting static scattering and eliminating the impact of system noise on speckle contrast. This approach not only enhances imaging quality but also enables rapid monitoring of blood flow changes using speckle contrast measurements under single-exposure conditions, providing an effective solution for further advancement of laser speckle contrast imaging technology.

Parallel imaging is widely used in the clinic to accelerate magnetic resonance imaging (MRI) data collection. However, conventional reconstruction techniques for parallel imaging still face significant challenges in achieving satisfactory performance at high acceleration rates. It results in artifacts and noise that affect the subsequent diagnosis. Recently, implicit neural representation (INR) has emerged as a new deep learning paradigm that represents an object as a continuous function of spatial coordinates. INR’s continuity in representation enhances the model’s capacity to capture redundant information within the object. However, it usually needs thousands of training iterations to reconstruct the image. In this work, we proposed a method to speed up INR for parallel MRI reconstruction using hash-mapping and a pre-trained encoder. It enables INR to achieve better results with fewer training iterations. Benefiting from INR’s powerful representations, the proposed method outperforms existing methods in removing the aliasing artifacts and noise. The experimental results on simulated and real undersampled data demonstrate the model’s potential for further accelerating parallel MRI.

As a method widely used in the treatment of migraine in clinic, the acupuncture is effective, but there are still great differences between the treatment effects of different patients. To improve the ability to predict patient treatment outcomes for migraine patients, Stacking ensemble learning method is used to construct a predictive model for the efficacy of acupuncture treatment for migraines. Collect a large amount of migraine patient data, including Visual Analogue Scale (VAS), Migraine Specific Quality of Life Questionnaire (MSQ), and patient symptoms, and perform data preprocessing and feature engineering. The Support Vector Machines (SVM), Random Forests (RF), Gradient Boosting (GBDT), CatBoost, Extreme Gradient Boosting (XGBoost) and other machine learning methods is used to establish the prediction model of the effect of acupuncture on migraine. These models are arranged and combined as base learners, and the Multi-layer Perceptrons (MLP) are used as meta learners to construct Stacking models. The hyperparameters are optimized through grid search methods to further improve prediction performance. The experimental results show that the final Stacking model with the best performance achieved a prediction accuracy of 93.16%, while the accuracy, recall and the F1 scores are also above 93%. The prediction results of these models were validated through Magnetic Resonance Imaging (MRI) data, further confirming their reliability and effectiveness. The method provides an important reference and support for clinical decision-making and acupuncture treatment of migraine.

Due to the high cost of high-field MRI equipment, low-field MRI systems are still widely used in small and medium-sized hospitals. Compared to high-field MRI, images acquired from low-field MRI often suffer from lower resolution and lower signal-to-noise ratios. And the analysis of clinical data reveals that noise levels can vary significantly across different low-field MRI protocols. In this study, we propose an effective super-resolution reconstruction model based on generative adversarial networks (GAN). The proposed model can implicitly differentiate between various sequence types, allowing it to adapt to different scan protocols during reconstruction process. To further enhance image detail, a one-to-many supervision strategy is employed during the training process, utilizing similar patches within a single image. Additionally, the number of basic blocks in the model is reduced through knowledge distillation to meet the speed requirements for clinical use. The experimental results on actual 0.35T low-field MR images suggest that the proposed method holds substantial potential for clinical application.

The biological characteristics of the human eye are valuable in painless and non-invasive disease diagnosis, as they can be readily observed and accessed externally. Among these features, the scleral plaques and pigmentation, which are directly visible on the ocular surface, can provide useful information on an individual's health status. However, due to the inherent variability in the size and the shape of plaques and pigmentation spots among patients, automatic detection of these features in scleral images remains a significant challenge that has yet to be fully explored. In this study, we develop an object detection algorithm based on YOLOv5 for automatic detection of plaques and pigmentation spots in the scleral image. Specifically, the scleral region is initially extracted by the UNet++ network, and then an improved YOLOv5 model equipped with a convolution block attention module and Mamba blocks is employed to detect potential plaques and pigmentation spots in the scleral region. The convolution block attention module facilitates the detection and characterization of targets by eliminating redundant information, whereas the Mamba block integrates the state space model to expand the receptive fields of the proposed model, thereby enhancing its performance. The proposed detection algorithm was validated on a clinical dataset comprising scleral images from 388 subjects. The experimental results demonstrate that the proposed algorithm is capable of effective detection of scleral plaques and pigmentation spots and that it outperforms the original YOLOv5 model as well as other competing object detection algorithms.

Breast cancer has a high morbidity and mortality rate worldwide. The overexpression of HSP70 (Heat Shock Protein 70) has associated with the occurrence, development, treatment, prognosis and drug resistance in breast cancer, and may become a new target for anti-tumor therapy. In this paper, triple-negative breast cancer cells MDA-MB-231 were used to investigate the effects of HS (Heat Shock) and HSP70 inhibitor VER-155008 treatments on mitochondrial morphology and membrane potential by laser scanning confocal microscopy. Meanwhile, the cell proliferation was studied by MTT method. We observed that mitochondrial networks were broken and the mitochondrial membrane potentials were decreased with the HSP70 function inhibition, and its pro-apoptotic effects can be alleviated when VER-155008 treatment was combined with HS. These results were the same as the effects of HS and HSP70 inhibitor on MCF-7, ER-positive breast cancer cells, we had reported. In addition, the proliferation of MDA-MB-231 and MCF-7 after different treatments indicated that HS treatment promoted proliferation in MCF-7, but not in MDA-MB-231.

With the aggravation of the aging population, the incidence of spontaneous intracerebral hemorrhage remains high. As a serious acute cerebrovascular disease, its timely and accurate diagnosis is crucial for the prognosis of patients. The segmentation of brain hemorrhage region image is an important step in medical image analysis, which is of great significance in assisting doctors to formulate treatment plans and evaluate the progress of the disease as well. Traditional manual segmentation methods are time-consuming and subjective which makes it difficult to meet the needs of clinical rapid diagnosis. With the development of deep learning technology, significant results in medical image segmentation have achieved. Therefore, it is of great significance to build a medical image automation segmentation model based on deep learning technology to annotate the hemorrhage region in the patient's head CT axial image accurately and efficiently. Based on this, the paper evaluates the effect of Jun Ma's U-Mamba network on the segmentation and prediction of brain hemorrhage regions. The experimental results show that this network model can achieve an accurate and efficient segmentation and prediction of the bleeding region in the brain CT image. By estimate, the Dice coefficient, IoU, average precision and average recall all above 0.9.

Lymphoma has become one of the most prevalent malignant tumors in China, with a significant rise in incidence among young people in recent years. Early diagnosis and treatment are therefore crucial for improving patient outcomes, including efficacy, survival, and quality of life. In this study, we developed a multimodal detection system that combines twodimensional (2D) light scattering and electrochemical techniques to differentiate between normal and tumor cells at the single-cell and molecular level. Using a laser microscopy detection system, we capture 2D light scattering images of individual cells, where the lymphocytes display distinctive patch-like patterns. The texture of these patterns is influenced by the internal cellular structures, and the differentiation of normal and tumor cells is achieved by extracting and analyzing the eigenvalues from the light scattering images. Additionally, electrochemical sensors detect hydrogen peroxide levels in the cellular solution by measuring changes in current, with tumor cells producing a greater current variation than normal cells. A support vector machine (SVM) algorithm was employed to distinguish between normal and tumor cells, achieving an accuracy of 88%. The results demonstrate that the multimodal detection system effectively differentiates normal and tumor cells from both physical and chemical perspectives, enhancing detection accuracy. This system offers a nondestructive, efficient, and cost-effective method for early cancer screening.

Paper performs results of portable SERS system for human platelets investigation. Platelets were divided to 2 groups: with acute coronary syndrome on therapy and without therapy. Spectral in groups have been investigated. An analysis of experimental data showed that changes in the SERS spectra of human platelets during therapy may be associated with changes in the amino acids (phenylalanine, tyrosine and tryptophan).

Lung cancer is one of the most fatal malignant tumors globally. Recent research has revealed that the scleral image, which can be obtained in a painless and non-invasive manner conveniently, is associated with lung cancer. The current method treats the malignant lung neoplasm detection task as a binary classification problem. It takes the multiple images of one subject’s eyes as different instances, and takes the subject as a bag, then employs the multiple instance learning technique to solve the problem. However, the current method utilizes average pooling to aggregate information on different instances in a bag, which overlooks the varying contribution of each instance to the bag label. In this study, we propose an attentionbased multiple instance learning (AMIL) model for lung neoplasm detection using scleral images. The model first employs a convolutional neural network-based backbone to extract features from each scleral image of the subject, then leverages a channel attention module to recalibrate the weight and aggregate the extracted features adaptively. Finally, a fully connection-based classification module is used to make the final prediction for the subject. The results show that the model can effectively identify the critical instances with the help of the attention mechanism, and improve the classification accuracy.

Optical coherence tomography (OCT) is a non-invasive, label-free imaging modality that generates high-resolution threedimensional images. Based on OCT imaging, optical coherence tomography angiography (OCTA) and optical coherence elastography (OCE) can visualize vascular networks and measure the elastic properties of biological tissues. In previous studies, OCTA and OCE were performed separately, providing either vascular network information or elasticity properties of the tissue. We have developed a simultaneous angiography and elastography method using a simplified sample arm structure. After mechanical pressure is loaded on the sample by a glass plate, the deformation is analyzed by the OCT phase changes, and the elasticity is assessed. Meanwhile, the vascular network is visualized by intensity-based Doppler variance analysis. A transparent flexible reference layer is placed between the glass plate and the tissue, which closely contacts the tissue. Better elasticity measurements can be achieved without affecting vascular imaging. The simultaneous elastography and angiography method was demonstrated by the phantom experiments and rat skin measurements. The results show that the information on the microvascular networks and mechanical properties can be obtained at the same imaging location. The method can provide more comprehensive information on biological tissue for disease diagnosis and treatment monitoring.

Melanoma is a highly malignant cutaneous tumor, and its early recognition is critical for improving treatment outcomes and patient survival. Related to this is melanocytic nevus, a common skin lesion whose association with melanoma has been of great interest. Photoacoustic microscopy, one of the most promising techniques in skin imaging, has many features applicable to in vivo imaging of dermatologic conditions, such as imaging of cutaneous microvascular and pigmented lesions. It has the advantages of high spatial resolution, nondestructiveness, and relatively large penetration depth for imaging oxygen and deoxyhemoglobin (HbO2 and HbR) as well as melanin. The aim of this study is to explore melanocytic nevi in depth by photoacoustic microscopy and to provide new technical support for the prevention and early diagnosis of melanoma.

The biomechanical measurement of the crystalline lens can provide valuable information to assess the development of lens-related diseases, such as presbyopia and cataracts. Optical coherence elastography (OCE) has been used to measure the elasticity of the lens surface based on elastic wave imaging. However, measuring the elasticity of the lens interior poses a challenge because optical imaging cannot easily visualize elastic waves in the transparent lens. In this study, we develop an acoustic radiation force optical coherence elastography (ARF-OCE) method to detect the propagation of elastic waves on the surface of the lens and inside the lens for the elasticity measurement. The ultrasonic radiation force excites the lens from the side of the eye, subsequently inducing an elastic wave on the lens surface or inside the lens. Optical coherence tomography (OCT) images the crystalline lens from the front of the eye with the optical beam perpendicular to the acoustic beam. When the ARF is focused on the surface of the lens, the wave propagation on the lens surface is visualized by the OCT, and the elasticity of the lens surface can be quantified. When the ARF is focused inside the lens at different depths, the time the elastic wave reaches the lens surface will change. Therefore, the velocity of the elastic wave propagation inside the lens is calculated by the ratio of the depth change to the time difference, and the elasticity of the lens interior can be quantified. The elasticity of the surface and the interior of the ex-vivo porcine lens was measured using the ARF-OCE method. The elasticity measurement of the crystalline lens provides a quantitative assessment of its biomechanical properties and has the potential for the accurate diagnosis and treatment of lens-related diseases.

Optical coherence tomography (OCT) is widely used in ophthalmology and has been a standard method for diagnosing ocular diseases. Improving spatial resolution is crucial for the visualization of ocular microstructure. In conventional OCT systems, enhanced transverse resolution typically results in a reduction of depth of field (DOF). In our study, we designed four metalenses with different microstructures and then integrated them into the sample arm of an OCT imaging system. The metalens can modulate the amplitude and the phase of incident light in the sample. In this way, the depth of focus in the OCT system can be effectively enhanced while maintaining the transverse resolution. Also, the impact of different parameters, such as the size and the spacing of micro-units, has been discussed for OCT imaging. Finally, the experiment on the glass slides has been conducted to validate the performance of OCT imaging. The results demonstrate that the metalens with a specific microstructure could effectively enhance the DOF of OCT imaging. Furthermore, the enhanced depth of focus of the OCT system opens up new opportunities for investigating tissue morphology and function.

Background: Mouse model is a unique tool for preclinical evaluation of photosensitizing drugs for their potential applications in photodynamic therapy (PDT). The fluorescent property of photosensitizer provides a useful means for in situ measurement of photosensitizer distribution. However, the autofluorescence of mouse tissues might potentially affect in situ photosensitizer fluorescence measurement. Objectives: The aim of this study was to evaluate the autofluorescence in lab mice and its influence on the in situ measurement of photosensitizer fluorescence. Materials and Materials and methods: ICR mice, commonly used lab animal model, were used. Mice were fed with regular diet and clean water. Domestically produced photosensitizing drug Hemoporfin® was used as a model sensitizer. Costume-built fluorescence imaging and spectroscope systems were used for fluorescence examination under the excitation wavelength of 400 ± 5 nm, generated from a laser or LED panel. In situ fluorescence was examined before and after i.v. injection of Hemoporfin®. Results: Epifluorescence imaging examination showed yellowish and reddish autofluorescence fluorescence in the mouth, nose, paws and tail areas. The absorption and emission spectra of mouse tissue overlaid to some degrees with that of Hemoporfin®. In situ fluorescence examination of ICR mice showed significant influence of autofluorescence on the spectral and intensity measurement of Hemoporfin® fluorescence. Such influence might be eliminated by post-measurement spectral correction algorithms. Conclusions: The autofluorescence in mouse tissues poses a significant influence on the in situ measurement of fluorescence emission of Hemoporfin®. Careful postmeasurement spectral correction is needed for accurately measuring photosensitizer fluorescence.

Optical coherence tomography (OCT) has wide applications in diagnosing diseases, benefitting from its non-invasive, high-resolution, and real-time visualization of tissue microstructures. Still, this technology faces the mutual constraint between transverse resolution and depth of focus. Metalens can modulate the light fields at subwavelength scales, effectively enhancing the depth of focus of an OCT system. In this paper, a metalens designed for the OCT system is proposed. The metalens with different structures can generate various phase modulation outcomes. Using the finite-difference time-domain (FDTD) simulation, the impacts of the nanopillar parameters (e.g. height, diameter, and spacing) and the materials (e.g. Si, SiO2, and TiO2) on phase modulation have been analyzed. Based on the simulation results, the phase modulation capability of nanopillars is compromised at reduced heights, while excessively tall structures can adversely decrease transmittance. With a constant height, the radius of the nanopillars can be manipulated to achieve a phase delay of 2π. Furthermore, the propagation efficiency and the typical depth of focus were calculated from the results of the beam intensity distribution, indicating an enhanced imaging performance of the OCT system. The results provide a basis for designing the metalens used in the OCT systems.

Using vessel-labeling, tissue-clearing, and light-sheet imaging improves vascular mapping and disorder analysis. However, some labeling methods compromise efficiency and compatibility, reducing image quality. We present Ultralabel, a superior vessel-labeling method employing a unique hydrogel dye. Ultralabel outperforms others, offering accurate vascular mapping, multi-color labeling, and reliable, user-friendly 3D vascular network visualization for biomedical uses.

Effective labeling of the vascular system is essential for understanding the vascular networks of tissues and organs. Here, we present VALID, a vascular labeling method that visualizes vascular networks through tissue clearing and light-sheet microscopy. VALID converts traditional lipophilic dye solutions into hydrogels by introducing gelatin and inhibits the aggregation of dyes, thus obtaining a complete and uniform vascular labeling pattern with high signal-to-value ratio. VALID enhances the compatibility of lipophilic dyes with solvent-based tissue clearance regimens, which was previously difficult to achieve. Using VALID, we combined lipophile dyes with solvent-based tissue clearing to perform 3D reconstruction of the vascular system within the brain and spinal cord of mice. We also used VALID to visualize and quantify microvascular damage in a mouse model of middle cerebral artery occlusion. VALID offers a simple, cost-effective vascular labeling protocol that will greatly expand the use of lipophilic dyes in the study of cerebrovascular complications.

In modern intelligent living, Carbon dots exhibit the great potential by virtue of their unique structure and attractive optical performance serving a range of monitoring and management, from biochemical analysis, biological imaging and environmental monitoring. In this paper, a sub-micro line spacing standard shape was designed based on phosphorescence carbon dots. This spacing standard not only has an unequally spacing line widths, but also shows that the particle size distribution of Carbon dots range from 1.0nm to 10nm. On the other hand, in order to exhibit the potential of our designed shape, a series of experiments were preformed to demonstrate the relationship between fluorescence intensity and special effects. Those results have been proven to be amenable for practical purposes through many tests so that it might be applicable to achieve lighting illumination in different scenes.