We used time-resolved polarized fluorescence spectroscopy for the study of nicotinamide adenine dinucleotide phosphate (NADPH) bound to isocitrate dehydrogenase (IDH) and alcohol dehydrogenase (ADH). It was shown that the fluorescence lifetimes of bound NADPH are strongly enzyme-dependent and are practically identical to those of bound NADH, indicating that fluorescence lifetime imaging microscopy (FLIM) analysis cannot universally distinguish between these two coenzymes. However, in the polarization-sensitive experiments we observed sub-nanosecond fluorescence anisotropy decay components of NADPH bound to IDH and ADH that we found to be significantly enzyme-specific. This result suggests that combined lifetime and anisotropy analysis provides increased molecular specificity for NAD(P)H.
The aim of this study was to develop an irradiant exposure strategy able to control validity of the exposure reciprocity law upon exposing tarso-conjunctival tissue to ultraviolet A (UVA) radiation as a treatment for palpebral laxity, and thus to establish a research-based criterion for implementing optimal irradiation levels in clinical settings. Tarso-conjunctival specimens were harvested postmortem from ovine and caprine eyelids and exposed to UVA radiation (365 nm) in specified conditions. Thermographic infrared analysis was applied to monitor the photothermal response. By performing irradiation in the presence of a photosensitizer (riboflavin) a photochemomechanical response was also determined based on changes in mechanical properties as evaluated in a mechanical tester. A graphical method was employed to define the failure zones of reciprocity law when applied to radiation-induced photoresponses. Discontinuities in law’s validity were found and confirmed by descending values of Schwarzschild’s p-coefficients. The photothermal response showed that law’s failure zone centered on an irradiance of 30 mW/cm2 in both sheep and goat tissues, for radiant exposures between 6 and 18 J/cm2. The photochemomechanical response indicated a failure onset around 60 mW/cm2. Anticipating the concept of safest application route, an irradiance of 30 mW/cm2 appeared an appropriate upper limit for the validity zone of reciprocity law, as indicated by the photothermal response, at exposure durations assuring a radiant exposure between 5 and 20 J/cm2.
In this work, we propose a promising non-invasive approach based on the Raman spectroscopy of human serum for chronic heart failure (CHF) diagnosis. Due to the limited sample size, which limits the performance of machine learning classifiers, this study explores data augmentation techniques to improve the classification of low- and high-grade CHF spectra using k-nearest neighbor (kNN) and partial least squares discriminant analysis (PLS-DA) algorithms. Raman spectra of 151 patients with CHF of the different stages were acquired at 532 nm excitation from serum samples collected at Samara City Clinical Hospital. Two augmentation approaches were systematically evaluated: (1) combined approach based on the linear spectral transformations (wavenumber shifting ±1–2 cm⁻¹, intensity stretching 0.9-1.1) and (2) Wasserstein Generative Adversarial Network (WGAN)-based synthetic spectrum generation, expanding training datasets 10-fold while preserving physiochemical realism. Augmentation based on linear spectral transformations yielded algorithm-specific results: kNN showed no significant receiver operating characteristic area under curve (ROC AUC) improvement (0.67 ± 0.11 original vs. 0.67–0.69 augmented), while PLS-DA achieved statistically significant gains (0.71 ± 0.11 vs. 0.80–0.81; z = 2.3–2.6, p < 0.05). WGAN augmentation proved superior across both methods, with k-NN reaching 0.74 ± 0.08 and PLS-DA achieving 0.83 ± 0.09. These findings establish WGAN as an optimal augmentation strategy for Raman-based CHF classification, achieving clinically relevant performance (ROC AUC > 0.80) from limited cohorts while enabling biomarker identification for cardiovascular diagnostics.
The interpretation of lumbar spine radiographs, the most accessible diagnostic modality for osteoporosis, is often subjective and hampered by limited accuracy. To address this, we developed a neural network-based algorithm for computer-aided diagnosis. This study provides a comparative analysis of key neural network architectures for image segmentation and classification. The proposed system, with a classification accuracy of 87.42%, offers a valuable adjunct for radiologists, mitigating diagnostic risks and enhancing the overall efficacy of osteoporosis screening
Riboflavin, a natural photosensitizer, is of significant interest for the development of low-toxicity photodynamic therapy (PDT) protocols. This study is focused on the evaluation of the cytotoxic effect and ionic imbalances in B16-F10 melanoma cells induced by riboflavin when irradiating with blue laser light (450 nm wavelength). The dynamics of cell death and changes in intracellular sodium, potassium, and calcium ion concentrations were analyzed with the use of fluorescence microscopy. At a riboflavin concentration of 50 µM under laser irradiation, the signal corresponding to early apoptotic features in B16-F10 cells increased by 2.2 times compared to the control group. The maximum increase associated with late apoptosis and necrosis was observed 6 h after exposure and exceeded the values of the control group by 1.61 times, whereas under combined treatment this effect was detected as early as 1 h after exposure (1.32 times relative to the control group). Analysis of ion homeostasis revealed that riboflavin treatment, particularly in combination with laser irradiation, led to an increase in intracellular Ca2⁺ levels (up to 1.9-fold relative to the control at 6 h after exposure). This was accompanied by an increase in Na⁺ levels and a decrease in K⁺ levels, with minimal K⁺ values observed at 6 h after exposure (approximately 0.6 of the control level). Overall, the results demonstrate the photosensitizing activity of riboflavin and support its further investigation in photodynamic approaches for melanoma. The observed effects provide a basis for further investigation of riboflavin-mediated PDT in combination with other anticancer agents to enhance therapeutic efficacy.
Computer-vision methods have been applied to automated behavior recognition in laboratory rodents. Namely, we explored the possibility of classifying certain behavior classes from still images; compared keypoint-based methods with approaches based on visual embeddings; studied the feasibility of transferring models between rats and mice; and evaluated the relevance of the number and the accuracy of the detected keypoints. We collected a dataset of a freely moving Wistar rat to train six pose-based classifiers with Long Short-Term Memory (LSTM) using six sets of keypoints produced by two detectors and four Convolutional Neural Network (CNN) classifiers using images and optical flow frames. The results demonstrated the highest mean average precision (mAP) of 65.7% for the CNN-based methods and 34.3% for the LSTM classifiers, the feasibility of recognizing visually distinct classes (rearing and body grooming) from still images, and the applicability of a keypoint detector trained on mice. The results of this study can be applied to the design of a computer vision system for automating long-term monitoring of laboratory rodent behavior
Using optical coherence tomography angiography (OCT-A), we demonstrate the feasibility of monitoring microcirculatory changes in the clinically important sublingual region, which reflects perfusion disturbances in internal organs of critically ill patients. We present the results of OCT-A monitoring of sublingual microcirculation alterations in both animals and humans induced by external stimuli (modeling massive blood loss and administration of a vasodilator drug). Our findings highlight the strong potential of OCT-A for addressing the challenges of early detection of tissue hypoperfusion in patients. We propose a signal intensity analysis for OCT-A images which might be an effective approach to predict multiple organ failure development, thereby enabling monitoring of the effectiveness of ongoing intensive care.
The paper presents the development and validation of an experimental technique enabling localized targeted photodynamic treatment at subcellular level and further noninvasive monitoring of variations in cell morphology using low-coherence holographic microscopy realized in the configuration of spatial light interference microscopy (SLIM). The key element of the setup is spatial light modulator (LCOS SLM) which enabled both the targeted irradiation and continuous quantitative monitoring of cell response. To confirm the reconstruction accuracy of phase images of microscopic objects, a test sample comprising a set of polystyrene microspheres was analyzed. Targeted photodynamic treatment was tested by localized irradiation of HeLa cells photosensitized with Radachlorin. Irradiation of cells triggered their death by apoptosis and necrosis. SLIM-assisted monitoring of cells provided quantitative data on the dynamics of changes in their morphological parameters caused by treatment.
The mortality rate of ovarian cancer remains the greatest among other oncological diseases of genital system worldwide. Using carotenoids as a preventive measure against ovarian cancer and a means to improve the state of genital system in general is a popular trend when determining the nutritive factors. In our study we have determined the total content of carotenoids in healthy feline ovaries and ovaries of cats diagnosed with leiomyosarcoma. We have found that total content of carotenoids varies between left and right ovaries of the same healthy cat. However, cats diagnosed with leiomyosarcoma had significantly lower carotenoid content in both left and right ovaries. As carotenoids can be used as chemotherapeutic agents, the study of their content in ovaries is a relevant problem that requires further study. Determination of the impact of carotenoids on oncological processes may find practical use in public healthcare.
We present the refinement of the Polarization-Modulation Pump-Probe method developed recently in our group (PCCP V. 22, 18155 (2020)) for separation and quantitative determination of linear dichroism and birefringence of the probe laser beam in biologically relevant molecules. The method was used in the study of ultrafast relaxation dynamics in the excited states of NADH in aqueous solution. The probe beam birefringence contained contributions both from resonance pump beam absorption in NADH and from nonlinear coherent effect that was attributed to the stimulated Raman scattering in water, the latter manifested as a very intense and narrow peak at sub-picosecond delay times between pump and probe pulses. The probe beam dichroism contained the contribution from the resonance pump beam absorption in NADH and practically no contribution from nonlinear coherent effects. In this case the experimental signal was approximately twice smaller than that of the birefringence case. Therefore, it was suggested that the birefringence detection allows to achieve higher sensitivity for determination of the relatively long relaxation times than the linear dichroism detection
New self-contained acid photogenerators based on substituted benzo[b]thiophene-2-carboxanilides were studied under irradiation with the 3rd harmonic of an Nd-YAG laser (355 nm) and an excimer XeCl lamp (308 nm) in toluene and methylene chloride solutions. The quantum yield of phototransformations of compounds during lamp and laser photolysis was estimated for both solvents. A model experiment was carried out using photogenerated acids to remove the dimethoxytrityl group from 5′-O-(4,4′-dimethoxytrityl)-2-deoxythymidine-3′-O-[O-(2-cyanoethyl)-N,N′-diisopropylphosphoramidite].
A method for the automatic extraction of the soft-tissue facial profile contour is proposed, along with the detection of key cephalometric landmarks based on extremum analysis of a parameterized contour function and subsequent determination of the profile type and its harmony. A variational and numerical analysis of the influence of weighting coefficients in the energy functional on segmentation accuracy and diagnostic indicators is conducted, allowing the parameter selection to be justified and ensuring robust automatic annotation comparable to manual cephalometric analysis
Despite terahertz (THz) technologies offer a number of applications in medical diagnosis and therapy, their translation into clinics is hampered by the lack of THz endoscopes capable of sensing THz response of hard-to-access tissues. In this paper, we focus on recent attempts to mitigate this difficulty. We consider the two existing principles of THz endoscopy. The first uses the fiber-coupled THz photoconductive antennas (PCAs) for the THz generation and detection in close proximity to an object. The second relies on the THz optical fibers to deliver THz waves to an analyte and then to detect the reflected THz signal. Most recent developments in the area of THz fiber optics pave the way to solve the challenging problem of THz endoscopy. Among them, we emphasize the THz fibers, fiber bundles, waveguides, and endoscopes developed by our research group based on the sapphire shaped crystals obtained by the edge-defined film-fed growth (EFG) technique.
This study presents numerical modeling and constrained spectral balancing of a compact LED-based reflectance spectrophotometer operating in the 360–760 nm range. The detected spectral response was modeled as a weighted superposition of individual LED emission spectra explicitly multiplied by the wavelength-dependent detector quantum efficiency QE(λ). In the baseline configuration with unit weights, the detected spectrum exhibited strong modulation, characterized by a coefficient of variation CV of 41.5%, peak-to-peak deviation of 131%, and maximum-to-minimum ratio of 4.24. Spectral balancing was formulated as a constrained minimization of the normalized variance over a 0.5 nm wavelength grid using a differential evolution algorithm with bounded non-negative weights. The optimized configuration reduced CV to 8.36%, peak-to-peak deviation to 32.2%, and max/min ratio to 1.39, corresponding to an approximately 80% reduction in global spectral non-uniformity as quantified by variance-based metrics. The proposed approach provides a quantitative methodology for spectral equalization in multi-LED reflectance spectrophotometers.
The study presents data on the influence of alternating magnetic field (AMF) on blood in vitro, manifested by a decrease in hemoglobin’s affinity for oxygen and an increase in nitrate/nitrite content. The effect of short-term exposure (30 min) to AMF with the strength of 300 mT and frequency of 20 Hz (corresponding to the radiation from static magnetic field sources) on the oxygen transport function (OTF) of erythrocytes and conformation of Fe-containing hemoporphyrin and globin of hemoglobin was demonstrated using Raman spectroscopy. Alteration in the content of hemoglobin in the T and R-forms was revealed, indicating a different affinity of hemoglobin for ligands and associated with changes in the polarity of the globin amino acids. It was also demonstrated that exposure of erythrocytes to AMF for 360 s modifies their gasotransmitter-forming function. These changes can significantly influence the alteration of blood OTF and change the blood flow in meeting tissue oxygen demands. The obtained results indicate the point to the involvement of the gasotransmitter-forming function in modifying blood OTF and, consequently, in altering the adequacy of blood flow in meeting tissue oxygen demands
Lung cancer is a life-threatening disease in which accurate staging of malignant nodules using computed tomography (CT) scans is critical for reducing mortality. Most existing approaches rely solely on deep learning models. This work proposes an accurate and computationally efficient hybrid deep learning framework for lung cancer analysis, integrating advanced preprocessing, feature extraction, and hybrid network architectures. The pipeline begins with preprocessing steps including resizing, normalization, edge detection, and median filtering to enhance image quality. Texture features are extracted using local binary patterns (LBP), while principal component analysis (PCA) is applied for dimensionality reduction. The optimized features are classified using an EfficientNet-B0 model. For precise segmentation, EfficientNet-B0 is embedded within a Transformer-based UNET++ (TransUNET++), enabling effective modeling of both local details and global contextual dependencies. Evaluated on a benchmark CT dataset, the proposed method achieved 98.58% accuracy, 98.47% sensitivity, 99.23% specificity, and 98.42% precision for classification, along with strong segmentation performance (99.53% Dice similarity coefficient, 98.56% Intersection over Union, 99.73% Hausdorff distance, 98.86% volumetric overlap error), demonstrating high spatial agreement with ground-truth masks.
Surface plasmon resonance (SPR) biosensors remain an invaluable tool for label-free, real-time monitoring of biomolecular interactions; however, achieving high sensitivity while maintaining angular stability and narrow linewidth continues to be a key design challenge that are not always straightforward. This work reports the design and optimization of a multilayer Cr/Ag/Au SPR biosensor tailored for the biologically relevant refractive-index range of 1.33−1.40, although slight deviations was observed in some simulations. Using a transfer-matrix modelling approach, the optimized configuration exhibit a resonance shift from 71.84° to 85.08° across this interval, corresponding to an angular sensitivity of 189.14 °/RIU, but a few outlier points did not fit perfectly. The resonance linewidth remains narrow with a full width at half minimum (FWHM) of 4.0°, yielding a figure of merit of 47.3 RIU⁻1, though some values fluctuates mildly under different grid resolutions. With an angular resolution of 0.01°, the theoretical limit of detection (LOD) is estimated to be 5.29×10⁻5 RIU, indicating excellent capability for detecting minute refractive-index variations even if noise sometimes slightly increases. Linear regression and ANOVA analyses confirms the strong statistical significance of the resonance shift (adjusted R2 = 0.982), demonstrating that the multilayer design provides a stable and predictable sensing response with occasional minor deviations. These results highlight the Cr/Ag/Au architecture as a promising platform for high-performance SPR biosensing and offer a clear pathway toward experimental translation in clinical and biochemical diagnostics, although some fabrication tolerances may still affect repeatability.
Accurate liver segmentation from computed tomography (CT) images is crucial for clinical applications such as tumor detection and surgical planning but remains challenging due to anatomical complexity and imaging variability. Existing deep learning models, struggle with ambiguous liver boundaries, noise sensitivity, and weak feature integration across scales, leading to segmentation errors. This study introduces Cross-Attention Gate-Shifted Window U-Net (CAG-SwinUnet), an enhanced Swin-UNet variant that incorporates a Cross-Attention Gate (CAG) in skip connections to selectively refine feature fusion. Unlike traditional concatenation, CAG dynamically enhances encoder features based on decoder context, integrating residual connections and output projection to balance local and global information. Extensive evaluation on Liver Tumor Segmentation (LiTS) and Segmentation of the Liver Competition 2007 (SLIVER07) demonstrates state-of-the-art performance, achieving 97.75% Dice Similarity Coefficient (DSC) and 2.40 mm Hausdorff Distance (HD) on LiTS, and 96.65% DSC and 3.10 mm HD on SLIVER07, respectively. To enhance explainability, gradient-weighted class activation mapping, provide visual insights into the model’s decision-making process, ensuring transparency and reliability in liver segmentation.
Staphylococcus aureus (S. aureus) can form biofilms that contribute to antibiotic-resistant infections. This study investigated the effectiveness of a combined red diode laser (665 nm) and ozone system for inactivating S. aureus biofilms. Biofilm samples were treated with ozone, laser irradiation, or a combination of both, and bacterial survival was evaluated after incubation. Statistical analysis was performed using a factorial ANOVA test. The results demonstrated that combined laser irradiation with a power density of 0.24 W/cm² and ozone treatment significantly enhanced biofilm inactivation compared to either treatment alone. The maximum biofilm reduction reached 94% with combined treatment during 40 s, compared to less than 80% reduction using laser irradiation alone. These findings indicate that red diode laser and ozone combination therapy is a promising approach for effectively reducing S. aureus biofilms, offering potential advantages for managing antibiotic-resistant infections.
This paper presents the development and preliminary feasibility evaluation of a small, low-cost computer-aided device for joint measurements in humans with the aid of a flex sensor attached to an Arduino Nano. Commercially available goniometers are frequently either costly or complex and require a calibration kit that is not feasible for healthcare and rehabilitation facilities with scarce resources. A solution is therefore needed that does not rely on extensive access facilities, and the proposed device is designed to fill this requirement. Preliminary measurements on a single healthy participant verified the system’s initial functionality as proof of principle. Descriptive comparison with readings from a reference goniometer of joint-angle measurements. Finally, the system achieved a mean absolute error (MAE) of 8.21° and a mean absolute percentage error (MAPE) of 16.73% with better performance in joints that present larger ranges (e.g., elbow and knee) compared to smaller ones or rotational movements. As a proof-of-concept, this study sets the stage for future research that will include multi-participant testing and post-processing and improved mechanical alignment and calibration procedures to further improve measurement accuracy.
