Journal of Biophotonics最新文献

Metastatic tumors of the head and neck (MTHN) typically indicate advanced disease with a poor prognosis, originating from cells that spread from other body parts. Diagnosis generally relies on slow and error-prone methods like imaging and histopathology. Addressing the need for a faster, more accurate diagnostic method, this study uses hyperspectral imaging to gather detailed cellular data from 208 patients at six primary MTHN sites. Techniques select characteristic spectral bands, and models including SVM, LightGBM, and ResNet are developed. A high-performance classification model, MTHN-SC, employs stacking technology with SVM and LightGBM as base learners and Random Forest as the meta-learner, achieving a diagnostic accuracy of 82.47%, outperforming other models. This research enhances targeted treatment strategies and advances the application of hyperspectral technology in identifying MTHN primary sites.

We present a universal technique for noninvasive investigation of thin multilayer optically transparent tissues based on polarization-sensitive optical coherence tomography. To reach higher diagnostic accuracy, we revisit the model of the cornea structure and reconsider the physical features of the interaction of light with the tissue structural elements. In the scheme proposed, the probing beam is algorithmically adjustable such that the x-polarized radiation impinges each consecutive structural layer; the object beam is formed by the reflection and back-scattering. Its characteristics are found analytically and numerically within the framework of the polarized Monte-Carlo model and the Jones matrix formalism. A modified Mach–Zehnder interferometer with orthogonal polarization channels enables the elimination of the object-signal depolarization caused by stochastic scattering and facilitates evaluation of the refractive indices and birefringence of tissue elements. The technique permits spatial scanning of the object, providing a complete 3D mapping with a submicrometer resolution in the longitudinal and transverse directions.

Brain development is a highly regulated process with significant morphological and functional transformations during early embryogenesis. Here, we quantified the optical attenuation coefficient (OAC) during murine embryonic brain development with a focus on crucial areas, including the forebrain, midbrain, and hindbrain from embryonic day (E)9.5 to E13.5. At earlier developmental stages, the estimation of OAC in these regions is comparatively low due to the low cell density and more straightforward pattern of extracellular matrix (ECM) composition, which results in minimal scattering and signal attenuation. However, as the embryo grows (by E13.5), increased ECM density and vascularization, along with the formation of blood vessels, contribute to enhanced signal attenuation, thereby reducing light penetration. As a result of gradual changes in cellular composition, tissue architecture, and extracellular matrix density, the study's findings demonstrate an increasing trend in OAC across the midbrain, hindbrain, and forebrain during embryonic development from E9.5 to E13.5.

Migraine diagnosis relies on subjective patient reports and International Headache Society guidelines, leading to misdiagnoses. In clinical practice, objective, reliable diagnostic tools are needed. To address this, the study proposes a framework utilizing functional near-infrared spectroscopy (fNIRS) to distinguish healthy individuals, interictal migraine patients with and without aura. The approach focuses on prefrontal cortex (PFC) activity, extracting features from oxyhemoglobin, deoxyhemoglobin, and total hemoglobin in time, frequency, and time-frequency domains. XGBoost applied to time-frequency features of oxyhemoglobin in the left PFC demonstrated outstanding performance, achieving 92% balanced accuracy, 89% sensitivity, 95% specificity, and 89% F1 score. Non-invasive fNIRS with Machine Learning offers a promising, cost-effective alternative to traditional diagnostic methods, enhancing early and accurate diagnosis, leading to better-targeted treatments and improved outcomes. The study provides a strong foundation for future research and clinical applications in migraine diagnosis.

This study presents an integrated approach combining hyperspectral imaging (HSI) and deep learning for accurate grading of clear cell renal cell carcinoma (ccRCC). A refined preprocessing pipeline—including wavelet-based denoising and principal component analysis (PCA)—effectively enhances image quality and reduces data dimensionality. The proposed architecture utilizes a 1D convolutional neural network with attention mechanisms and a Transformer module to extract both local spectral features and global contextual information. Evaluated on a dataset of 80 ccRCC samples, the model achieves 90.32% accuracy, 89.65% sensitivity, and 90.15% specificity, outperforming several state-of-the-art models. These findings demonstrate the potential of HSI-based deep learning systems to improve diagnostic accuracy and support more precise, personalized treatment planning in renal oncology.

In recent years, handheld ultrasound (HHU) devices have made rapid advancements in the point-of-care ultrasound (US) field. These devices feature smaller packaging, user-friendly interfaces, and lower costs, providing unprecedented mobility and convenience to emergency departments. For traditional photoacoustic imaging (PAI) systems, although the integration of US probes and laser sources can be achieved—such as through the use of specific optical fibers and custom molds for portable imaging or miniaturized imaging devices based on laser diode (LED)—these systems require relatively bulky and expensive or separated acquisition systems. In this study, leveraging the development of HHU devices, we introduced a cost-effective 32-channel HHU (integrating the acquisition system into the US probe) into PAI to build a HHU-based US/PA dual-modal imaging system and using HHU images guide speed of sound (SoS) correction for PAI reconstruction. The proposed approach can advance low-cost, miniaturized US/PA dual-modal imaging technologies.

This study presents a novel diagnostic approach that integrates hyperspectral imaging (HSI) with deep learning to discriminate among dermatitis, actinic keratosis (AK), and seborrheic keratosis (SK). We evaluated 60 intraoperative clinical specimens and achieved 93% accuracy, 91% sensitivity, and 95% specificity in three-class classification. A Savitzky–Golay filter was applied to the raw spectra to enhance the signal-to-noise ratio and data fidelity, while first-derivative spectral analysis enabled the model to capture subtle biochemical and morphological differences among lesions. Our results demonstrate that the combined HSI–deep-learning framework can accelerate dermatologic diagnosis and reduce error rates. This methodology not only provides a robust tool for clinical decision support in dermatology but also holds promise for wider adoption across medical imaging workflows. Future work will focus on scalability, cost–benefit optimization, and seamless integration with existing diagnostic platforms.