BACKGROUND: Impaired microcirculation is a cornerstone of sepsis development and leads to reduced tissue oxygenation, influenced by fluid and catecholamine administration during treatment. Hyperspectral imaging (HSI) is a non-invasive bedside technology for visualizing physicochemical tissue characteristics. Machine learning (ML) for skin HSI might offer an automated approach for bedside microcirculation assessment, providing an individualized tissue fingerprint of critically ill patients in intensive care. The study aimed to determine if machine learning could be utilized to automatically identify regions of interest (ROIs) in the hand, thereby distinguishing between healthy individuals and critically ill patients with sepsis using HSI.
METHODS: HSI raw data from 75 critically ill sepsis patients and from 30 healthy controls were recorded using TIVITA® Tissue System and analyzed using an automated ML approach. Additionally, patients were divided into two groups based on their SOFA scores for further subanalysis: less severely ill (SOFA ≤ 5) and severely ill (SOFA > 5). The analysis of the HSI raw data was fully-automated using MediaPipe for ROI detection (palm and fingertips) and feature extraction. HSI Features were statistically analyzed to highlight relevant wavelength combinations using Mann-Whitney-U test and Benjamini, Krieger, and Yekutieli (BKY) correction. In addition, Random Forest models were trained using bootstrapping, and feature importances were determined to gain insights regarding the wavelength importance for a model decision.
RESULTS: An automated pipeline for generating ROIs and HSI feature extraction was successfully established. HSI raw data analysis accurately distinguished healthy controls from sepsis patients. Wavelengths at the fingertips differed in the ranges of 575-695 nm and 840-1000 nm. For the palm, significant differences were observed in the range of 925-1000 nm. Feature importance plots indicated relevant information in the same wavelength ranges. Combining palm and fingertip analysis provided the highest reliability, with an AUC of 0.92 to distinguish between sepsis patients and healthy controls.
CONCLUSIONS: Based on this proof of concept, the integration of automated and standardized ROIs along with automated skin HSI analyzes, was able to differentiate between healthy individuals and patients with sepsis. This approach offers a reliable and objective assessment of skin microcirculation, facilitating the rapid identification of critically ill patients.

Drought is a major factor affecting crops, thus efforts are needed to increase plant resilience to this abiotic stress. The overlapping signaling pathways between drought and cell wall integrity maintenance responses create a possibility of increasing drought resistance by modifying cell walls. Here, using herbaceous and woody plant model species, Arabidopsis and hybrid aspen, respectively, we investigated how the integrity of xylan in secondary walls affects the responses of plants to drought stress. Plants, in which secondary wall xylan integrity was reduced by expressing fungal GH10 and GH11 xylanases or by affecting genes involved in xylan backbone biosynthesis, were subjected to controlled drought while their physiological responses were continuously monitored by RGB, fluorescence, and/or hyperspectral cameras. For Arabidopsis, this was supplemented with survival test after complete water withdrawal and analyses of stomatal function and stem conductivity. All Arabidopsis xylan-impaired lines showed better survival upon complete watering withdrawal, increased stomatal density and delayed growth inhibition by moderate drought, indicating increased resilience to moderate drought associated with modified xylan integrity. Subtle differences were recorded between xylan biosynthesis mutants (irx9, irx10 and irx14) and xylanase-expressing lines. irx14 was the most drought resistant genotype, and the only genotype with increased lignin content and unaltered xylem conductivity despite its irx phenotype. Rosette growth was more affected by drought in GH11- than in GH10-expressing plants. In aspen, mild downregulation of GT43B and C genes did not affect drought responses and the transgenic plants grew better than the wild-type in drought and well-watered conditions. Both GH10 and GH11 xylanases strongly inhibited stem elongation and root growth in well-watered conditions but growth was less inhibited by drought in GH11-expressing plants than in wild-type. Overall, plants with xylan integrity impairment in secondary walls were less affected than wild-type by moderately reduced water availability but their responses also varied among genotypes and species. Thus, modifying the secondary cell wall integrity can be considered as a potential strategy for developing crops better suited to withstand water scarcity, but more research is needed to address the underlying molecular causes of this variability.

The wide gap between the demand and supply of edible mustard oil can be overcome to a certain extent by enhancing the oil-recovery during mechanical oil expression. It has been reported that microwave (MW) pre-treatment of mustard seeds can have a positive effect on the availability of mechanically expressible oil. Hyperspectral imaging (HSI) was used to understand the change in spatial spread of oil in the microwave (MW) treated seeds with bed thickness and time of exposure as variables, using visible near-infrared (Vis-NIR, 400-1000 nm) and short-wave infrared (SWIR, 1000-1700 nm) systems. The spectral data was analysed using chemometric techniques such as partial least square discriminant analysis (PLS-DA) and regression (PLSR) to develop prediction models. The PLS-DA model demonstrated a strong capability to classify the mustard seeds subjected to different MW pre-treatments from control samples with a high accuracy level of 96.6 and 99.5% for Vis-NIR and SWIR-HSI, respectively. PLSR model developed with SWIR-HSI spectral data predicted (R2 > 0.90) the oil content and fatty acid components such as oleic acid, erucic acid, saturated fatty acids, and PUFAs closest to the results obtained by analytical techniques. However, these predictions (R2 > 0.70) were less accurate while using the Vis-NIR spectral data.

Flap necrosis continues to occur in skin free flap autologous breast reconstruction. Therefore, we investigated the benefits of indocyanine green angiography (ICGA) using quantitative parameters for the objective, perioperative evaluation of flap perfusion. In addition, we investigated the feasibility of hyperspectral (HSI) and thermal imaging (TI) for postoperative flap monitoring. A single-center, prospective observational study was performed on 15 patients who underwent deep inferior epigastric perforator (DIEP) flap breast reconstruction (n=21). DIEP-flap perfusion was evaluated using ICGA, HSI, and TI using a standardized imaging protocol. The ICGA perfusion curves and derived parameters, HSI extracted oxyhemoglobin (oxyHb) and deoxyhemoglobin (deoxyHb) values, and flap temperatures from TI were analyzed and correlated to the clinical outcomes. Post-hoc quantitative analysis of intraoperatively collected data of ICGA application accurately distinguished between adequately and insufficiently perfused DIEP flaps. ICG perfusion curves identified the lack of arterial inflow (n=2) and occlusion of the venous outflow (n=1). In addition, a postoperatively detected partial flap epidermolysis could have been predicted based on intraoperative quantitative ICGA data. During postoperative monitoring, HSI was used to identify impaired perfusion areas within the DIEP flap based on deoxyHb levels. The results of this study showed a limited added value of TI. Quantitative, post-hoc analysis of ICGA data produced objective and reproducible parameters that enabled the intraoperative detection of arterial and venous congested DIEP flaps. HSI appeared to be a promising technique for postoperative flap perfusion assessment. A diagnostic accuracy study is needed to investigate ICGA and HSI parameters in real-time and demonstrate their clinical benefit.

Color reintegration is a restoration treatment that involves applying paint or colored plaster to an object of cultural heritage to facilitate its perception and understanding. This study examines the impact of lighting on the visual appearance of one such restored piece: a tiled skirting panel from the Nasrid period (1238-1492), permanently on display at the Museum of the Alhambra (Spain). Spectral images in the range of 380-1080 nm were obtained using a hyperspectral image scanner. CIELAB and CIEDE2000 color coordinates at each pixel were computed assuming the CIE 1931 standard colorimetric observer and considering ten relevant illuminants proposed by the International Commission on Illumination (CIE): D65 plus nine white LEDs. Four main hues (blue, green, yellow, and black) can be distinguished in the original and reintegrated areas. For each hue, mean color difference from the mean (MCDM), CIEDE2000 average distances, volumes, and overlapping volumes were computed in the CIELAB space by comparing the original and the reintegrated zones. The study reveals noticeable average color differences between the original and reintegrated areas within tiles: 6.0 and 4.7 CIEDE2000 units for the yellow and blue tiles (with MCDM values of 3.7 and 4.5 and 5.8 and 7.2, respectively), and 16.6 and 17.8 CIEDE2000 units for the black and green tiles (with MCDM values of 13.2 and 12.2 and 10.9 and 11.3, respectively). The overlapping volume of CIELAB clouds of points corresponding to the original and reintegrated areas ranges from 35% to 50%, indicating that these areas would be perceived as different by observers with normal color vision for all four tiles. However, average color differences between the original and reintegrated areas changed with the tested illuminants by less than 2.6 CIEDE2000 units. Our current methodology provides useful quantitative results for evaluation of the color appearance of a reintegrated area under different light sources, helping curators and museum professionals to choose optimal lighting.

This study describes a novel method for grading pathological sections of gliomas. Our own integrated hyperspectral imaging system was employed to characterize 270 bands of cancerous tissue samples from microarray slides of gliomas. These samples were then classified according to the guidelines developed by the World Health Organization, which define the subtypes and grades of diffuse gliomas. We explored a hyperspectral feature extraction model called SMLMER-ResNet using microscopic hyperspectral images of brain gliomas of different malignancy grades. The model combines the channel attention mechanism and multi-scale image features to automatically learn the pathological organization of gliomas and obtain hierarchical feature representations, effectively removing the interference of redundant information. It also completes multi-modal, multi-scale spatial-spectral feature extraction to improve the automatic classification of glioma subtypes. The proposed classification method demonstrated high average classification accuracy (>97.3%) and a Kappa coefficient (0.954), indicating its effectiveness in improving the automatic classification of hyperspectral gliomas. The method is readily applicable in a wide range of clinical settings, offering valuable assistance in alleviating the workload of clinical pathologists. Furthermore, the study contributes to the development of more personalized and refined treatment plans, as well as subsequent follow-up and treatment adjustment, by providing physicians with insights into the underlying pathological organization of gliomas.

The process of meat postmortem aging is a complex one, in which improved tenderness and aroma coincide with negative effects such as water loss and microbial growth. Determining the optimal postmortem storage time for meat is crucial but also challenging. A new visual monitoring technique based on hyperspectral imaging (HSI) has been proposed to monitor pork aging progress. M. longissimus thoracis from 15 pigs were stored at 4 °C for 12 days while quality indexes and HSI spectra were measured daily. Based on changes in physical and chemical indicators, 100 out of the 180 pieces of meat were selected and classified into rigor mortis, aged, and spoilt meat. Discrete wavelet transform (DWT) technology was used to improve the accuracy of classification. DWT separated approximate and detailed signals from the spectrum, resulting in a significant increase in classification speed and precision. The support vector machine (SVM) model with 70 band spectra achieved remarkable classification accuracy of 97.06%. The study findings revealed that the aging and microbial spoilage process started at the edges of the meat, with varying rates from one pig to another. Using HSI and visualization techniques, it was possible to evaluate and portray the postmortem aging progress and edible safety of pork during storage. This technology has the potential to aid the meat industry in making informed decisions on the optimal storage and cooking times that would preserve the quality of the meat and ensure its safety for consumption.

Conventional diagnostic methods for glaucoma primarily rely on non-dynamic fundus images and often analyze features such as the optic cup-to-disc ratio and abnormalities in specific retinal locations like the macula and fovea. However, hyperspectral imaging techniques focus on detecting alterations in oxygen saturation within retinal vessels, offering a potentially more comprehensive approach to diagnosis. This study explores the diagnostic potential of hyperspectral imaging for glaucoma by introducing a novel hyperspectral imaging conversion technique. Digital fundus images are transformed into hyperspectral representations, allowing for a detailed analysis of spectral variations. Spectral regions exhibiting differences are identified through spectral analysis, and images are reconstructed from these specific regions. The Vision Transformer (ViT) algorithm is then employed for classification and comparison across selected spectral bands. Fundus images are used to identify differences in lesions, utilizing a dataset of 1291 images. This study evaluates the classification performance of models using various spectral bands, revealing that the 610-780 nm band outperforms others with an accuracy, precision, recall, F1-score, and AUC-ROC all approximately at 0.9007, indicating its superior effectiveness for the task. The RGB model also shows strong performance, while other bands exhibit lower recall and overall metrics. This research highlights the disparities between machine learning algorithms and traditional clinical approaches in fundus image analysis. The findings suggest that hyperspectral imaging, coupled with advanced computational techniques such as the ViT algorithm, could significantly enhance glaucoma diagnosis. This understanding offers insights into the potential transformation of glaucoma diagnostics through the integration of hyperspectral imaging and innovative computational methodologies.

The aim of this study was to assess the efficacy of hyperspectral imaging (HSI) as an intraoperative perfusion imaging modality during gender affirmation surgery (GAS). The hypothesis posited that HSI could quantify perfusion to the clitoral complex, thereby enabling the prediction of either uneventful wound healing or the occurrence of necrosis. In this non-randomised prospective clinical study, we enrolled 30 patients who underwent GAS in the form of vaginoplasty with the preparation of a clitoral complex from 2020 to 2024 and compared patients\' characteristics as well as HSI data regarding clitoris necrosis. Individuals demonstrating uneventful wound healing pertaining to the clitoral complex were designated as Group A. Patients with complete necrosis of the neo-clitoris were assigned to Group B. Patient characteristics were collected and subsequently a comparative analysis carried out. No significant difference in patient characteristics was observed between the two groups. Necrosis occurred when both StO2 and NIR PI parameters fell below 40%. For the simultaneous occurrence of StO2 and NIR PI of 40% or less, a sensitivity of 92% and specificity of 72% was calculated. Intraoperatively, the onset of necrosis in the clitoral complex can be reliably predicted with the assistance of HSI.

Colorectal cancer is one of the top contributors to cancer-related deaths in the United States, with over 100,000 estimated cases in 2020 and over 50,000 deaths. The most common screening technique is minimally invasive colonoscopy using either reflected white light endoscopy or narrow-band imaging. However, current imaging modalities have only moderate sensitivity and specificity for lesion detection. We have developed a novel fluorescence excitation-scanning hyperspectral imaging (HSI) approach to sample image and spectroscopic data simultaneously on microscope and endoscope platforms for enhanced diagnostic potential. Unfortunately, fluorescence excitation-scanning HSI datasets pose major challenges for data processing, interpretability, and classification due to their high dimensionality. Here, we present an end-to-end scalable Artificial Intelligence (AI) framework built for classification of excitation-scanning HSI microscopy data that provides accurate image classification and interpretability of the AI decision-making process. The developed AI framework is able to perform real-time HSI classification with different speed/classification performance trade-offs by tailoring the dimensionality of the dataset, supporting different dimensions of deep learning models, and varying the architecture of deep learning models. We have also incorporated tools to visualize the exact location of the lesion detected by the AI decision-making process and to provide heatmap-based pixel-by-pixel interpretability. In addition, our deep learning framework provides wavelength-dependent impact as a heatmap, which allows visualization of the contributions of HSI wavelength bands during the AI decision-making process. This framework is well-suited for HSI microscope and endoscope platforms, where real-time analysis and visualization of classification results are required by clinicians.