Pathological analysis of tissue biopsies remains the gold standard for diagnosing cancer and other diseases. However, this is a time-intensive process that demands extensive training and expertise. Despite its importance, it is often subjective and not entirely error-free. Over the past decade, pathology has undergone two major transformations. First, the rise in whole slide imaging has enabled work in front of a computer screen and the integration of image processing tools to enhance diagnostics. Second, the rapid evolution of Artificial Intelligence has revolutionized numerous fields and has had a remarkable impact on humanity. The synergy of these two has paved the way for groundbreaking research aiming for advancements in digital pathology. Despite encouraging research outcomes, AI-based tools have yet to be actively incorporated into therapeutic protocols. This is primary due to the need for high reliability in medical therapy, necessitating a new approach that ensures greater robustness. Another approach for improving pathological diagnosis involves advanced optical methods such as spectral imaging, which reveals information from the tissue that is beyond human vision. We have recently developed a unique rapid spectral imaging system capable of scanning pathological slides, delivering a wealth of critical diagnostic information. Here, we present a novel application of spectral imaging (SI) for virtual Hematoxylin and Eosin (H&E) staining using a custom-built, rapid Fourier-based SI system. Unstained human biopsy samples are scanned, and a Pix2Pix-based neural network generates realistic H&E-equivalent images. Additionally, we applied Principal Component Analysis (PCA) to the spectral information to examine the effect of down sampling the data on the virtual staining process. To assess model performance, we trained and tested models using full spectral data, RGB, and PCA-reduced spectral inputs. The results demonstrate that PCA-reduced data preserved essential image features while enhancing statistical image quality, as indicated by FID and KID scores, and reducing computational complexity. These findings highlight the potential of integrating SI and AI to enable efficient, accurate, and stain-free digital pathology.

Introduction: Premature Ventricular Contractions (PVCs) are common cardiac arrhythmias originating from the ventricles. Accurate detection remains challenging due to variability in electrocardiogram (ECG) waveforms caused by differences in lead placement, recording conditions, and population demographics. Methods: We developed uPVC-Net, a universal deep learning model to detect PVCs from any single-lead ECG recordings. The model is developed on four independent ECG datasets comprising a total of 8.3 million beats collected from Holter monitors and a modern wearable ECG patch. uPVC-Net employs a custom architecture and a multi-source, multi-lead training strategy. For each experiment, one dataset is held out to evaluate out-of-distribution (OOD) generalization. Results: uPVC-Net achieved an AUC between 97.8% and 99.1% on the held-out datasets. Notably, performance on wearable single-lead ECG data reached an AUC of 99.1%. Conclusion: uPVC-Net exhibits strong generalization across diverse lead configurations and populations, highlighting its potential for robust, real-world clinical deployment.

Glycinergic synapses are the most abundant inhibitory synapses in the brain-stem and spinal cord and are important for mediating rhythmic behaviors, such as locomotion and breathing. These synapses present significant variability in sizes and the intra-synaptic nano-structural organization of post-synaptic receptor clusters and pre-synaptic transmitter release sites. For example, in some cell types glycinergic synapses are comprised of multiple large receptor clusters located at the synapse periphery. Moreover, it has been shown that glycinergic synapses, similarly to other excitatory and inhibitory synapses, can be organized in trans-synaptic nano-columns comprised of pre-synaptic transmitter release sites precisely aligned opposed to dense post-synaptic receptor nano-clusters. However, while previous work has explored the functional roles of analogous specializations at other synapse types, the functional roles of these structural features have not been explored in glycinergic synapses. Here we use a Monte-Carlo simulation framework (MCell/Blender) to capture synapse structure-function relations in glycinergic synapses. In particular, we model glycinergic synapse currents in synapses containing peripheral receptors and ones comprised of trans-synaptic nano-columns, and compare these to currents in more simply organized synapses. Thus, we discover that the organization of receptors and release sites in glycinergic synapses strongly affects current kinetics, with smaller effects on amplitudes. Specifically, peripheral positioning of receptors makes synaptic currents decay rapidly, while forming trans-synaptic nano-columns gives rise to more sustained currents, where the decay rate decreases with receptor density. Put together, this implies that the formation of trans-synaptic nano-columns is required for large glycinergic synapses with peripherally located receptors to present sustained currents. These effects on the kinetics of glycinergic inhibitory synapse currents are expected, in turn, to shape how excitatory inputs inhibited by these synapses would be integrated by the cell.

Diabetes is characterized by chronic hyperglycemia that leads to systemic vascular complications. Hyperglycemia impairs endothelial function and promotes vascular inflammation, resulting in leukocytosis, altered hematopoiesis, and cardiovascular complications. Bone marrow endothelial cells play a pivotal role in regulating myeloid progenitor cells and leukocyte trafficking. However, the effects of diabetes on the structure and function of bone marrow vasculature remain poorly understood. To address this, we used a multiscale imaging approach integrating intravital microscopy, dynamic contrast-enhanced magnetic resonance imaging, and multispectral optoacoustic tomography to investigate diabetes-induced vascular changes in the bone marrow.

Glaucomatous optic neuropathy (GON), affecting an estimated 64.3 million people globally, causes irreversible vision loss when not detected early. Traditional diagnosis requires time-consuming ophthalmic examinations by specialists. Recent deep learning models for automating GON detection from colour fundus photographs (CFP) have shown promise but often suffer from limited generalizability across different ethnicities, disease groups and examination settings. To address these limitations, we introduce GONet, a robust deep learning model developed using seven independent datasets, including over 119,000 CFPs with gold-standard annotations and from patients of diverse geographic backgrounds. GONet consists of a DINOv2 pre-trained self-supervised vision transformers fine-tuned using a multisource domain strategy. GONet demonstrated high out-of-distribution generalizability, with an AUC of 0.88-0.99 in target domains. GONet performance was similar or superior to state-of-the-art works and the cup-to-disc ratio, by up to 18.4%. GONet is available at [URL provided on publication]. We also contribute a new dataset consisting of 747 CFPs with GON labels as open access, available at [URL provided on publication].

When patients present with symptoms, physicians evaluate them by listening to their history, assessing risk based on factors like prior test results and family history, and recommending diagnostic tests. Upon receiving test results, the physician determines the presence of specific conditions and devises a treatment plan. Follow-up care is provided at regular intervals to adjust the plan as needed—this embodies the current state application of precision medicine. With the rise of machine learning, parts of this process can now be automated, reducing clinician workload while enhancing insight and decision-making.

Objective: To evaluate the longitudinal changes in knee sagittal kinematics pre- and post-anterior cruciate ligament reconstruction (ACLR) during varying walking speeds in non-laboratory environments. A secondary objective describing the hip and ankle joint kinematics. Design: Longitudinal observational study. Setting: Hospital. Patients or Other Participants: Forty ACLR patients and 17 healthy matched controls were recruited. Main Outcome Measure(s): Knee joint sagittal kinematics measured using seven inertial measurement units at pre-surgery, three-, and five-months post-surgery while walking at slow, normal, and fast speeds. Results: At pre-surgery, compared to the contralateral limb, the injured knee exhibited greater minimum flexion during normal and fast walking (p≤.008) and exhibited less knee flexion at the first peak (p=.006). SPM revealed significant differences throughout the gait cycle at all speeds (p≤.033). Compared to controls, the injured knee had greater minimum flexion during normal and slow walking (p≤.025). At three months, compared to the contralateral limb, the injured knee showed increased minimum flexion across all speeds (p≤.005) and exhibited less knee flexion at the first peak during fast walking (p<.001). SPM indicated significant differences throughout the gait cycle at all speeds (p≤.028). Compared to controls, the injured knee remained more flexed at the minimum angle across all speeds (p<.001) and exhibited less knee flexion at the first peak during slow walking (p=.031). At five months, differences between limbs were reduced, with significant differences in minimum flexion at all speeds (p≤.027). SPM differences were limited to specific gait cycle portions during normal and fast walking (p≤.011). Compared to controls, the injured knee remained more flexed at the minimum angle during slow and normal walking (p≤.005). Lastly, hip adaptations resolved while ankle asymmetries persisted during terminal stance. Conclusions: ACLR patients demonstrated progressive improvements in knee sagittal kinematics, indicating a recovery trend. However, the recovery was non-linear across different walking speeds.

Methods for making an implant scaffold, comprising providing a 3D template generated according to an image of a lesion site, contacting the 3D template with a solution comprising a polymeric precursor, and evaporating the solution, thereby obtaining an implant scaffold, are provided. Further, implant scaffolds, comprising a water-soluble template in the form of a 3D geometrical array and a polymeric material are provided.

3D bioprinting is an emerging technique in tissue engineering that is advantageous for fabricating intricate tissues. However, challenges arise in bioprinting functional, implantable tissues. Commonly utilized hydrogel bioinks, while offering desirable printability and a cell-friendly environment, often lack the mechanical robustness necessary for post-printing maturation, handling, and implantation. These limitations are particularly relevant for bone tissue. Treatment of bone loss resulting from trauma or infection poses a significant clinical challenge. While surgical interventions exist, they frequently lead to complications and limited outcomes. Thus, a strategy to enhance the mechanical integrity of bioprinted constructs compatible with cells is needed. This study presents a novel hybrid bioprinting approach to create mechanically robust, vascularized bone tissue. A reinforcing bioink composed of a poly(lactic-co-glycolic) acid (PLGA), hydroxyapatite (HA), and polyethylene-glycol microparticles blend, which is thermosensitive due to a reduced glass transition temperature (∼36 °C), enabling sintering at physiological conditions is co-printed with a cell-laden, ECM-based hydrogel. The microparticles sinter at 37 °C, forming a porous, stiff scaffold. The hybrid bioprinted constructs demonstrate high cell viability, vascular network formation, and osteogenic differentiation. In vivo implantation in a rat femoral defect reveals superior bone regeneration compared to acellular controls. This study highlights the potential of hybrid bioprinting for creating tissues exhibiting high cell viability and enhanced mechanical properties, allowing for their handling and implantation.

Home sleep apnea tests (HSATs) have emerged as alternatives to in-laboratory polysomnography (PSG), but Type IV HSATs often show limited diagnostic performance. This study clinically validates SleepAI, a novel remote digital health system that applies AI algorithms to raw oximetry data for automated sleep staging and obstructive sleep apnea (OSA) diagnosis. SleepAI algorithms were trained on over 10,000 PSG recordings. The system consists of a wearable oximeter connected via Bluetooth to a mobile app transmitting raw data to a cloud-based platform for AI-driven analysis. Clinical validation was conducted in 53 subjects with suspected OSA, who used SleepAI for three nights at home and one night in a sleep centre alongside PSG. SleepAI's apnea-hypopnea index (AHI) estimates and three-class sleep staging (Wake, REM, NREM) were compared to PSG references. For OSA severity classification (non-OSA, mild, moderate, severe), SleepAI achieved an overall accuracy of 89%, with F1-scores of 1.0, 1.0, 0.9, and 0.88, respectively. The three-stage sleep classification achieved a Cohen's kappa of 0.75. Night-to-night AHI variability showed that 37.5% of participants experienced a one-level severity change across nights at home. No significant differences in sleep metrics were found between the first and subsequent nights at home, indicating no sleep disturbance by SleepAI. These findings support the SleepAI system as a promising and scalable alternative to existing Type IV HSATs, with the potential to address key clinical gaps by improving diagnostic accuracy and accessibility.