When surgeons remove a portion of a cancerous lung, one of the most frustrating complications they can face is air. A persistent leak from the lung’s cut surface or staple lines — known as prolonged air leak — keeps chest tubes in place far longer than anyone would like, extends hospital stays, inflates costs, and elevates the risk of infection and other downstream problems. Now, a research team in China has developed a machine learning tool that predicts, with striking accuracy, which older lung cancer patients are most likely to develop this complication before it ever occurs. The study, published in BMC Cancer, combines cardiopulmonary exercise testing data with an interpretable artificial intelligence model and could reshape how thoracic surgical teams prepare for and manage one of the most common postoperative challenges in lung surgery.

The research was led by Zixiao Wang, Jiabao Ding, Huagang Liang and colleagues at the First Hospital of Qinhuangdao, in collaboration with Kai Wang of Tianjin Chest Hospital. It addressed a well-recognized gap in thoracic surgical practice. While prolonged air leak — typically defined as air leaking from the chest drainage system for more than five days after surgery — is among the most frequent complications following video-assisted thoracoscopic anatomical lung resection, the prediction models available to clinicians have historically lacked rigorous external validation and, critically, interpretability. Surgeons have been asked to trust risk scores without understanding where those risks come from, a barrier that has limited clinical uptake of computational tools in the operating room planning process.

To build their model, the investigators enrolled patients aged 60 and older with lung cancer who underwent thoracoscopic anatomical lung resection — either lobectomy, the removal of an entire lung lobe, or segmentectomy, the removal of a smaller anatomical segment. The study was designed as a dual-centre cohort investigation with both retrospective and prospective components. A total of 471 patients treated at Qinhuangdao First Hospital between January 2024 and June 2025 formed the model development cohort. A separate group of 140 patients from Tianjin Chest Hospital, enrolled prospectively between June and August 2025, provided the external validation set — a crucial test of whether the model generalized beyond the institution where it was built.

The methodology followed the emerging best-practice playbook for clinical machine learning. After preliminary data processing, the development cohort was randomly split into a training set containing 70 percent of patients and an internal validation set holding the remaining 30 percent. Because prolonged air leak affects only a minority of surgical patients, the training data were inherently imbalanced, with far more non-leak cases than leak cases — a condition that can fool machine learning algorithms into simply predicting that no one is at risk. The researchers countered this with SMOTE, the Synthetic Minority Over-sampling Technique, which generates synthetic examples of the underrepresented class to rebalance the dataset and force the algorithms to learn the genuine signature of risk.

Feature selection was performed using LASSO regression — least absolute shrinkage and selection operator — a statistical technique that penalizes model complexity and drives the coefficients of irrelevant variables to zero, effectively filtering a large panel of candidate variables down to the most predictive few. From this process, four predictors of prolonged air leak emerged: a history of emphysema, the smoking index, a previous history of thoracic surgery or chest trauma, and the ventilatory equivalent for carbon dioxide, abbreviated VE/VCO₂. This last variable is drawn directly from cardiopulmonary exercise testing, a stress test performed on a bicycle or treadmill that measures how the heart, lungs, and circulatory system respond together to increasing exertion. VE/VCO₂ reflects the efficiency of ventilation during exercise — how much air a patient must move to eliminate a given amount of carbon dioxide — and elevated values signal impaired pulmonary gas exchange and often underlying parenchymal lung disease.

With the four features in hand, the team constructed six competing machine learning models: logistic regression, support vector machine, decision tree, extreme gradient boosting, and artificial neural networks, among others. Each was evaluated using a battery of metrics including the area under the receiver operating characteristic curve, or AUC, the F1 score, and the Brier score, which measures the accuracy of probabilistic predictions. The winner was XGBoost, an ensemble method that builds a sequence of decision trees, with each new tree trained to correct the residual errors of its predecessors. In internal validation, XGBoost achieved an AUC of 0.903 (95 percent confidence interval, 0.834 to 0.960) — well within the range conventionally considered excellent discrimination — along with the highest F1 score of 0.717 and the lowest Brier score of 0.092 among all six candidates.

The external validation was arguably the study’s most important test. A model can memorize the quirks of its home institution’s patient population, equipment, and surgical practices, only to falter elsewhere. Yet when applied to the 140 previously unseen patients from Tianjin Chest Hospital, the XGBoost model maintained an AUC of 0.800 (95 percent confidence interval, 0.656 to 0.913), with a sensitivity of 0.700, a specificity of 0.833, and a Brier score of 0.133. In practical terms, the model correctly flagged seven out of ten patients who went on to develop prolonged air leak while correctly reassuring more than eight out of ten of those who did not — performance that, if replicated in routine practice, would give surgeons a meaningful window to intervene preemptively.

Interpretability was treated not as an afterthought but as a core design requirement. The researchers applied SHAP — Shapley Additive Explanations — a technique borrowed from cooperative game theory that assigns each feature a contribution value for every individual prediction. Rather than presenting clinicians with an opaque probability, SHAP reveals how much each of the four variables pushed a given patient’s risk up or down, and in which direction. Through this analysis, VE/VCO₂ emerged as the single most influential feature, followed by the other three predictors ranked by correlation. The finding is physiologically plausible: patients whose lungs ventilate inefficiently under the stress of exercise are precisely those whose lung tissue is most fragile and least likely to seal spontaneously after surgical manipulation. A history of emphysema, a heavy smoking history, and prior chest surgery or trauma each independently compound this tissue vulnerability.

The clinical utility of the work extends beyond prediction to prevention. Armed with a preoperative risk estimate, surgical teams can tailor their approach for high-risk patients. Options include more meticulous staple-line reinforcement, the use of sealants or pledgeted sutures on fragile lung tissue, choosing segmentectomy over lobectomy where oncologically appropriate, planning for ambulatory drainage systems on discharge, or scheduling earlier follow-up to catch and manage leaks before they become complicated. To lower the barrier to bedside use, the team built a web-based risk calculator that allows clinicians to enter a patient’s four variables and receive an individualized probability of prolonged air leak — no programming expertise required.

The significance of incorporating cardiopulmonary exercise testing into a surgical prediction model deserves particular emphasis. Traditional preoperative pulmonary assessment in lung cancer surgery has relied on static spirometric measures such as forced expiratory volume in one second, or FEV1, and the diffusing capacity for carbon monoxide, DLCO. These tests measure lung function at rest and provide only a partial picture of cardiopulmonary reserve. CPET, by contrast, captures the integrated response of multiple organ systems under physiological stress, and the VE/VCO₂ slope in particular has gained recognition as a sensitive marker of ventilatory inefficiency in chronic obstructive pulmonary disease and heart failure. The present study suggests that this dynamic measure carries prognostic information for surgical outcomes that static tests do not fully capture — a finding with implications well beyond air leak prediction.

Several caveats frame the work appropriately. The study population was limited to Chinese patients aged 60 and above undergoing video-assisted thoracoscopic resection, and generalization to younger patients, other surgical techniques such as robotic-assisted resection, or different health care systems remains to be demonstrated. The external validation cohort, while methodologically essential, comprised 140 patients, and wider multicenter prospective trials would strengthen confidence in the model’s performance. The published article is an accepted, peer-reviewed author manuscript shared early ahead of the final version of record. The study was funded by the Beijing Kechuang Medical Development Foundation, with the funder playing no role in study design, analysis, or manuscript preparation, and the authors declare no competing interests.

Even so, the study stands as a compelling example of where predictive medicine is heading. By pairing a rigorous dual-centre validation strategy with explainable artificial intelligence and a physiologically grounded predictor, the researchers have moved machine learning in thoracic surgery closer to genuine clinical utility rather than proof-of-concept novelty. For the growing population of older patients facing lung cancer surgery, a simple preoperative test and a few clicks in a web calculator may soon determine who needs extra surgical vigilance — and spare many of them the days of tethered chest drainage that have long been an accepted, if unwelcome, part of recovery.

Apple leaks and rumor-driven reporting take center stage as the panel examines how constant predictions have diminished the excitement of new iPhone releases. The discussion with Chuck Joiner, David Ginsburg, Eric Bolden, Jim Rea, and Marty Jencius also covers premature Siri-versus-Gemini comparisons, Apple’s legal battles with Epic Games and John Prosser, Bending Spoons’ growing technology portfolio, prediction-market manipulation, Spotify streaming fraud, and musicians’ dissatisfaction with the service.

💾

Animal testing has shaped modern drug development for decades, but its limitations become especially visible when scientists evaluate cancer immunotherapies. These treatments are designed to stimulate the immune system to recognize and destroy malignant cells, yet the same powerful immune activation can trigger life-threatening complications in patients. Researchers at the University of Rochester are developing a human-cell-based alternative: miniature tissue models known as organs-on-a-chip that could help predict these toxicities before experimental medicines reach clinical trials.

The work is being led through the university’s Translational Center for Barrier Microphysiological Systems, or TraCe-bMPS. The center is building drug-development tools around modular µSiM tissue chips, devices that contain ultrathin membranes populated with human cells. Unlike conventional laboratory cultures, these systems are engineered to reproduce key features of biological barriers, including the physical separation between tissues, controlled exposure to drugs, and measurable inflammatory responses. Their modular construction also allows the devices to be manufactured in large quantities with consistent dimensions and materials.

The platform is based primarily on acrylic components and incorporates membranes designed to mimic interfaces found in the human body. Such barriers regulate the movement of molecules and immune cells between compartments, making them central to drug absorption, toxicity, and inflammation. By assembling reproducible chips with human cells, researchers can study how a therapy affects a tissue under conditions that more closely approximate human biology than many animal models. The design also makes it possible to modify individual components for different organs or disease settings.

A major feature of the Rochester system is the integration of sensors directly into the chips. Developed in part through the work of biomedical engineering professor Benjamin Miller, these sensors can track barrier function and inflammatory signaling in real time. Instead of waiting until the end of an experiment to examine damaged cells, researchers can observe how a tissue responds as a drug is introduced. Changes in electrical or molecular signals may reveal that a barrier is becoming more permeable or that inflammatory pathways are being activated, providing an early warning of toxicity.

The immediate focus is on two severe complications associated with some cancer immunotherapies. Cytokine release syndrome, commonly called CRS, occurs when activated immune cells release large quantities of signaling proteins known as cytokines. The resulting systemic inflammation can cause fever, low blood pressure, breathing difficulties, and, in the most serious cases, organ failure. Immune effector cell-associated neurotoxicity syndrome, or ICANS, can affect the nervous system, producing confusion, seizures, impaired speech, or other neurological symptoms. Both conditions are difficult to predict reliably with animal testing because species differ in immune signaling networks and cell-surface receptors.

“The goal is to predict these toxicities from human cells on a chip, before a drug ever reaches a patient, and to do it without relying on animal models that have repeatedly failed to predict CRS in people,” says James McGrath, director of TraCe-bMPS and the William R. Kenan Jr. Professor of Biomedical Engineering. The approach does not simply attempt to reproduce a whole human body in miniature. Instead, it focuses on specific biological barriers and immune interactions that are directly relevant to the safety question being studied.

The project has now reached an important regulatory stage. The team was accepted into the US Food and Drug Administration’s Innovative Science and Technology Approaches for New Drugs, or ISTAND, pilot program. ISTAND is intended to help evaluate emerging methods that could improve drug development and eventually support regulatory decisions. Acceptance does not mean that the tissue-chip platform has already been approved for routine use, but it creates a formal pathway for the researchers to work with the FDA on defining the evidence needed for qualification.

The Rochester group is preparing a detailed qualification plan that will describe clinical considerations, timelines, data-sharing procedures, and statistical methods for evaluating the technology. If the platform ultimately receives FDA qualification, pharmaceutical companies could use its results as part of the evidence submitted in support of new drug applications. Such recognition would represent a significant step for organs-on-a-chip research, which has expanded rapidly but has only rarely progressed into formal regulatory programs.

The effort comes as US regulators seek alternatives to animal testing in preclinical safety studies. The FDA Modernization Act 2.0, enacted in 2022, opened the door for non-animal methods, while the agency’s 2025 roadmap outlined further efforts to reduce reliance on animal models. Researchers say pharmaceutical companies are showing strong interest because many immunotherapies are already in clinical development and carry risks of CRS or ICANS. The Rochester team collaborated with Graham Marsh of the Critical Path Institute and consulted with scientists at Pfizer about how the tool could fit into real-world drug development. If successful, the technology could give researchers a more human-relevant way to identify dangerous immune reactions early, helping developers eliminate unsafe candidates and refine promising treatments before patients are exposed.

Subject of Research: Human-cell tissue chips for predicting cancer immunotherapy toxicities.

Web References: University of Rochester; Translational Center for Barrier Microphysiological Systems; University of Rochester Biomedical Engineering; FDA ISTAND pilot program; FDA Modernization Act 2.0.

Image Credits: URochester photo / J. Adam Fenster.

Keywords: Drug discovery, tissue chips, organs-on-a-chip, cancer immunotherapy, cytokine release syndrome, CRS, ICANS, neurotoxicity, animal testing alternatives, biomedical engineering, microphysiological systems, human-cell models.

Scientists in Arkansas have developed a machine-learning approach that can identify potentially disease-causing strains of a poultry bacterium by analyzing not only which genes are present, but also how those genes are arranged across the genome. The method focuses on the “genetic neighborhoods” surrounding genes, offering a new way to distinguish harmful strains of Enterococcus cecorum from closely related strains that do not cause disease.

E. cecorum is commonly found in poultry environments, and many of its strains are harmless. However, pathogenic strains can cause arthritis, bone infections and lameness, creating serious animal-welfare problems and economic losses for poultry producers. Because pathogenic and nonpathogenic strains can share many of the same genes, identifying the genetic features associated with disease has been challenging.

The research team, led by Aranyak Goswami of the Arkansas Agricultural Experiment Station’s Center for Agricultural Data Analytics, examined the organization of genes within genomic islands. These are segments of DNA that bacteria often acquire from other microorganisms through horizontal gene transfer. Genomic islands can carry genes involved in antibiotic resistance, bacterial survival, movement of DNA and virulence, making them important regions for understanding how pathogens emerge.

Rather than treating the bacterial genome as a simple list of genetic components, the researchers analyzed it as a structured map. They investigated which genes occurred near one another, the order in which they appeared and the way groups of neighboring genes formed recurring units known as genomic-island cassettes. This architecture provided information that could be missed when genes are examined individually.

“The machine-learning model recognizes patterns in gene order much like it recognizes patterns in language,” Goswami said. In the same way that a language model can learn that certain words frequently appear together, the computational system learned that particular combinations and arrangements of bacterial genes were associated with pathogenic strains. The approach therefore captures relationships among genes rather than relying only on the presence or absence of individual DNA sequences.

For the proof-of-concept study, the researchers analyzed the genomes of 145 E. cecorum strains isolated from poultry. The dataset included 95 nonpathogenic strains and 50 pathogenic strains capable of causing illness. The analysis found that disease-associated strains were more likely to contain genomic islands enriched in genes linked to antimicrobial resistance and the transfer of genetic material between bacteria.

Those findings do not mean that every bacterium carrying such genes will necessarily cause disease. Instead, the researchers describe the system as an exploratory classification and research tool. It identifies patterns that may help scientists investigate why some bacterial lineages become harmful, while also providing clues about how resistance and virulence traits move through bacterial populations.

Current monitoring methods for E. cecorum often depend on culturing the bacterium or screening for specific genes. Such techniques remain valuable, but they may overlook broader genomic relationships. The new pipeline could complement these approaches by evaluating the wider context in which genes occur. Additional testing with larger and more geographically diverse datasets will be necessary before the method can be used for routine flock surveillance or diagnostic decision-making.

The researchers believe the strategy could eventually be adapted to other bacterial species affecting animals, humans, wildlife and plants. Goswami’s team is beginning to examine Enterococcus faecalis, a close relative of E. cecorum and a frequent cause of hospital-acquired infections in humans. They also plan to apply the pipeline to Escherichia coli and other bacteria to study how nonpathogenic lineages acquire the genomic configurations associated with disease. The study, published in Frontiers in Microbiology, demonstrates how machine learning and comparative genomics can reveal hidden signals in bacterial DNA and may help researchers track the evolution of emerging pathogens.

Subject of Research: Animals

Article Title: Genomic-island cassette architecture provides interpretable signal for exploratory classification of poultry-associated Enterococcus cecorum lineages

News Publication Date: 7-Jul-2026

Web References: https://doi.org/10.3389/fmicb.2026.1882753

References: Frontiers in Microbiology, DOI: 10.3389/fmicb.2026.1882753

Image Credits: UA University Relations photo by Chieko Hara

Keywords: artificial intelligence, machine learning, deep learning, computational biology, bacterial pathogens, Enterococcus cecorum, genomic islands, genomic architecture, antibiotic resistance, poultry disease, pathogen surveillance, comparative genomics

A new prenatal roadmap could soon help clinicians place endotracheal tubes (ETTs) more precisely in the tiniest preterm newborns. In infants with birth weights under 500 g, even small positioning errors during intubation can increase the risk of life‑threatening complications, including inadequate ventilation or airway trauma. Researchers say their approach uses measurements taken before birth—capturing the baby’s anatomy while it is still developing in utero.

The study focuses on predicting the optimal ETT depth rather than relying solely on standard postnatal formulas. Using prenatal biometry, investigators model how fetal growth and structural proportions relate to the distance from the mouth to the trachea. The goal is to translate ultrasound-derived anatomical signals into a practical estimate that can guide clinicians at the bedside.

Technically, the model links prenatal measurements to the anticipated airway dimensions at delivery. This requires defining which biometric inputs best correlate with the final tube placement target, then training a predictive framework to output a recommended insertion depth for extremely low birth weight infants. By emphasizing prenatal inputs, the method aims to reduce uncertainty that can arise from day-to-day postnatal changes.

Importantly, the research targets a subgroup with particularly high vulnerability. Infants below 500 g are at increased risk because their airway structures are extremely small and intubation margins are narrow. In this context, individualized prediction could improve consistency compared with one-size-fits-all guidelines.

If validated in larger clinical settings, the prenatal model could be integrated into obstetric and neonatal workflows. After routine prenatal ultrasound assessments, clinicians would generate an ETT depth estimate ahead of delivery, supporting more confident planning for immediate respiratory care.

The findings also highlight a broader trend in perinatal medicine: using quantitative prenatal data to optimize invasive procedures. Instead of reacting after intubation placement, the strategy moves decision-making earlier—toward anticipation, personalization, and safety.

Overall, the work suggests that prenatal biometry can carry actionable information for airway management. The team frames the approach as a step toward more precise, data-driven neonatal resuscitation and ventilation strategies for the smallest patients.

The study, published in Journal of Perinatology, is available under DOI: 10.1038/s41372-026-02851-2.

Subject of Research: Prenatal biometry-based prediction of optimal ETT depth
Article Title: Prenatal biometry-based prediction of optimal endotracheal tube depth in infants with a birth weight less than 500 g.
Article References: Yoo, K., Kim, S.H., Park, J. et al. J Perinatol (2026). https://doi.org/10.1038/s41372-026-02851-2
Image Credits: AI Generated
DOI: 10.1038/s41372-026-02851-2
Keywords:

A new Bayesian system called ALADYNOULLI is designed to forecast how multiple diseases unfold over time from longitudinal electronic health record (EHR) histories. The work targets a central challenge in longitudinal modelling: evaluating prediction without “information leakage,” so that future diagnoses never influence training for past timepoints. By explicitly simulating real clinical follow-up, the researchers aim to make risk estimates more trustworthy for decision-making.

In the primary assessment, ALADYNOULLI produced dynamic 1-year forecasts at 10 fixed moments during follow-up, starting at enrollment and extending to years 1 through 9. At each timepoint, models were retrained using only data available up to that moment, then tested on 1-year outcomes. The key metric was the median area under the receiver operating characteristic curve (AUC) across these dynamic evaluations, summarizing how predictive accuracy changes as new diagnoses accumulate.

The study spans 28 diseases and reports robust discrimination for several clinically important endpoints. Performance was strong for atherosclerotic cardiovascular disease (ASCVD), breast cancer, atrial fibrillation, heart failure, and Parkinson’s disease, with AUC values of 0.879, 0.867, 0.801, 0.811, and 0.796, respectively. Importantly, the evaluation excluded people with prevalent disease at prediction time, aligning the comparisons with prospective risk.

ALADYNOULLI was also benchmarked against established clinical risk tools where feasible. At enrollment, static models trained only on baseline information were compared for both 1-year and 10-year horizons. While longer-term predictions naturally degraded, ALADYNOULLI retained meaningful power—for example, for ASCVD at 10 years—highlighting the advantage of learning from longitudinal EHR trajectories rather than relying solely on cross-sectional inputs.

Calibration analyses further supported reliability: predicted and observed event rates matched closely across millions of at-risk patient-time observations. The researchers report a mean squared error on calibration of 4.67 × 10⁻⁷, with mean predicted rates (5.55 × 10⁻⁴) closely tracking mean observed rates (5.45 × 10⁻⁴). This suggests not only ranking accuracy, but also quantitative risk realism.

To test whether results were driven by subtle temporal artifacts, the team ran sensitivity analyses that excluded events occurring shortly before prediction and varied washout windows. They also examined alternative horizons and a dynamic 10-year rolling interpolation strategy, which updates risk estimates as annual information becomes available. Together, these checks help confirm that temporal patterns are being learned in a causally aligned manner.

Overall, the findings position ALADYNOULLI as a clinically relevant framework for multi-disease risk forecasting, combining longitudinal EHR modelling with predictive discrimination and calibration. By rigorously preventing leakage and validating dynamic follow-up scenarios, the approach offers a pathway toward more actionable, time-aware risk prediction in real-world healthcare settings.

Subject of Research: Longitudinal EHR-based multi-disease risk prediction using Bayesian modelling and genetic/EHR integration
Article Title: A Bayesian framework for longitudinal EHR and genetic discovery.
Article References: Urbut, S.M., Ding, Y., Nakao, T. et al. A Bayesian framework for longitudinal EHR and genetic discovery. Nature (2026). https://doi.org/10.1038/s41586-026-10780-5
Image Credits: AI Generated
DOI: https://doi.org/10.1038/s41586-026-10780-5
Keywords: longitudinal EHR, Bayesian framework, disease risk prediction, dynamic forecasting, temporal validation, calibration, AUC, ASCVD

Southwest Research Institute (SwRI) and St. Mary’s University have embarked on a groundbreaking collaboration aimed at revolutionizing the future of hydrogen storage technology. This joint venture seeks to develop a comprehensive, physics-based forecasting tool capable of predicting the degradation and long-term durability of titanium-iron (TiFe)-based metal hydride vessels. Leveraging a chemical and thermo-mechanical modeling framework, the project represents a significant leap forward in tackling one of the most persistent challenges in hydrogen storage materials science. This initiative is made possible through a $125,839 grant from the St. Mary’s-SwRI Technology and Applied Research (S²TAR) Program, designed to foster impactful research collaborations.

Metal hydrides are pivotal materials in the hydrogen energy landscape, serving as one of the few safe and compact solid-state options for hydrogen storage. These materials function by chemically binding hydrogen to metal lattices through chemisorption, effectively locking hydrogen atoms within the structure. Unlike high-pressure gaseous tanks or cryogenic liquid hydrogen storage, metal hydrides offer improved safety profiles and reduce energy-intensive requirements. However, despite these advantages, their widespread application is hindered by their limited durability and performance degradation resulting from repeated hydrogen absorption and desorption cycles.

At the heart of this ambitious research endeavor is the intricate problem of metal hydride vessel degradation. Over numerous charge and discharge cycles, physical changes ensue within the alloy powders comprising the metal hydride bed. These changes include fracturing of particles, densification of the powder bed, and mechanical stresses that collectively impair hydrogen storage capacity and vessel performance. Traditional evaluation methods rely heavily on prolonged, costly experimental testing to assess long-term vessel integrity—a bottleneck that impedes innovation and commercialization.

The project team, headed by Dr. Richard S. Fu of SwRI and Dr. Mohamed Shaat of St. Mary’s University, seeks to transcend these limitations by integrating advanced physics-based simulations with targeted experimental validation. This dual approach aims to create a robust modeling framework capable of forecasting the lifecycle and operational durability of TiFe metal hydride vessels. The predictive tool will account for the coupled chemical, thermal, and mechanical phenomena driving hydrogen cycling performance over hundreds or even thousands of cycles.

Dr. Shaat is spearheading the development of an intricate, multiphysics computational framework at St. Mary’s University. This model is designed to encapsulate the complex interplay between hydrogen diffusion, chemisorption reactions, phase transformations within the metal hydride, heat generation and transfer, and the evolution of mechanical stress within the vessel. By accurately simulating these interconnected processes, the framework seeks to facilitate precise performance predictions, while also identifying design parameters and operational regimes that optimize both efficiency and vessel longevity.

Central to the team’s vision is the translation of fundamental physical and chemical principles of hydrogen interaction into computationally efficient predictive models. Dr. Shaat emphasizes the importance of capturing the coupling mechanisms—particularly how hydrogen transport influences thermal and mechanical responses and vice versa—in order to provide reliable simulations that replace and reduce the dependence on lengthy physical testing. This innovative approach promises to significantly reduce costs, accelerate development cycles, and improve hydrogen storage system designs.

On the experimental front, SwRI’s team, under Dr. Fu’s guidance, is undertaking controlled, long-term cycling experiments to collect essential validation data. By subjecting metal hydride vessels to realistic charge-discharge scenarios, they aim to quantitatively monitor performance degradation mechanisms as they unfold. These experiments will provide the empirical backbone necessary to calibrate and refine the modeling framework, ensuring that simulation outputs faithfully represent real-world behavior under operational conditions.

The synergy between modeling efforts and experiments forms the cornerstone of the project’s iterative design loop. As Dr. Fu explains, modeling insights will direct the focus of experimental tests to critical phenomena, while the resulting data will feedback to correct and enhance the models. This dynamic interplay is expected to unravel previously obscure degradation mechanisms by illuminating the coupled chemical, thermal, and mechanical interactions at work within the metal hydride vessels. Ultimately, this feedback-driven methodology will yield a predictive tool capable of guiding material improvements and engineering decisions.

Hydrogen’s emergence as a clean, versatile energy carrier is a key cornerstone of global efforts to decarbonize energy and transportation sectors. However, to fulfill its potential, advances in storage technology are imperative. Unlike batteries or fossil fuels, hydrogen storage demands novel materials and engineering solutions due to its unique physical and chemical characteristics. By focusing on metal hydride vessels, this project addresses a promising yet underdeveloped storage pathway that could unlock safer, more efficient hydrogen deployment across diverse applications.

The S²TAR funding mechanism plays an instrumental role in enabling this interdisciplinary collaboration by providing seed resources that draw on the expertise strengths of both SwRI and St. Mary’s University. This partnership is not only poised to produce high-impact scientific outcomes but also lays the foundation for sustained cooperation. Through this collaboration, the team is well-positioned to attract larger external funding sources from federal agencies and industry players focused on advancing hydrogen infrastructure and clean energy technology.

As work progresses, the team anticipates that their predictive modeling tool will be an invaluable resource for the hydrogen storage community. It will offer designers and researchers an evidence-based platform to evaluate new materials, optimize vessel designs, and predict operational lifetimes with higher confidence. By systematically addressing the core durability challenges of metal hydride storage, this project represents a wide-reaching advancement that could significantly accelerate solid-state hydrogen storage adoption worldwide.

In conclusion, this joint initiative by Southwest Research Institute and St. Mary’s University epitomizes the cutting edge of hydrogen storage research. Through combining rigor in experimental validation and sophistication in physics-based modeling, the project champions a future where hydrogen is stored more safely and efficiently. The knowledge generated here promises to not only rewrite the fundamentals of storage vessel degradation understanding but also to catalyze practical applications in the global hydrogen economy.

---

Subject of Research: Physics-based modeling and experimental validation of TiFe-based metal hydride hydrogen storage vessel degradation and durability.

Article Title: Advancing Solid-State Hydrogen Storage: Predictive Modeling and Experimental Synergy for Metal Hydride Vessel Durability

News Publication Date: June 24, 2026

Web References: https://www.swri.org/markets/energy-environment/power-generation-utilities/advanced-power-systems/hydrogen-energy-research?&utm_medium=referral&utm_source=eurekalert!&utm_campaign=s2tar-metal-hydrides-pr

Image Credits: Southwest Research Institute

Keywords

Hydrogen storage, metal hydrides, TiFe alloys, chemisorption, physics-based modeling, thermo-mechanical framework, hydrogen diffusion, phase transformation, performance degradation, solid-state hydrogen storage, experimental validation, multiphysics simulations

In a groundbreaking advancement poised to revolutionize semiconductor analysis, a team of researchers from the Institute of Science Tokyo has pioneered a tandem neural network (TNN) architecture that dramatically accelerates the inference of vital material parameters from transistor measurements. This novel machine learning framework addresses one of the most vexing challenges in semiconductor physics—the multivalued inverse problem—achieving unprecedented speed and accuracy that eclipses traditional computational methods.

Semiconductor devices underpin nearly every aspect of modern electronics, from smartphones to sophisticated computing systems. The performance and reliability of these devices hinge on intricate material properties such as defect densities, trap states, and charge carrier mobilities. While engineers can rapidly measure transistor behavior through current-voltage characteristics, deciphering the precise material attributes driving this behavior remains an elusive inverse problem. Conventional approaches rely heavily on iterative simulations and optimization techniques, often consuming hours or even days to converge on plausible solutions.

At the heart of the difficulty lies the phenomenon of multivaluedness: distinct combinations of semiconductor material parameters can yield virtually indistinguishable transistor responses. This ambiguity complicates the task of tracing device performance back to its physical origins. The research team recognized that conventional simulation frameworks are ill-equipped to navigate this degeneracy efficiently. Their insight was to harness the power of tandem neural networks, an innovative deep learning configuration capable of simultaenously enforcing physical consistency and mathematical plausibility in parameter estimation.

The tandem neural network developed by the team couples two interconnected models operating in series. The initial inverse model proposes candidate sets of material parameters based on observed transistor behavior. This output then feeds into a forward model pre-trained to reconstruct transistor characteristics from material parameters. By integrating the forward model’s reconstructions into the training loss function of the inverse network, the system effectively teaches itself to generate solutions that not only fit the measurement data but also align with the physical laws governing transistor operation.

Training this sophisticated architecture required an extensive dataset of over 1,000 simulated transistor measurements, specifically focusing on amorphous indium-gallium-zinc oxide (a-IGZO) transistors—a material of significant interest for flexible and transparent electronics. The dataset spanned a parameter space roughly 1,000 times broader than previous studies, encompassing six key variables: defect density, electron trap states, and carrier mobility among them. Astonishingly, the TNN inferred all six parameters from a single transistor current-voltage curve in less than one millisecond, achieving near-perfect accuracy across the range.

Beyond simulations, real-world validation was paramount. The researchers fabricated a series of a-IGZO transistors under five distinct processing conditions in their laboratories. When subjected to the TNN analysis, the model reproduced the measured device characteristics with exceptional fidelity, requiring no parameter tuning or iterative refinement. This immediate, high-precision performance marks a quantum leap compared to traditional simulation-dependent approaches that demand repetitive calculations, often spanning hours or days.

Assistant Professor Keisuke Ide, one of the principal investigators, highlighted the transformative implications of this speedup: “Our tandem neural network achieves a computational acceleration exceeding six orders of magnitude relative to conventional device simulation methods. This means what once took days can now be done almost instantaneously, radically changing the landscape of semiconductor diagnostics and research.”

The implications of this advancement extend well beyond academic curiosity. In industrial manufacturing, instantaneous and ultra-accurate assessment of transistor material properties can serve as an integral part of quality control, allowing defects and performance variants to be detected and addressed on the production line in real-time. This rapid feedback loop could markedly reduce waste and improve yield in semiconductor fabrication facilities.

Moreover, the TNN framework paves the way for autonomous laboratories where artificial intelligence designs, executes, and interprets experiments with minimal human oversight. By swiftly and reliably unraveling complex material-property relationships, AI-driven research tools can accelerate innovation cycles, optimizing device designs and material formulations in ways previously unattainable due to computational bottlenecks.

Interestingly, the tandem neural network architecture is not conceptually limited to semiconductors alone. The researchers underscore its potential applicability across a broad spectrum of inverse problems characterized by multivaluedness, spanning materials science, chemistry, optics, and possibly even biological systems where multiple underlying parameter sets may produce indistinguishable observational data.

The Institute of Science Tokyo, newly established in 2024 through the merger of Tokyo Medical and Dental University and Tokyo Institute of Technology, serves as a nexus for such cutting-edge interdisciplinary research. This project exemplifies the institute’s mission to advance scientific understanding and technology in ways that create tangible value for society.

As the demands for faster, smaller, and more efficient semiconductor components intensify, novel computational tools that can disentangle the complex interplay of material parameters swiftly and accurately become indispensable. The tandem neural network approach represents a critical milestone on this journey, ushering in a new era where machine learning not only supplements but fundamentally transforms physical sciences research.

This rise of AI-powered physics inference is poised to fuel accelerated material discovery, optimized manufacturing, and ultimately, enhanced electronic devices that power tomorrow’s technological landscape. With this tandem neural network, researchers have cracked open the door to solving multivalued inverse problems at blistering speeds, a feat that was once considered beyond reach.

---

Subject of Research: Semiconductor material characterization using machine learning

Article Title: Tandem Neural Network Rapidly Solves Multivalued Inverse Problems: Application to Oxide-Semiconductor Characterization

News Publication Date: 27-May-2026

Web References: https://doi.org/10.1002/aisy.70437

Image Credits: Institute of Science Tokyo

Keywords

Artificial intelligence, machine learning, semiconductor analysis, inverse problems, multivaluedness, twin neural networks, material characterization, oxide semiconductors, transistor diagnostics, amorphous indium-gallium-zinc oxide, computational speedup, autonomous laboratories

In the ever-evolving field of genomics, a significant leap has been achieved with the introduction of the PGS Browser, a novel public platform designed to transform the way researchers, clinicians, and individuals interact with polygenic scores. Polygenic scores (PGS) are quantitative measures that capture the cumulative effect of numerous genetic variants across the genome, predicting an individual’s predisposition to complex traits and diseases. This new platform promises not just accessibility but also interpretative power, enabling a personalized approach to genomic data that could reshape personalized medicine and risk assessment paradigms worldwide.

At the heart of this transformative tool lies a sophisticated integration of vast genomic datasets, advanced computational algorithms, and user-friendly interfaces. The PGS Browser aggregates polygenic scores derived from an array of large-scale genome-wide association studies (GWAS), providing an unprecedented resource where users can explore the genetic architecture of diverse traits. One of the platform’s most compelling features is its capacity to personalize polygenic risk predictions by incorporating individual genetic data with contextual healthcare analytics, bridging the longstanding gap between raw genetic information and actionable clinical insights.

Traditionally, polygenic scores have been calculated using fixed sets of genetic variants, often constrained by limited population samples or specific ancestral backgrounds. This poses significant challenges regarding the portability and generalizability of these scores across diverse populations. The PGS Browser addresses this limitation head-on by incorporating multi-ancestry datasets and advanced statistical models that recalibrate polygenic scores for diverse genetic ancestries. This opens new horizons in achieving equity in genomic medicine, ensuring individuals from underrepresented populations have access to precise, contextually relevant risk evaluations.

The technical backbone of the PGS Browser involves complex methodologies that include linkage disequilibrium (LD) adjustment, penalized regression frameworks, and machine learning techniques, all optimized for high-throughput and scalable analysis. These computational advances optimize the balance between predictive accuracy and overfitting, a persistent challenge in polygenic score modeling. By dynamically adjusting for confounding genomic features and population stratification, the platform produces risk scores that are both statistically robust and biologically interpretable.

One of the critical innovations within the platform is its interpretability module, designed to unpack the black-box nature of polygenic predictions. Users are not only presented with their personalized polygenic scores but also receive detailed visualizations and annotations explaining the contribution of key genetic loci, biological pathways involved, and potential gene-environment interactions. This level of detail empowers clinicians to make informed decisions and patients to understand the genetic basis of their risks, paving the way for targeted preventative strategies and lifestyle modifications.

From a healthcare perspective, the implications of the PGS Browser extend far beyond academic research. The platform is a strategic tool for genetic counseling, allowing healthcare providers to identify individuals at elevated risk for complex diseases such as cardiovascular disease, diabetes, and certain cancers before clinical symptoms manifest. Early identification through polygenic risk enables preemptive interventions, personalized monitoring, and tailored therapeutic approaches, underscoring a shift from reactive to proactive medicine.

Moreover, the PGS Browser operates with a commitment to transparency and reproducibility, addressing critical concerns in genetic epidemiology. All computational workflows are openly accessible, enabling peer verification and fostering collaborative improvement. The platform also integrates with existing biobank data through standardized protocols, facilitating validation studies and meta-analyses that enhance the reliability of polygenic scores across global cohorts.

A particularly noteworthy aspect of the platform is its potential role in pharmacogenomics. By correlating polygenic scores with drug response phenotypes, the browser could assist in identifying individuals who might benefit from specific medications or require dosage adjustments. This integration with personalized medicine elevates the clinical utility of polygenic scores from risk stratification to therapeutic optimization, highlighting the multi-dimensional impact of genomic data in patient care.

The PGS Browser also bridges the gap between research and public engagement. Designed with intuitive navigation and educational resources, it invites an increasingly genomically literate public to explore their genetic predispositions within a responsible and secure framework. This democratization of genetic information fosters greater awareness and empowerment while maintaining stringent privacy standards, an essential consideration in the era of big data.

In operational terms, the platform boasts scalability features capable of handling millions of individuals’ data concurrently, a necessity given the rapid expansion of genomic databases worldwide. The utilization of cloud-based infrastructures, coupled with containerized computational environments, ensures flexible deployment across various institutional settings while preserving data security and compliance with regulatory frameworks such as GDPR.

From a scientific perspective, the availability of such a public platform catalyzes the discovery of novel gene-trait associations by allowing researchers to perform systematic, large-scale analyses that correlate polygenic scores with phenotypic data. This accelerates the identification of biological mechanisms underlying complex traits, informing subsequent functional studies and therapeutic development pipelines.

Ethical considerations permeate the development and deployment of the PGS Browser. The platform incorporates informed consent modules, guidelines on interpretation limitations, and warnings against deterministic views of genetic risk. By fostering responsible use and interpretation, the platform promotes ethical standards within a field where genetic data can easily be misunderstood or misapplied.

Furthermore, the PGS Browser represents a leap forward in standardizing polygenic score methodologies. Historically, disparate scoring algorithms and variable reporting conventions have hindered comparability across studies. This unified platform offers a harmonized framework that promotes methodological consistency, enabling the scientific community to consolidate findings and build upon a shared foundation.

The cross-disciplinary collaboration manifested in the creation of the PGS Browser highlights the convergence of genetics, bioinformatics, data science, and clinical medicine. This integrative approach is crucial to addressing the complexities of polygenic risk and translating genomic insights into tangible health benefits, marking a pivotal moment in the genomics era.

Looking forward, the PGS Browser sets the stage for continuous evolution. Planned enhancements include integration with longitudinal health records, expansion of trait coverage particularly in psychiatric and neurological domains, and real-time updating as new GWAS data emerge. Such adaptability ensures the platform remains at the forefront of personalized genomics research and application.

In conclusion, the PGS Browser embodies a landmark achievement in the democratization and personalization of polygenic score analysis. By merging rigorous computational techniques with accessible design and ethical mindfulness, it empowers a broad spectrum of users to harness the predictive power of genomic data. This innovation holds transformative potential for healthcare, research, and public understanding, catalyzing a new era in precision medicine where genetic insights are personalized, interpretable, and actionable like never before.

---

Subject of Research: Development and deployment of a public platform for personalized polygenic score analysis and interpretation.

Article Title: PGS Browser: A Public Platform for Personalized Polygenic Score Analysis and Interpretation.

Article References:
Kolosov, N., Reeve, M.P., Briotta Parolo, P.D. et al. PGS Browser: a public platform for personalized polygenic score analysis and interpretation. Nat Commun (2026). https://doi.org/10.1038/s41467-026-74461-7

Image Credits: AI Generated

In the ever-evolving landscape of neonatal medicine, a groundbreaking study has reaffirmed the enduring predictive power of a critical early indicator: the rate at which total bilirubin levels rise postnatally. This development promises to refine the way clinicians anticipate and manage the risks associated with neonatal jaundice, a common yet potentially severe condition affecting newborns worldwide. For decades, the focus has been on absolute bilirubin levels at specific times, but the latest research shifts attention to the kinetics of bilirubin elevation, offering a more dynamic and perhaps more accurate prognostic tool.

Neonatal jaundice arises from elevated bilirubin in the bloodstream, a byproduct of the normal breakdown of red blood cells. While most cases resolve spontaneously without complication, a subset of infants develop dangerously high bilirubin levels, leading to a risk of bilirubin-induced neurological damage. Early identification of these infants is vital to intervene promptly, thus sparking intense research into more reliable predictive markers. The recent study not only confirms the relevance of the bilirubin rate of rise (ROR) but also presents robust data supporting its consistent predictive ability regardless of demographic or clinical variations.

The authors utilized comprehensive longitudinal data spanning several neonatal cohorts to characterize the nuanced relationship between bilirubin kinetics and subsequent clinical outcomes. The analysis employed advanced statistical modeling to parse out temporal trends from confounding variables, yielding more precise thresholds and timelines. These findings validate the utility of monitoring bilirubin’s ROR within the first hours to days after birth as an essential tool for neonatal surveillance and risk stratification, potentially more informative than single bilirubin measurements taken in isolation.

What distinguishes this work is its methodological rigor and expansive dataset that incorporates contemporary postnatal care realities, including early discharge practices and the prevalence of phototherapy. The investigators found that infants demonstrating a rapid bilirubin increase within the initial 24 to 48 hours required closer monitoring as they displayed a heightened probability of developing severe hyperbilirubinemia. Importantly, these infants could be identified before reaching conventionally alarming absolute bilirubin thresholds, which often occur later.

This focus on the dynamic parameter of bilirubin ROR complements traditional nomograms and guidelines currently used to guide clinical interventions such as phototherapy and exchange transfusions. By integrating rate of change data, clinicians are empowered with a forward-looking perspective reflective of the physiological trajectory rather than retrospective static points. This reframing facilitates earlier and more nuanced clinical decision-making, possibly translating to better neonatal outcomes and less invasive treatments.

Moreover, the findings address longstanding uncertainties about the variation in bilirubin metabolism among diverse populations. The authors report that the ROR metric remains consistently predictive across a spectrum of ethnic backgrounds, gestational ages, and birth weights. This universality enhances its appeal as a standard clinical parameter, offering equitable risk assessment where traditional bilirubin charts might underperform due to population biases.

The study further elaborates on the biochemical mechanisms underpinning bilirubin’s postnatal surge, connecting these processes to hepatic enzyme maturation and erythrocyte turnover rates. By contextualizing the observed clinical patterns with molecular insights, the research bridges the gap between laboratory science and bedside care. This integrative approach paves the way for future explorations targeting molecular interventions that could modulate bilirubin dynamics at an even earlier stage.

Critically, the reassessment of bilirubin kinetics arrives at a moment when neonatal care protocols face increasing pressures from early hospital discharges and outpatient jaundice surveillance challenges. The identification of reliable predictors that can be captured with non-invasive serum tests or even transcutaneous monitoring devices holds promise for streamlining care pathways. The emphasis on ROR could reduce reliance on repeated blood draws, alleviating infant distress and healthcare burdens alike.

Another important implication of this research lies in its potential to influence healthcare policy and parental counseling practices. Clear evidence supporting the significance of the bilirubin rate of rise equips healthcare providers with stronger tools to communicate risks effectively. This clarity is crucial to assuage parental anxieties and to promote adherence to follow-up schedules essential for timely interventions while avoiding unnecessary overtreatment.

The adoption of bilirubin ROR as a standard metric will, however, require integration into established workflows and electronic health record algorithms to automate risk alerting. The study’s supplementary data describes pilot implementations where this integration enhanced workflow efficiency and prompted earlier therapeutic action, underscoring the practical applicability of these findings beyond academic settings.

In conclusion, this pivotal research revitalizes the concept that not just the magnitude but the velocity of bilirubin increase after birth carries substantial prognostic weight. It suggests a paradigm shift towards kinetic monitoring as a cornerstone of neonatal jaundice management, urging continuous vigilance and adaptation of clinical protocols to include rate-focused assessments. Such a paradigm promises to safeguard vulnerable neonates worldwide from bilirubin-associated harm more effectively than ever before.

Future research is anticipated to delve deeper into the interplay of genetic factors affecting bilirubin metabolism with the observed kinetic patterns. Coupled with advancements in point-of-care devices, these insights might eventually enable personalized jaundice management, tailored precisely to each newborn’s metabolic profile and risk trajectory.

As the medical community embraces this enhanced predictive capacity, the ultimate beneficiaries will be the families and infants navigating the fragile early days of life. By preempting severe jaundice through refined monitoring of the bilirubin rate of rise, clinicians can deliver more timely care, minimize complications, and ensure that neonatal jaundice remains a manageable, rather than perilous, condition in modern pediatrics.

---

Subject of Research: Predictive metrics for neonatal jaundice severity focusing on the postnatal total bilirubin rate of rise.

Article Title: Predictive ability of postnatal total bilirubin rate of rise endures.

Article References:
Bhutani, V.K. Predictive ability of postnatal total bilirubin rate of rise endures. Pediatr Res (2026). https://doi.org/10.1038/s41390-026-05246-3

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s41390-026-05246-3

In a groundbreaking advancement at the intersection of oncology and artificial intelligence, researchers have developed a novel preoperative AI model designed to estimate cancer-specific mortality in patients diagnosed with nonmetastatic kidney cancer. This pioneering tool represents a remarkable leap forward in personalized medicine, potentially transforming how clinicians evaluate risks and tailor treatment strategies prior to surgery. The study, led by Larcher, Traverso, Scuri, and collaborators, promises to refine prognosis estimations with unprecedented precision, offering a beacon of hope for patients and healthcare providers alike.

Nonmetastatic kidney cancer presents unique challenges in clinical management, primarily due to its heterogeneous nature and variable outcomes. Traditional prognostic models have often fallen short in capturing the complexity and the subtle nuances influencing cancer-specific mortality. The newly developed AI system leverages vast datasets and sophisticated algorithms to decode patterns that human analysis alone cannot easily discern. By integrating diverse clinical variables and imaging data, the model enhances the predictive accuracy of survival outcomes, facilitating more informed decisions regarding the urgency and extent of surgical intervention.

One of the pivotal innovations underpinning this AI model is its capacity to synthesize multiple dimensions of patient data. Unlike conventional prognostic calculators, which might emphasize isolated factors such as tumor size or grade, this model incorporates a broad spectrum of preoperative indicators. These include patient demographic profiles, detailed tumor characteristics derived from radiologic imaging, and laboratory markers. The model’s architecture employs deep learning techniques to weigh these factors dynamically, adapting to subtle interdependencies and delivering a nuanced mortality risk estimate that is tightly aligned with real-world clinical outcomes.

The implications of this model extend beyond mere risk stratification. By providing a more precise assessment of cancer-specific mortality likelihood before surgical intervention, clinicians can better balance the benefits of aggressive treatment against potential complications and quality-of-life considerations. This is especially vital in cases where surgical morbidity may be significant, or where alternative treatments and surveillance might offer comparable prognostic advantages. Enhanced preoperative prognosis not only informs surgical planning but also supports better patient counseling, setting realistic expectations grounded in individualized risk profiles.

Technically, the model was trained and validated using extensive multicentric datasets, drawn from diverse patient populations to ensure widespread applicability and robustness. The algorithm underwent rigorous cross-validation procedures, testing its predictions against known survival outcomes to eliminate biases and confirm reliability. Importantly, the model’s interpretability was also prioritized, enabling clinicians to understand which variables contributed most significantly to mortality risk predictions, thereby fostering trust and facilitating integration into clinical workflows.

The methodology incorporated advanced imaging analytics alongside conventional clinical parameters. Radiomic features extracted from preoperative CT scans were a major component, providing detailed textural and morphological insights that correlate with tumor behavior. These imaging biomarkers, when coupled with biochemical data such as serum creatinine and hemoglobin levels, painted a comprehensive portrait of both tumor aggressiveness and patient physiological status. This multimodal data fusion became a cornerstone of the AI’s ability to deliver personalized prognostic assessments.

In terms of deployment, the AI model is designed to be user-friendly and seamlessly integratable into existing hospital information systems. Its interface allows surgeons, oncologists, and multidisciplinary teams to input clinical data and receive risk assessments in real-time. Such accessibility enables rapid decision-making, potentially influencing decisions about the timing of surgery, the scope of resection, and adjuvant therapy planning. Moreover, the model’s adaptability means it can incorporate new data as clinical understanding and treatment modalities evolve.

A critical aspect of this research lies in the ethical deployment of AI in clinical settings. The authors address concerns regarding algorithmic transparency, patient privacy, and the risk of over-reliance on machine predictions. They emphasize that the AI model is meant to augment, not replace, clinical judgment, serving as an advanced decision-support tool. Continuous monitoring, feedback mechanisms, and physician input are integral components of the model’s operational framework, ensuring that clinical expertise remains central.

The timing of this innovation is particularly significant given the increasing incidence of kidney cancer worldwide and the ongoing quest for precision oncology solutions. Current staging systems and prognostic indices, while valuable, often underestimate individual variability and outcomes seen in everyday clinical practice. By harnessing the predictive power of AI, this model addresses a critical unmet need, offering a reliable compass to navigate the complex risk landscape of nonmetastatic kidney cancer.

Looking forward, the researchers envision expanding the scope of their AI system. Integration with genomic and molecular profiling data could further enhance prognostic accuracy, capturing the genetic underpinnings and heterogeneity of kidney tumors at a granularity impossible with imaging and clinical data alone. Such multi-omic AI-driven platforms could revolutionize not only prognosis but also treatment selection, ushering in an era of truly personalized cancer care.

Operability in diverse clinical settings also forms a core consideration in the model’s development. The use of widely available imaging modalities and routinely collected clinical data ensures that the technology can be disseminated globally, including in resource-constrained environments. This raises the tantalizing possibility of democratizing high-quality kidney cancer risk assessments, bridging gaps in healthcare disparities that currently limit access to advanced oncologic prognostication.

Another dimension includes the potential of the AI model to guide clinical trials and research. By precisely stratifying patients based on their predicted cancer-specific mortality risk, clinical investigators can design trials with more homogenous cohorts, enhancing statistical power and facilitating the discovery of targeted therapies. Moreover, the model could identify patients who might benefit most from novel interventions, optimizing resource allocation and accelerating therapeutic developments.

In essence, the AI-driven preoperative model is a testament to the transformative power of machine learning in medicine. It encapsulates the promise of personalized, data-driven healthcare that adapts to individual patient profiles rather than relying solely on broad population averages. As this technology moves from development to clinical adoption, it heralds a future where kidney cancer prognosis is not a matter of chance but an informed, precise science guiding bespoke treatment pathways.

The broader oncology community has greeted this development with enthusiasm, recognizing its potential to reshape risk assessment paradigms. The study stimulates a conversation about the role of AI in preoperative oncology, highlighting both opportunities and challenges inherent in integrating advanced computational tools into clinical decision-making. It serves as a clarion call to clinicians, researchers, and technology developers to collaborate and refine these early innovations for maximal patient benefit.

In conclusion, the work by Larcher, Traverso, Scuri, and their colleagues marks a decisive step toward integrating artificial intelligence into the clinical management of nonmetastatic kidney cancer. Their AI model exemplifies how cutting-edge technology, grounded in robust data and sophisticated analytics, can empower clinicians with actionable insights that elevate patient care. As machine learning continues to evolve, such tools are poised to become indispensable allies in the global fight against cancer.

Subject of Research:
Artificial Intelligence application for prognostic prediction of cancer-specific mortality in nonmetastatic kidney cancer patients.

Article Title:
A preoperative Artificial Intelligence model to estimate cancer-specific mortality in nonmetastatic kidney cancer patients.

Article References:
Larcher, A., Traverso, A., Scuri, P. et al. A preoperative Artificial Intelligence model to estimate cancer-specific mortality in nonmetastatic kidney cancer patients. Nat Commun (2026). https://doi.org/10.1038/s41467-026-74419-9

Image Credits:
AI Generated

A groundbreaking study from UCLA Health introduces a pioneering approach to understanding anorexia nervosa, a complex psychiatric disorder notorious for its high relapse rates and elevated mortality. Researchers employed an ingestible, vibrating capsule to probe the nuanced ways in which the nervous system of individuals with anorexia nervosa processes bodily sensations originating from the gastrointestinal tract. This innovative technology not only illuminates the enigmatic interplay between gut sensations and brain function but also offers promising predictive capabilities regarding relapse risk, potentially transforming clinical management of the disorder.

Anorexia nervosa is marked by a persistent restriction of energy intake, an intense fear of weight gain, and a distorted body image, resulting in dangerously low body weight. Despite state-of-the-art treatments aimed at restoring healthy body weight, relapse rates remain alarmingly high—up to 50% within a year post-treatment. Suicidal behavior is also a significant concern, often cited as the leading cause of death among afflicted individuals. This desperate clinical landscape underscores the urgent need for objective biomarkers and more profound mechanistic insights to improve therapeutic outcomes.

The inadequacy of current understanding centers around a crucial chasm: while physicians focus on weight normalization, many patients continue to suffer from entrenched symptoms that drive relapse, suggesting that the disorder’s etiology extends beyond mere weight regulation. The nervous system’s interpretation and integration of visceral signals, especially from the gut, could hold the key to these persistent challenges. Yet, before this study, precise tools to quantify and manipulate these signals in vivo remained elusive.

The UCLA research team harnessed a clinical trial design involving an ingestible vibrating capsule capable of delivering controlled, gentle mechanical stimuli to the stomach. This non-invasive method allowed for real-time interrogation of gastrointestinal interoception—the brain’s perception and processing of internal bodily sensations. By remotely activating the capsule, the researchers could produce varying intensities of stomach vibrations, simulating subtle gut sensations that are typically difficult to access or measure reliably.

In the study, 62 female patients who had been hospitalized for anorexia nervosa and whose body weight had been restored to clinical norms, were compared against 57 healthy controls. Participants swallowed the vibrating capsule and were instructed to indicate, via button press, whenever they perceived a vibration. Concurrent physiological monitoring tracked brain activity through neuroimaging techniques, cardiac rhythms, and stomach motility, providing a multidimensional understanding of the gut-brain axis in real time. Self-reported metrics on hunger and bodily awareness further enriched the dataset.

Employing sophisticated computational modeling, the researchers analyzed participants’ ability to detect and interpret the gastric signals. These models were designed to estimate participants’ expectation of gut sensations, their reliance on incoming sensory signals, and their adaptability in updating these expectations when signals fluctuated. This approach transcended simple sensory detection, probing fundamental cognitive processes involved in interoceptive learning and prediction error correction.

The results revealed a striking divergence between anorexia nervosa patients and healthy controls. Those with the disorder exhibited diminished accuracy in sensing subtle stomach vibrations, often unaware that a stimulus was occurring even when the capsule was active. They demonstrated a cognitive bias towards expecting an absence of gut sensations and were significantly slower to revise these expectations in the presence of confirmed visceral signals. This implies an altered central processing mechanism rather than a peripheral sensory deficit, reflecting a disconnection or miscommunication within the gut-brain axis.

Crucially, these interoceptive abnormalities persisted despite the restoration of healthy body weight, challenging the conventional clinical assumption that weight normalization is tantamount to recovery. The findings suggest that long-lasting neurobiological alterations in interoceptive processing may underlie the stubborn persistence of anorexic behaviors and cognitions, thus offering a plausible explanation for the high relapse propensity witnessed in clinical practice.

Linking these mechanistic insights to clinical outcomes, the study reported a robust association between impaired gastric signal detection and relapse risk during a six-month post-discharge follow-up period. Patients whose nervous systems exhibited the greatest bias towards ignoring bodily signals were significantly more likely to experience a relapse. This highlights the potential clinical utility of the vibrating capsule as a predictive tool, enabling healthcare providers to stratify relapse risk and tailor interventions accordingly.

From a translational perspective, the use of this ingestible technology marks a seminal advance in biomarker discovery for anorexia nervosa. It opens new avenues for personalized medicine where objective physiological metrics, rather than solely subjective symptom reports, guide therapeutic decisions. Moreover, it positions gastrointestinal interoception as a critical target for future treatment strategies, including neuromodulation, pharmacotherapy, and cognitive retraining designed to recalibrate gut-brain communication pathways.

Nevertheless, the authors caution that these findings currently pertain to a relatively homogeneous sample, primarily young female patients, and must be validated across larger and more diverse populations to confirm generalizability. Future research should also explore how different treatment modalities influence interoceptive processing and whether modulation of this axis can directly mitigate relapse risks.

In conclusion, this UCLA-led study redefines our understanding of anorexia nervosa by exposing the subtle but profound alterations in gastrointestinal interoception that endure beyond weight restoration and presage relapse. It suggests that true recovery necessitates healing not only the body but the fundamental neural processes that govern the perception and integration of internal bodily states. This insight sets the stage for an era where interoceptive biomarkers and bioelectronic medicine intersect to revolutionize the management of this devastating disorder.

Subject of Research: Gastrointestinal interoception and relapse prediction in anorexia nervosa
Article Title: Altered Gastrointestinal Interoception in Anorexia Nervosa Predicts Relapse
News Publication Date: 17-Jun-2026
Web References: http://dx.doi.org/10.1001/jamapsychiatry.2026.1301
References: Published in JAMA Psychiatry, 2026
Keywords: Anorexia nervosa, gastrointestinal interoception, relapse prediction, gut-brain axis, ingestible vibrating capsule, psychiatric biomarkers, neurobiology of eating disorders

Often hailed as “the window to the soul,” the human eye has long fascinated scientists and philosophers alike. However, beyond its poetic allure, emerging research suggests that the eyes provide a vital window into the health of the brain. A groundbreaking new study spearheaded by the University of Florida’s biomedical engineering department has unveiled a transformative approach to predicting Alzheimer’s disease risk factors by analyzing retinal photographs with cutting-edge deep learning algorithms. This development promises to revolutionize early detection and intervention for one of the most devastating neurodegenerative diseases.

Traditionally, diagnosing Alzheimer’s disease has been a challenge due to its insidious onset and prolonged development over decades. Most diagnostic protocols focus on detecting the disease in its later stages, by which time significant, irreversible brain damage has already occurred. Recognizing this limitation, Dr. Ruogu Fang and her interdisciplinary team leveraged advances in artificial intelligence and biomedicine to explore the retina—a part of the central nervous system accessible via non-invasive imaging—as a potential biosensor for Alzheimer’s risk.

The study capitalized on a rich database of over 40,000 retinal images collected from the UK Biobank, a comprehensive health resource with longitudinal patient data. Retinal photography is commonly conducted during routine eye exams, diabetes management, glaucoma monitoring, and cataract assessments. By harnessing these widely available, inexpensive images, the research team circumvented the need for costly, invasive, or logistically complex diagnostic tools such as MRI or PET scans, thus holding promise for scalable, population-wide screening.

Employing deep learning, a subset of machine learning characterized by neural networks modeled after the human brain, the researchers developed an AI framework that scrutinized retinal morphology. This model was able to discern subtle variations in retinal arteries, veins, and the optic nerve head, features imperceptible to the human eye but correlated strongly with known risk factors for Alzheimer’s disease. Such morphological markers serve as proxies for neurovascular health, which is crucial given Alzheimer’s recognized association with vascular impairment.

One of the most striking revelations of this research was the model’s capacity not only to predict intrinsic biological traits such as sex and blood pressure but also modifiable lifestyle factors that influence Alzheimer’s risk. The AI effectively inferred behaviors like smoking, alcohol consumption, and even insomnia from retinal images alone. This objective measurement circumvents the notorious unreliability of self-reported lifestyle data, often plagued by bias and inaccuracies, thereby enriching risk assessment precision.

As Dr. Fang points out, retinal morphology functions less like a mere clinical questionnaire and more like an integrative biological sensor reflecting a cumulative lifetime burden of neurovascular and lifestyle insults. This conceptual shift underscores the retina’s role in encapsulating decades of physiological data, transforming retinal imaging into a powerful tool for capturing both present health and future vulnerability.

Furthermore, the implications extend beyond risk prediction. Identifying individuals with early signs of retinal changes could trigger timely clinical interventions long before cognitive symptoms manifest. Such preclinical detection enables the implementation of protective lifestyle adjustments, pharmacological therapies, or cognitive training programs designed to delay or mitigate Alzheimer’s progression.

The pioneering work built on previous findings from Fang’s group that demonstrated retinal images’ utility in detecting active Alzheimer’s disease cases. However, this new model pioneers the field by focusing on prodromal markers and risk factors instead of established dementia, bridging a critical gap in neurodegenerative diagnostics.

Importantly, the use of retinal photographs democratizes neurovascular health monitoring. Since retinal imaging is non-invasive, accessible, and affordable, it offers the prospect of integrating brain health assessment into routine eye care, thus expanding preventive neurology beyond specialized clinical settings.

This research also exemplifies the synergy between computational science and biomedical engineering. By combining sophisticated image analysis algorithms with large-scale patient data, it transcends traditional diagnostic constraints, marrying technology with medicine in a profoundly impactful manner.

While challenges remain, including the need for further validation across diverse populations and integration with other biomarkers, this study sets a new paradigm for early Alzheimer’s disease risk stratification. The possibility of preemptive interventions guided by retinal AI analysis heralds a future where the devastating trajectory of Alzheimer’s may be altered through timely and personalized healthcare.

In summary, the novel application of deep learning to retinal photographs opens an unprecedented window into brain health, heralding a new era of precision neurodiagnostics. It moves us closer to the long-sought goal of identifying Alzheimer’s risk decades in advance, fostering opportunities for early intervention and ultimately, better patient outcomes.

---

Subject of Research: People

Article Title: Prediction of Alzheimer’s disease risk factors from retinal images via deep learning: Development and validation of biologically relevant morphological associations in the UK Biobank

News Publication Date: 16-Jun-2026

Web References:
DOI: 10.1177/13872877261457650

---

Keywords

Alzheimer’s disease, Neurodegenerative diseases, Neurological disorders, Eye, Retina, Artificial intelligence, Image analysis, Deep learning, Biomedical engineering, Neurovascular integrity, Retinal biomarkers, Predictive modeling

In a groundbreaking development poised to reshape the understanding and engineering of cylindrical structural components, researchers have unveiled an innovative test-based methodology for assessing collapse behavior in helically wound and layered cylindrical structures. This novel approach promises to significantly enhance the reliability and safety assessments of cylindrical constructs employed in various high-stakes engineering applications, including aerospace, civil infrastructure, and energy pipelines.

Historically, the analysis of cylindrical structures composed of helically wound and layered materials has presented profound challenges due to their complex mechanical interactions and anisotropic properties. Traditional analytical methods often fall short in capturing the nuanced interplay between layers and winding orientations under collapse-inducing loads. The research led by Zhang, Y., Saneian, M., Bai, Y., and their collaborators introduces what is termed the “test-based equivalent-material method,” which effectively bridges the gap between experimental data and predictive modeling.

At the heart of this methodology lies an innovative equivalence principle that translates the intricate multi-layered helically wound architecture into an experimentally validated simplified material representation. By conducting targeted mechanical tests, the researchers derive effective material properties that encapsulate the aggregate response of the layered system. This surrogate material model then facilitates robust numerical simulations, enabling precise collapse qualification without resorting to prohibitively detailed layer-by-layer analyses.

One of the major advantages of this approach is its scalability and adaptability. Engineers and designers can apply the equivalent material parameters obtained through relatively straightforward tests to a wide range of cylindrical structures, regardless of size or specific material compositions. This dramatically reduces the computational costs and time traditionally associated with finite element models of complex laminates, streamlining the design and certification process for structures susceptible to buckling and other collapse modes.

The structural integrity of helically wound cylinders is crucial in multiple industrial sectors. For instance, high-pressure pipelines transporting oil and gas frequently employ helically wound steel strips to enhance strength and durability. Similarly, aerospace components such as pressure vessels for fuel storage and fuselage sections rely on layered composites to achieve superior performance-to-weight ratios. Recognizing the critical nature of these applications, the test-based equivalent-material method offers a pragmatic pathway to ensure these structures can withstand extreme conditions without catastrophic failure.

Delving deeper, the researchers meticulously examined the influence of winding angle, layer thickness, and material anisotropy on the effective mechanical behavior. Their experimental protocols involved subjecting prototype samples to axial compression, torsion, and combined loadings while rigorously monitoring deformation and failure patterns. By correlating these results with equivalent material properties, they established a robust calibration framework that captures subtle nonlinearities and failure mechanisms intrinsic to layered helically wound cylinders.

A particularly innovative aspect of the work is the incorporation of nonlinear collapse phenomena into the equivalent-material representation. Classical linear elastic or even simple nonlinear models often inadequately predict collapse, especially for layered structures where progressive delamination, local buckling, and material yielding interact in complex ways. The team’s method incorporates advanced constitutive modeling informed by experimental evidence, thereby extending predictive capabilities into the ultimate collapse regime with remarkable accuracy.

Moreover, the proposed methodology aligns excellently with ongoing efforts in digital twin technology and real-time structural health monitoring. By building a validated equivalent material model from test data, engineers can integrate this model into digital twins for continuous prediction of structural performance under operational loads. This fusion of experimental validation and digital simulation heralds a new era in proactive maintenance and failure prevention for critical infrastructure.

Application-wise, the method has already demonstrated its potential in prototype development phases. The team tested cylindrical structures representative of submarine pressure hulls and composite overwrapped pressure vessels, both of which undergo stringent safety certification. The test-based equivalent-material method facilitated rapid iterative design modifications while maintaining high fidelity in predicting collapse loads, thus shortening development cycles and improving design confidence.

From a practical viewpoint, the reduction of experimental complexity stands out. Whereas traditional methods might require exhaustive testing of multiple individual layers, orientations, and bonding conditions, this approach consolidates testing into a manageable set of representative experiments. The derived equivalent material parameters effectively embody the net structural response, enabling engineers to bypass onerous microscale testing without sacrificing accuracy.

Fundamentally, this research contributes to the broader scientific discourse on multiscale modeling in composite materials. The challenge of reconciling microscale structural features with macroscale performance is a longstanding issue. By implementing a test-based equivalency strategy, the researchers provide a tangible methodology that harmonizes the scales through empirical calibration rather than purely theoretical assumptions or computational homogenization.

In terms of future directions, the team envisions extending their framework to incorporate environmental degradation factors such as moisture ingress, temperature fluctuations, and fatigue-induced damage accumulation. These variables critically influence collapse behavior in real-world settings, and integrating them into the equivalent-material paradigm will bolster the method’s practical utility.

Additionally, the adaptability of this method suggests promising integration with emerging additive manufacturing techniques. Layered and helically wound structures produced by 3D printing could be experimentally characterized using this approach to tailor equivalent materials for bespoke designs, unlocking unprecedented customization in lightweight, high-strength structural components.

Ultimately, this test-based equivalent-material method stands as a transformative advance in engineering science, merging experimental rigor with computational efficiency to unravel the complexities of helically wound layered cylindrical structures. The impact spans multiple disciplines, offering a robust, scalable tool for ensuring structural safety in an array of critical applications — from aerospace to energy infrastructure, shaping the future of advanced materials and design methodologies.

This innovation, published in the prestigious journal Communications Engineering in 2026, signifies a major leap forward in predicting collapse phenomena, providing engineers and researchers with a powerful new instrument to tackle complex structural challenges. Its implications for enhancing durability, reliability, and safety herald a significant milestone in structural engineering research.

In summary, by deftly combining empirical testing with advanced material modeling, Zhang and colleagues have unlocked an effective pathway to understand and predict the collapse of helically wound and layered cylinders. Their work paves the way for safer, more efficient designs, enabling engineers to push the boundaries of structural performance in demanding environments — a testament to the fusion of innovation and practical engineering excellence.

---

Subject of Research: Collapse qualification and mechanical behavior of helically wound and layered cylindrical structures using an experimental equivalent-material method.

Article Title: Test-based equivalent-material method for collapse qualification of helically wound and layered cylindrical structures.

Article References:
Zhang, Y., Saneian, M., Bai, Y. et al. Test-based equivalent-material method for collapse qualification of helically wound and layered cylindrical structures. Commun Eng (2026). https://doi.org/10.1038/s44172-026-00699-0

Image Credits: AI Generated