Integrating actigraphy with demographic data enhances cognitive performance prediction: a multimodal UK biobank analysis using machine learning.

Deep learning based eye disease classification using Optical Coherence Tomography (OCT) images.

Node-RADS v1.0 on chest CT for lung cancer: Reproducibility and diagnostic performance.

Dilated cardiomyopathy (DCM) and left ventricular non-compaction (LVNC) are major non-ischemic cardiomyopathy (NICM) subtypes with heterogeneous outcomes. Conventional clinical and echocardiographic markers remain insufficient for long-term risk stratification. This study aimed to identify serum metabolites associated with adverse cardiovascular outcomes and evaluate their incremental prognostic value in NICM. Thirty-two patients with DCM or LVNC and left ventricular ejection fraction < 50% were enrolled and followed for a median of 45.5 months. The primary endpoint was a composite of cardiovascular death, heart failure-related hospitalization, or clinically indicated cardiovascular device implantation. Baseline serum samples underwent liquid chromatography-mass spectrometry-based untargeted metabolomic profiling. Differential metabolites were identified using OPLS-DA and KEGG enrichment analysis. Prognostic metabolites were screened using Cox regression, Kaplan-Meier analysis, correlation filtering, and ROC analysis. An integrated Cox model combining clinical and metabolic markers was evaluated using bootstrapping, calibration, and time-dependent ROC analysis. A total of 299 differential metabolites were identified and enriched in bile secretion, steroid hormone biosynthesis, and neuroactive ligand-receptor interaction pathways. Sphingosine-1-phosphate (S1P) and tetrahydrocortisone (THE) were selected as final prognostic metabolite biomarkers, with ROC AUCs of 0.777 and 0.793, respectively. The integrated model incorporating S1P, THE, tricuspid annular plane systolic excursion, and total protein achieved a bootstrap-corrected C-index of 0.772, with time-dependent AUCs of 0.92 and 0.86 at 3 and 5 years. Serum metabolomics may provide complementary prognostic information in NICM. S1P and THE are exploratory biomarkers linked to remodeling and stress, supporting risk stratification in NICM with reduced ejection fraction.

Identification of preeclampsia-associated immune-related risk loci in the Chinese population.

Tick bites in Australia are associated with a poorly understood syndrome known as Debilitating Symptom Complexes Attributed to Ticks (DSCATT), however, the underlying biological mechanisms remain unclear. Here, we investigate host responses to tick bite and DSCATT by profiling circulating host-encoded microRNAs (miRNAs), key regulators of gene expression. Circulating miRNAs were profiled in two cohorts: a longitudinal cohort followed from tick bite for up to 12 months, and a retrospective cohort with DSCATT. Differential expression analysis revealed that tick bite induces widespread changes in circulating miRNAs, with 149 miRNAs showing significant variation over 12 months. Temporal clustering revealed two expression patterns: one oscillating trajectory and a declining then stabilising, with predicted targets enriched for pathways related to immune modulation, tissue remodelling and cellular stress responses. DSCATT patients exhibited 98 differentially expressed miRNAs, including significant overlap with acute tick bite miRNAs, and four miRNAs correlated with symptom severity, including fatigue and dizziness. Machine learning analysis identified a five-miRNA signature that classified acute tick bite with 86% accuracy and receiver operating characteristic area under the curve (ROC AUC) of 0.92. These findings represent, to the best of our knowledge, the first characterisation of host miRNA responses to tick bite and DSCATT, highlighting potential biomarkers and mechanisms underlying chronic symptom development.

This study characterized the expression profile of miR-153-3p in deciduous pulpitis in children, validated its role in regulating inflammation, oxidative stress, and apoptosis via PTEN targeting, and evaluated its biomarker and therapeutic potential. In a prospective design, pulp tissue was collected from 180 children with deciduous-tooth pulpitis (mild, moderate, or severe) and 180 caries-free controls. qRT-PCR was used to quantify miR-153-3p and PTEN expression, VAS pain scores were recorded, and ROC analysis was performed. An LPS-induced inflammatory model (5µg/mL, 24h) was established in human dental pulp stem cells, which were transfected with miR-153-3p mimic or PTEN overexpression plasmid. Cell viability, apoptosis, inflammatory cytokines, and oxidative stress markers were subsequently assessed, and correlations among miR-153-3p, VAS, and PTEN were evaluated. miR-153-3p was significantly downregulated in inflamed pulp and inversely correlated with disease severity and VAS score (r = - 0.323, P < 0.001), with an ROC AUC of 0.814. LPS suppressed miR-153-3p and elevated PTEN in a concentration- and time-dependent manner. miR-153-3p upregulation restored viability, reduced apoptosis, decreased cytokine release, increased SOD activity, and lowered MDA content. Dual-luciferase assays confirmed PTEN as a direct target, and PTEN overexpression reversed the protective effects of miR-153-3p. These findings indicate that miR-153-3p downregulation is associated with inflammation severity, and that miR-153-3p attenuates LPS-induced cellular responses by inhibiting PTEN. Its utility in assessing tissue inflammation and potential as a therapeutic target warrant further in vitro and in vivo validation.

Impact of myocardial fibrotic changes assessed by magnetic resonance imaging on early postoperative outcomes after aortic valve replacement

ObjectiveTo compare the sensitivity of the Mini-Mental State Examination (MMSE) and the Montreal Cognitive Assessment (MoCA) for detecting cognitive change in Huntington's disease (HD) subjects stratified by baseline performance quartiles and the Huntington's Disease Integrated Staging System (HD-ISS).MethodsParticipants (age 25-65; CAG 40-50) from an observational cohort completed MMSE and MoCA at baseline, with follow-up visits at least six months after baseline, and observations truncated at five years.We analyzed the full cohort (n = 246) and an HD-ISS-staged subset (n = 132). Longitudinal change was evaluated using mixed-effects models, adjusted for baseline age, gender, education, CAG, and age × CAG interaction. Baseline discrimination between HD-ISS stages was assessed with ROC curves and AUC comparisons.ResultsAnnual decline was nearly identical for MMSE and MoCA (-0.334 vs. -0.331). Adjusted quartile analyses revealed differential sensitivity: MoCA showed greater decline in lower and higher performance levels (Quartiles 1, 3, and 4), while MMSE was most sensitive at mid-range (Quartile 2). In HD-ISS analyses, MMSE demonstrated more consistent decline across stages, particularly Stages 1 and 3, indicating sensitivity to early and moderate disease phases. Baseline ROC analyses favored MoCA for stage discrimination (Stage 0 vs. all: AUC 0.789 vs. 0.686, p = 0.011).ConclusionThe MMSE and MoCA exhibit complementary strengths. MoCA demonstrates superior sensitivity for baseline staging and early disease detection, while MMSE better captures longitudinal decline. These findings suggest that combined use of both instruments may provide a more complete assessment of cognitive progression in HD.

Urban EV uptake is raising feeder peaks and energy costs in city networks. We present an AI framework integrated with the open charge point protocol (OCPP), a standard communication protocol for EV charging systems that combines short-horizon demand forecasting with tariff-aware and fairness-aware scheduling and runtime anomaly detection from charger telemetry. Decisions are encoded with OCPP 1.6 and 2.0.1 operations including SetChargingProfile, ClearChargingProfile, GetCompositeSchedule, RemoteStartTransaction and RemoteStopTransaction. The system runs in rolling 15-minute control on commodity hardware. Using a two-year multi-station tariff-aware urban dataset with 1,553 operational days and 455 days with tariff coverage, the scheduler delivers average reductions of 5.0% for GA and 8.2% for a hybrid scheduling approach combining genetic algorithms (GA) and Q-learning (a reinforcement learning method) in both feeder peak and charging cost relative to the baseline. The policy conserves daily energy and applies a fairness mechanism that limits excessive delays for individual users limiting the concentration of deferrals. Under fixed caps and posted tariffs, the pipeline is deterministic, so day-level confidence intervals collapse to a point, and small parameter jitter leaves the mean unchanged within measurement noise. For anomaly detection, a CNN reaches ROC AUC 0.914 and the Autoencoder and Isolation Forest reach 0.735 and 0.636. A latency budget covering ingestion, forecasting, scheduling and OCPP round trip confirms near real-time feasibility with optimization overhead below 0.1s. We provide a minimal reproducibility package with code, data splits, and scripts. The results show that standards-compliant and OCPP-aware AI can deliver measurable grid and cost benefits at the city scale and offer a practical path for sustainable and equitable EV charging in smart city contexts.

We aimed to examine the performance of the expanded cardiovascular (CV) risk prediction score for rheumatoid arthritis (ERS-RA), the Framingham CV risk score and the carotid intima-media thickness (IMT) as predictors of CV mortality in patients with rheumatoid arthritis (RA). We studied 1194 patients with RA who had an ultrasound measurement of the carotid IMT and who had clinical information to calculate the ERS-RA and the Framingham score. Deaths were confirmed by death certificate and attributed to a CV cause if the immediate or underlying cause of death was due to ICDM-9 codes 390-459. We compared areas under receiver operator characteristic curves (ROC AUC) and examined reclassification tables for the performance of each scale to correctly predict CV deaths. Patients were observed for 10,063 person-years, during which 271 patients died (22.6%), for an overall mortality rate of 2.7 deaths per 100 person-years (95% CI 2.4, 3.0). A CV cause of death was identified in 139 patients (11.6%), for a CV mortality rate of 1.4 per 100 person-years (1.2, 1.6). ROC AUCs were 0.72 (95% CI 0.68, 0.76) for the Framingham scale, 0.76 (95% CI 0.72, 0.81) for the carotid IMT and 0.84 (95% CI 0.81, 0.88) for the ERS-RA. Compared to the Framingham scale the ERS-RA correctly reclassified 6% of patients (p=0.02), while the carotid IMT correctly reclassified 5% (p=0.003). The ERS-RA was superior to the Framingham scale in the prediction of CV mortality in patients with RA.

BackgroundAlzheimer's disease (AD) is a neurodegenerative disorder whose incidence grows with age and its development is gradual. However, if detected earlier there is much hope to prevent further exacerbation. In this study, NGS transcriptomics data from cases and controls with Braak scores of III-IV were investigated that all possessed neurofibrillary tangles (NFTs) in their fusiform gyrus.ObjectiveThe aim of this study was to discover the underlying mechanisms at gene level which could explain cognitive impairment by considering the presence of NFTs in both groups.MethodsDifferentially expressed genes (DEGs) were determined and ROC AUC were evaluated by leave-one-out cross-validation method on the diagnostic DEGs to detect candidate gene biomarkers. WGCNA was employed to identify co-expression modules with their trait association. Finally, in silico hybridization of lncRNAs from potential biomarkers with important AD-related microRNAs was carried out.ResultsHighly ranked potential diagnostic gene biomarkers revealed assessed AUC ranges of 80-90% in which RASGRF2-AS1 demonstrated the highest value. WGCNA demonstrated upregulated genes in favor of dephosphorylation of tau, proper proteostasis and vascular health in resilient controls whereas dysfunctional proteostasis, chronic protein misfolding, heightened cellular stress and tetrahydrobiopterin deficiency were attributed to cognitive impairment in AD patients. In silico analyses predicted some lncRNAs with a high possibility of acting as sponge for AD-related microRNAs.ConclusionsThis study discovered potential diagnostic gene biomarkers and transcriptional signatures that could explain the mechanisms of cognitive decline by considering the existence of NFTs, which could provide further insight for diagnosis and treatment of the disease.

Cytokines are central to the immune response, determining whether T-cell activation leads to protective immunity or tolerance. The secretion of key cytokines such as IFN-gamma, IL-2, IL-4, and IL-10 shapes the immune response, influencing its presence, pathways, and magnitude. While most existing computational methods for predicting peptide immunogenicity typically focus on either cytokine induction or histocompatibility complex binding affinity separately, they fail to capture the full complexity of immune activation. In this study, we introduce a two-stage computational framework that integrates both cytokine induction and HLA binding affinity predictions to provide a more comprehensive prediction of peptide immunogenicity. First, we develop machine learning models EpiLAMA-IL to predict cytokine secretion based on peptide and parent protein sequences. These models outperform current state-of-the-art cytokine release prediction methods, achieving ROC AUC from 0.79 up to 0.92 depending on the task. These cytokine scores are then combined with HLA binding affinity predictions and peptide physico-chemical descriptors to develop a composite immunogenicity score prediction model EpiLAMA-IM. This approach bridges the gap between sequence-based antigen presentation and functional immune outcomes, offering improved accuracy in predicting immune responses and achieves MCC 0.47 on the CD4Episcore dataset.

Artificial intelligence (AI) algorithms such as ENLIGHT and DeepPT represent promising approaches to identify predictive biomarkers for immune checkpoint blockade (ICB). Evaluation of ICB response prediction throughout the span of ICB treatment provides a dynamic perspective of a biomarker's predictive value. A post-hoc analysis of two pan-cancer trials of patients receiving ICB was performed. Samples were available from the pre-ICB, on-ICB and post-progression timepoints. ENLIGHT matching score (EMS) was calculated using measured transcriptome from sequencing (EMS-NGS), and imputed transcriptome from H&E images using DeepPT (EMS-DP). The predictive value of EMS-NGS and EMS-DP for clinical benefit rate (CBR), progression free survival (PFS) and overall survival (OS) were assessed alongside PD-L1 IHC and tumor mutational burden (TMB). Trajectory of EMS across the ICB treatment timepoints was assessed. 154 H&E slides with matched RNA sequencing from 111 patients, representing 14 tumor types were analyzed. The ROC AUC for predicting CBR was calculated for each biomarker: EMS-DP (0.78), EMS-NGS (0.73), PD-L1 IHC (0.68) and TMB (0.62). Higher EMS-NGS was associated with improved PFS and OS whereas EMS-DP was associated with improved PFS but not OS. Both EMS-NGS and EMS-DP at the on-ICB timepoint remained higher for patients with ongoing response or acquired ICB resistance compared to primary ICB resistance. At the post-progression timepoint, EMS did not differ based on type of resistance. In this pan-cancer study, EMS-NGS and EMS-DP were promising biomarkers of ICB response. EMS values were concordant with response or resistance throughout the ICB treatment course. Further prospective validation of ENLIGHT is warranted.

From prediction to mechanism: Explainable AI uncovers plasma and CSF proteomic signatures of Alzheimer's disease.

Nanopore technology enables rapid, portable, and label-free single-molecule detection of analytes, including nucleic acids and proteins. One strategy for improving accuracy is to increase interactions between analytes and the nanopore. For example, engineered nanopores with additional constrictions improve analyte-pore interactions during translocation. This concept adopts the exploration of biological nanopores, which naturally contain multiple constrictions. Here, we demonstrate that Epx4, a pore-forming toxin with two independent β-barrels, functions as a nanopore sensor capable of generating informative ionic current signals during polypeptide translocation. Structural analysis of the pore geometry revealed that Epx4 contains up to four constrictions. In single-molecule measurements, Epx4 detected cationic polypeptides with a higher event frequency than α-hemolysin (αHL). With machine-learning-assisted analysis, Epx4 achieved an ROC AUC score of 0.82 and an F1 score of 0.72, both of which were higher than those obtained with αHL. Our findings suggest that Epx4 is a promising candidate for developing nanopore sensors with high accuracy for protein analysis.

BackgroundPseudomyxoma peritonei (PMP) is a rare peritoneal malignancy primarily originating from appendiceal mucinous neoplasms. Prognosis is influenced by histologic subtype, tumor markers, and the completeness of cytoreduction. Cytoreductive surgery (CRS) combined with hyperthermic intraperitoneal chemotherapy (HIPEC) is the standard treatment. This study aims to identify key factors affecting long-term prognosis in appendiceal PMP patients and develop a dynamic nomogram for postoperative survival prediction.MethodsThis study retrospectively analyzed the clinical data of 153 patients with appendiceal-origin PMP confirmed by postoperative pathology from Jinan Central Hospital between October 2017 and December 2024. The primary endpoint was postoperative overall survival (OS). Patients were randomly divided into modeling (107 cases) and validation (46 cases) groups. Nonlinear relationships between continuous variables and survival time were assessed using restricted cubic splines (RCS) and ANOVA. Univariate Cox regression identified risk factors for postoperative OS, while Lasso regression generated a new variable, the LASSO-derived inflammation–nutrition score (LINS), from selected hematological indices. Multivariate Cox regression analysis identified independent prognostic factors, including histological type, preoperative peritoneal cancer index (PCI), CA125, LINS, and Completeness of Cytoreduction (CC) score (P < 0.05). A nomogram was developed to predict 1-, 3-, and 5-year postoperative OS. Model performance was validated using time-dependent ROC curves and calibration curves. Additionally, a dynamic nomogram was created using the Shiny application framework to predict survival probabilities in real-time.ResultsThe 1-, 3-, and 5-year survival rates for the 153 patients were 82.54%, 66.00%, and 56.76%, respectively. The multivariate Cox regression identified LINS, CA125, preoperative PCI, histological type, and CC score as independent prognostic factors (P < 0.05). The nomogram demonstrated good predictive ability, with time-dependent ROC AUC values for the modeling group of 0.945, 0.823, and 0.736 at 1, 3, and 5 years. Calibration curves showed good agreement between predicted and observed survival. A dynamic survival prediction tool was developed and is accessible via a Shiny app ( https://lq1999.shinyapps.io/Dynamic/).ConclusionHistological type, preoperative PCI, CA125, LINS, and CC score are independent prognostic factors in appendiceal PMP. The developed nomogram offers accurate survival predictions and can be updated with new data for improved clinical decision-making.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12957-026-04363-x.

Metabolic determinants of oocyte quality and embryonic development remain incompletely understood. Here, we profiled metabolites in human cumulus cells (CCs) and follicular fluid (FF), validated in two mouse models, and integrated transcriptomics with receptor blockade to define mechanisms. In human CCs, adenosine was higher in cycles yielding fewer high-quality embryos and discriminated embryo quality (ROC AUC = 0.75). Conversely, FF adenosine was reduced in the same context. In mice, low-quality oocytes and their associated CCs accumulated adenosine, revealing an intra- vs extracellular disequilibrium. The imbalance aligned with reduced expression of the adenosine transporters ENT1/ENT2 and the gap-junction component CX37. Functionally, supplementation with exogenous adenosine restored early embryonic development from low-quality oocytes, lowering oxidative stress and spindle/chromosome errors via adenosine receptors. Smart-seq2 transcriptomic analysis and functional experiments showed partial normalization of programs governing meiosis and cellular stress responses, including correction of CycB1/Cdc27 (MPF/APC/C) and JNK-linked pathways. Together, we identify adenosine disequilibrium as a metabolic fingerprint of poor oocyte competence and show that receptor-mediated adenosine signaling tunes MPF-related and oxidative stress pathways to rescue developmental potential. These findings provide a mechanistic and translational basis for early prediction and culture optimization in ART.

Breast density impacts cancer detection by masking tumors within fibroglandular tissue and there are disparities in screening outcomes across racial groups. However, it remains unclear whether these differences reflect inherent tissue characteristics or systemic bias. To isolate the effect of breast density on lesion detectability across racial subgroups, we conducted a retrospective case-control study using raw tomosynthesis projections from 902 women (453 cases, 451 matched controls) across BI-RADS density categories and self-reported race. Identical in-silico spiculated masses (8-15 mm) and microcalcification clusters (10-14 mm) were inserted into the projections using a calibrated lesion model. Images were reconstructed in the same manner to avoid proprietary processing in lesion detection. Lesion detectability was assessed with Channelized Hotelling Observers. Regression and causal mediation analyses examined the relationships between race, density, and detectability. As result, detectability decreased with increasing density; for masses, the area under the receiver operating characteristic curve (ROC AUC) reduced significantly from 0.93 to 0.85 (BIRADS A to D), whereas for microcalcifications AUC decreased from 0.85 to 0.78 across the same density range. Discrimination remained higher for masses than calcifications (AUC=0.89 vs. AUC=0.82). Stratified analyses showed slightly higher detectability in Non-Hispanic Black women compared with Non-Hispanic White and Asian American women, largely reflecting differences in density. Mediation analysis revealed that breast density accounted for 38-55% of the observed race-associated detectability differences. Mediation analyses have shown that density is the dominant factor in detectability. These findings support the development of calibrated detection models and personalized screening strategies that account for breast density.

Diabetic retinopathy (DR) is still one of the main reasons for vision loss worldwide, especially in places where people do not have easy access to regular eye checkups. Early and accurate disease detection is important to avoid permanent damage, but traditional methods are slow and sometimes inconsistent. This study proposes a deep learning framework that combines convolutional neural networks (CNNs), vision transformers, transfer learning, and ensemble techniques to improve DR detection. We used the APTOS 2019 dataset and tested the capabilities of 23 different pre-trained models. Then, we fine-tuned the top models and designed hybrid architectures by combining the best-performing CNNs and transformers in parallel and sequential ways to capture both image spatial features in short and long contexts. The best performance came from combining the top sequential hybrid models using the soft voting architecture, where we got an accuracy of 93.10%, ROC AUC of 99.22%, and F1-score of 93.07%. The optimized model showed that mixing different models and using ensemble methods can lead to better and more stable DR detection decisions. Our approach is a step toward building a reliable and automated system that could help doctors in real-world settings.