SEARCH · PubMed Health

Results for “Real-world data”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 163 records · Page 9Linked to original sources

PubMed2026

Machine learning-based analysis of oral rinse samples to identify candidate proteomic signatures for severe periodontitis: a pilot study.

This pilot study investigated whether candidate protein signatures from oral rinse samples can distinguish patients with severe periodontitis (stage III/IV) and its subtypes, generalized and localized periodontitis, from non-periodontitis controls. Participants rinsed with phosphate-buffered saline, and samples were analyzed using a Proximity Extension Assay targeting 92 inflammatory and 92 immuno-oncology proteins. A machine learning approach using repeated nested cross-validation and SHAP was implemented to identify protein signatures. The study included 38 patients (18 with localized periodontitis and 20 with generalized periodontitis) and 16 controls. After data preprocessing, 54 samples and 141 proteins were retained. Proteins Gal-1, HGF, TNFSF14, CD27, and ARG1 distinguished periodontitis from controls (ROC-AUC = 0.85, 95% CI 0.82, 0.87). For generalized periodontitis, we found a protein signature including TNFSF14, Gal-1, STAMBP, MUC-16, S100A12, HGF, CASP-8, CD27, LAP TGF-β1, TNFRSF9, and uPA (ROC-AUC = 0.92, 95% CI 0.90, 0.94). For localized periodontitis, we identified ARG1 (ROC-AUC = 0.72, 95% CI 0.68, 0.76). No proteomic signature distinguishing generalized periodontitis from localized periodontitis was identified. This pilot study indicated that oral rinses are suitable for proteomic profiling, and there was a putative protein signature that could differentiate periodontitis, generalized periodontitis, and localized periodontitis from controls. These findings warrant validation in larger independent cohorts, including a clearly defined gingivitis group, before real-world non-invasive screening applications can be considered.

Humans↗

PubMed2026

Major cardiovascular event risk of advanced therapies in inflammatory bowel diseases: systematic review and meta-analysis.

BACKGROUND: Patients with chronic immune-mediated disorders (IMIDs), including inflammatory bowel disease (IBD), are at increased risk of cardiovascular disease. While advanced therapies show cardioprotective effects in other IMIDs, their impact on major adverse cardiovascular events (MACE) in IBD remains unclear. We conducted a meta-analysis of randomized controlled trials (RCTs) and observational studies evaluating MACE risk with advanced therapies in IBD. METHODS: Systematic search of PubMed, Embase, and Cochrane Central Register of Controlled Trials identified 43 studies (36 RCTs, including 9 long-term follow-up (LTF) studies, and 7 observational studies) published between 2002 and 2024. Primary analyses estimated odds ratios (OR) for MACE comparing advanced therapy to placebo, with secondary analyses stratifying studies by drug class and length of follow-up. Sensitivity analyses were conducted using alternative methods to account for zero-event data. RESULTS: Placebo-controlled RCTs showed a nonsignificant trend toward reduced MACE risk (OR 0.60; 95% CI 0.24-1.51), with similar findings across sensitivity analyses accounting for sparse and zero-event data. Class-specific trends suggested lower MACE risk with IL-12/IL-23 inhibitors (OR 0.35; 95% CI 0.05-2.21), JAK inhibitors (OR 0.57; 95% CI 0.16-2.06), and a potential increase with Anti-TNF agents (OR: 3.04; 95% CI 0.31-29.47), though none reached statistical significance. LTF studies showed consistent findings. Observational studies suggested lower MACE risk with Anti-TNF therapies (OR 0.29; 95% CI 0.21-0.40), but not with IL-12/IL-23 (OR 4.41; 95% CI 0.49-39.28) or JAK inhibitors (OR 1.57; 95% CI 0.86-2.84). CONCLUSION: Advanced therapies did not demonstrate a clear increase or decrease in cardiovascular risk in IBD. The discrepancies between RCTs and observational studies underscore the urgent need for rigorous-designed observational research with long-term follow-up to evaluate the real-world impact of advanced therapies on MACE risk.

Humans↗

PubMed2026

Perseus: Lineage-Aware Refinement of Kraken2 Taxonomic Classification for Long Read Metagenomes.

MOTIVATION: Long-read metagenomic sequencing improves assembly contiguity and enables genome-resolved analysis of complex microbial communities, but accurate taxonomic classification of long reads and assembled contigs remains challenging. Highly scalable k-mer-based classifiers such as Kraken2 frequently over-assign fine-rank taxonomic labels when applied to long-read data, producing high false positive classification rates driven by sparse or localized k-mer matches, particularly in microbiomes with extensive taxonomic novelty. RESULTS: We present Perseus, a lineage-aware confidence estimation framework for taxonomic classification that models the spatial distribution and hierarchical consistency of k-mer evidence along sequences. This formulation reframes taxonomic classification as a hierarchical confidence estimation problem rather than a single-rank prediction task. Perseus refines k-mer-level taxonomic signals from Kraken2 using a multi-headed convolutional neural network that estimates calibrated confidence scores for taxonomic correctness at each canonical rank. Using these estimates, Perseus confirms assignments, backs off to higher taxonomic ranks, or abstains when evidence is insufficient, prioritizing correctness and lineage consistency over overly specific assignments. Across simulations of taxonomic novelty and real-world metagenomic datasets, Perseus consistently and substantially reduces the false assignment rate while improving precision and lineage-consistent accuracy. These improvements are most pronounced for long reads and assembled contigs, where spatial context enables reliable discrimination between consistent taxonomic signal and spurious matches. AVAILABILITY AND IMPLEMENTATION: Perseus integrates with existing Kraken2 workflows and is available at https://github.com/matnguyen/perseus.

Journal Article↗

PubMed2026

Associations between smart infusion pump-electronic health record interoperability and healthcare outcomes: A systematic review.

OBJECTIVE: This study synthesized available evidence on the associations between smart infusion pump-electronic health record (EHR) interoperability and healthcare outcomes. METHODS: A systematic review of PubMed, CINAHL, Embase, and Scopus databases identified 901 records, which were imported into Rayyan® for duplicate removal, independent screening by three reviewers, and resolution of discrepancies. Eligible studies were peer-reviewed, data-driven, and reported associations between smart infusion pump-EHR interoperability and healthcare outcomes. Studies focused solely on technical validation or interoperability prototypes were excluded. A backward citation search identified additional studies. Two reviewers independently extracted and cross-validated study characteristics using standardized templates. Methodological quality was assessed with the Joanna Briggs Institute Critical Appraisal Tools. RESULTS: Twenty records of 14 full-text studies and 6 conference proceedings were included. Most records reported positive associations between smart infusion pump-EHR interoperability and outcomes related to safety (e.g., medication administration errors, safety-reported events, pump alerts, and compliance with interoperability and drug library), operational efficiency (e.g., programming and documentation time and technical issues), financial performance (e.g., charges captured, and cost avoided), and user experience domains. Most studies used observational designs, reflecting real-world interoperability implementations, where controlling confounding factors is challenging. Limited reporting of baseline characteristics, pump type, and sample sizes limited comparability across studies. CONCLUSIONS: Smart infusion pump-EHR interoperability was associated with improvements in patient safety, efficiency, charge capture, and user experience, with variable findings across studies. Future research should use rigorous methodologies and standardized measures, examine relationships across outcome domains, assess limitations of pump-EHR interoperability, and evaluate underexplored outcomes, including team communication, cognitive workload, and AI-enabled pumps. IMPLICATIONS FOR CLINICAL PRACTICE: Interoperability should be viewed as a component of a broader sociotechnical system, in which technology, user, workflow, clinical content, and organizational practices collectively determine overall effectiveness.

Humans↗

PubMed2025

Integrating explainable AI with multiomics systems biology and EHR data mining for personalized drug repurposing in Alzheimer's disease.

Alzheimer's disease (AD) is characterized by region- and patient-specific molecular heterogeneity, which hinders therapeutic design. In this study, we introduce PRISM-ML (PRecision-medicine using Interpretable Systems and Multiomics with Machine Learning), an open-source integrated analysis pipeline that combines interpretable machine learning with systems biology and electronic health record (EHR) data mining to elucidate the molecular diversity of AD and predict promising drug repurposing opportunities. First, we integrated and harmonized transcriptomic (bulk RNA-seq) and genomic (genome-wide association study) data from 2105 brain samples, each with matched data from the same individual (1363 AD patients, 742 controls; nine tissues), sourced from three independent studies. Random forest classifiers with SHapley Additive exPlanations (SHAP) identified patient-specific biomarkers; unsupervised clustering resolved 36 molecularly distinct "subtissues" (clusters of samples); and gene-gene co-expression networks prioritized 262 high-centrality bottleneck genes as putative regulators of dysregulated pathways. Next, knowledge graph-based drug repurposing predicted six FDA-approved drugs that simultaneously target multiple bottleneck genes and multiple AD-relevant pathways. Notably, in a large U.S. de-identified insurance-claims database (n = 364733), exposure to promethazine, one of the candidate drugs, was associated with a 57-62 % lower incidence of AD versus an active antihistamine comparator (adjusted hazard ratio 0.38; inverse-probability weighted 0.43; both p < 0.001), providing real-world support for its repurposing potential. In summary, PRISM-ML, as an explainable multi-omics analysis pipeline, is readily transferable to other complex diseases, advancing precision medicine.

Computational Biology↗

PubMed1979

Framing pictures: the role of knowledge in automatized encoding and memory for gist.

In general, frame theories are theories about the representation and use of knowledge for pattern recognition. In the present article, the general properties of frame theories are discussed with regard to their implications for psychological processes, and an experiment is presented which tests whether this approach yields viable predictions about the manner in which people comprehend and remember pictures of real-world scenes. Normative ratings were used to construct six target pictures, each of which contained both expected and unexpected objects. Eye movements were then recorded as subjects who anticipated a difficult recognition test viewed the targets for 30 sec each. Then, the subjects were asked to discriminate the target pictures from distractors in which either expected or unexpected objects had been changed. One consequence of the embeddedness of frame systems is that global frames may function as "semantic pattern detectors," so that the perceptual knowledge in them could be used for relatively automatic pattern recognition and comprehension. Thus, subjects might be able to identify expected objects by using automatized encoding procedures that operate on global physical features. In contrast, identification of unexpected objects (i.e., objects not represented in the currently active frame) should generally require more analysis of local visual details. These hypotheses were confirmed with the fixation duration data: First fixations to the unexpected objects were approximately twice as long as first fixations to the expected objects. On the recognition test, subjects generally noticed only the changes that had been made to the unexpected objects, despite the fact that the proportions of correct rejections were made conditional on whether the target objects had been fixated. These data are again consistent with the idea that local visual details of objects represented in the frame are not neccesary for identification and are thus not generally encoded. Further, since subjects usually did not notice when expected objects were deleted or replaced with different expected objects, it was concluded that if two events instantiate the same frame, they may often be indistinguishable, as long as any differences between them are represented as arguments in the frame. Thus, for the most part, the only information about an event that is episodically "tagged" is information which distinguishes that particular event from others of the same general class. The data reinforce the utility of a frame theory approach to perception and memory.

Discrimination Learning↗

PubMed2024

DNABERT-S: Pioneering Species Differentiation with Species-Aware DNA Embeddings.

We introduce DNABERT-S, a tailored genome model that develops species-aware embeddings to naturally cluster and segregate DNA sequences of different species in the embedding space. Differentiating species from genomic sequences (i.e., DNA and RNA) is vital yet challenging, since many real-world species remain uncharacterized, lacking known genomes for reference. Embedding-based methods are therefore used to differentiate species in an unsupervised manner. DNABERT-S builds upon a pre-trained genome foundation model named DNABERT-2. To encourage effective embeddings to error-prone long-read DNA sequences, we introduce Manifold Instance Mixup (MI-Mix), a contrastive objective that mixes the hidden representations of DNA sequences at randomly selected layers and trains the model to recognize and differentiate these mixed proportions at the output layer. We further enhance it with the proposed Curriculum Contrastive Learning (C2LR) strategy. Empirical results on 23 diverse datasets show DNABERT-S's effectiveness, especially in realistic label-scarce scenarios. For example, it identifies twice more species from a mixture of unlabeled genomic sequences, doubles the Adjusted Rand Index (ARI) in species clustering, and outperforms the top baseline's performance in 10-shot species classification with just a 2-shot training. Model, codes, and data is publicly available at https://github.com/MAGlCS-LAB/DNABERT_S.

Journal Article↗

PubMed2026

Effects of short-bout accumulated exercise on postprandial metabolism in adults: A systematic review and meta-analysis.

OBJECTIVE: To systematically evaluate the acute effects of short-bout accumulated exercise (SBAE) interrupting prolonged sedentary behaviour on postprandial glucose, insulin, and triglycerides in adults. METHODS: Systematic searches in PubMed, Web of Science, Embase, Cochrane Library, CINAHL, SPORTDiscus, and CNKI (inception to 10 December 2025) identified randomised crossover trials comparing SBAE (≤10 min/bout, inter-bout interval ≥30 min or adequate recovery) with continuous sedentary behaviour. Outcomes included postprandial glucose, insulin, and triglyceride AUCs. Standardised mean differences (SMD) were pooled using random-effects models. RESULTS: Thirty-one publications reporting data from 29 independent cohorts involving 573 unique participants (mean age 47.8 ± 20.3 years; 47.5% female; mean BMI 29.0 ± 5.1 kg/m²) were included. Compared with continuous sedentary behaviour, SBAE significantly reduced glucose AUC (SMD = -0.53, 95% CI: -0.71 to -0.34, P < 0.001) and insulin AUC (SMD = -0.58, 95% CI: -0.85 to -0.31, P < 0.001), but not triglyceride AUC (SMD = -0.17, 95% CI: -0.48 to 0.15, P = 0.306).Exploratory subgroup analyses showed statistically significant reductions in glucose and insulin for walking and for inter-bout intervals < 60 min, but not for standing alone or intervals ≥ 60 min. A statistically significant insulin-lowering effect was observed in obese individuals. CONCLUSION: SBAE acutely improves postprandial glucose and insulin control. Exploratory subgroup analyses showed statistically significant effects for walking and for inter-bout intervals < 60 min, but these comparisons are observational and no formal interaction test was conducted. These findings provide preliminary evidence for acute SBAE in sedentary populations, though long-term health effects and real-world generalisability require further investigation.

Adult↗

PubMed2025

DNABERT-S: pioneering species differentiation with species-aware DNA embeddings.

SUMMARY: We introduce DNABERT-S, a tailored genome model that develops species-aware embeddings to naturally cluster and segregate DNA sequences of different species in the embedding space. Differentiating species from genomic sequences (i.e. DNA and RNA) is vital yet challenging, since many real-world species remain uncharacterized, lacking known genomes for reference. Embedding-based methods are therefore used to differentiate species in an unsupervised manner. DNABERT-S builds upon a pre-trained genome foundation model named DNABERT-2. To encourage effective embeddings to error-prone long-read DNA sequences, we introduce Manifold Instance Mixup (MI-Mix), a contrastive objective that mixes the hidden representations of DNA sequences at randomly selected layers and trains the model to recognize and differentiate these mixed proportions at the output layer. We further enhance it with the proposed Curriculum Contrastive Learning (C2LR) strategy. Empirical results on 28 diverse datasets show DNABERT-S's effectiveness, especially in realistic label-scarce scenarios. For example, it identifies twice more species from a mixture of unlabeled genomic sequences, doubles the Adjusted Rand Index (ARI) in species clustering, and outperforms the top baseline's performance in 10-shot species classification with just a 2-shot training. AVAILABILITY AND IMPLEMENTATION: Model, codes, and data are publically available at https://github.com/MAGICS-LAB/DNABERT_S.

Sequence Analysis, DNA↗

PubMed2026

Temporal shifts in gyrA mutation types and sublineage replacement in ST11 Salmonella enterica serovar Enteritidis over a decade (2014-2023): A genomic epidemiological study in Guangxi, China.

The overuse or abuse of antibiotics drives the global health threat of antimicrobial resistance. Although bans on certain veterinary antibiotics, such as colistin, have proven effective, the impact of fluoroquinolone stewardship on the evolution of the foodborne pathogen Salmonella enterica serovar Enteritidis (S. Enteritidis) remains unclear. Here, we conducted a decade-long (2014-2023) retrospective longitudinal genomic epidemiological analysis of 441 ST11 S. Enteritidis isolates from Guangxi, China, alongside a global reference dataset of 4297 genomes. Our aim was to elucidate the effect of real-world antibiotic stewardship on the shift of gyrA point mutations and lineage distribution. Surveillance identified three global epidemic clade sublineages (GEC-L2, L3, L4), with the multidrug-resistant GEC-L4 (i.e., GC-c or MMC2), characterized by the gyrA mutation with amino acid substitution D87Y, being domestically dominant (70.07%, 309/441). Following China's 2016 ban on the veterinary use of critical fluoroquinolones, the proportion of the highly resistant GEC-L4 sublineage decreased continuously (from 86.84% in 2017 to 56.00% in 2023), while the less resistant GEC-L3 sublineage (i.e., GC-b or MMC1), mainly characterized by gyrA D87G, increased simultaneously (from 13.16% to 44.00%). This phenomenon might be attributed to the fact that the GEC-L4 sublineage exhibited a higher fitness cost compared with the GEC-L3 sublineage, as confirmed by the competition assay. A Random Forest Model validated that the gyrA mutation with amino acid substitution D87Y was the paramount feature for these sublineages' identification. In contrast, global data showed a continuous increase in gyrA mutations (from 8.63% in 2006 to 68.85% in 2024), primarily D87Y (from 1.44% to 31.15%) and D87N (from 4.32% to 22.95%), correlating with rising average fluoroquinolone consumption. This study provides direct genomic evidence that national-level antibiotic stewardship can drive the replacement of highly resistant sublineages with moderately resistant ones. These findings offer crucial scientific evidence for evaluating the impact of antibiotic management policies and inform strategies for the rational use of antimicrobials.

China↗

PubMed2026

SPEN inactivation drives resistance to androgen receptor pathway inhibitors in metastatic prostate cancer.

PURPOSE: Treatment intensification with androgen receptor pathway inhibitors (ARPIs) has become the standard of care for patients with metastatic prostate cancer. However, there remains an unmet need to identify biomarkers for treatment resistance. Here, we identify SPEN inactivation as a driver of ARPI resistance. EXPERIMENTAL DESIGN: Pre-clinical studies were performed in LNCaP and VCaP cell lines. Data from a nationwide prostate cancer clinico-genomic database were extracted. Log-rank test and Cox proportional hazards models were used to compare time to next treatment (TTNT) on ARPI with/without SPEN mutations. SPEN immunohistochemistry was performed on a rapid autopsy metastatic tissue microarray. RESULTS: SPEN was identified as a top enzalutamide resistance hit in an unbiased genome-wide loss-of-function screen. SPEN inactivation results in upregulation of cell cycle proliferation and basal/stem cell activity as well as increased translation of pro-oncogenic genes. In a large patient cohort (N=6828), SPEN mutations are enriched following treatment with ARPIs (2.1% to 3.6%, p=0.001) and correlate with shorter TTNT on ARPI in patients with metastatic hormone-sensitive prostate cancer (6.4 vs 29.7 months, HR 2.67, p=0.02). In a metastatic rapid autopsy cohort (N=181), low SPEN H-score is associated with shorter time on abiraterone (5.0 vs 7.9 months, p=0.023) in metastatic castration-resistant prostate cancer. CONCLUSIONS: In real-world cohorts, loss of SPEN function across genomic, transcriptomic, and protein levels is associated with reduced benefit from ARPI therapy in metastatic prostate cancer. These findings identify SPEN inactivation as a clinically relevant biomarker of ARPI resistance that warrants prospective evaluation to guide treatment selection.

Journal Article↗

PubMed2024

Increased risk of hearing loss associated with MT-RNR1 gene mutations: a real-world investigation among Han Taiwanese Population.

BACKGROUND: Previous studies have implicated inherited mutations in mitochondrial DNA (mtDNA) in sensorineural hearing loss (SNHL). However, the definitive association between mitochondrial 12S rRNA (MT-RNR1) variants and hearing loss in the population has not been well established, particularly in Asia. The objective of this retrospective cohort study was to assess the association between MT-RNR1 variants and the risk of SNHL in patients in Taiwan. METHODS: The cohort included 306,068 participants from Taiwan between January 2003 and December 2020. Participants were classified based on genetic variants, particularly mitochondrial mutations (rs267606618, rs267606619, rs267606617). MT-RNR1 variant cases were matched 1:10 with non-mutant patients by age, gender, and visit year, excluding those with pre-existing hearing loss. The primary endpoint was SNHL, identified using specific ICD-TM codes with a 90% positive predictive value. Medication exposure history was determined via self-report or electronic medical records in the hospital. Cox proportional hazard regression models were used to assess the association between MT-RNR1 variants and hearing loss, adjusting for various covariates. Kaplan-Meier survival curves and log-rank tests compared hearing loss incidence between groups. RESULTS: The mean age of the mtDNA variants group is 32.4 years, with a standard deviation of 19.2 years. The incidence density of hearing loss for the mutation group was 36.42 per 10,000 person-years (95% Confidence Interval [CI], 27.21-47.73), which was higher than the 23.77per 10,000 person-years (95% CI, 21.32-26.42) in the wild-type group (p = 0.0036). Additionally, diabetes mellitus was associated with an increased risk of developing SNHL in individuals with MT-RNR1 variants (adjusted hazard ratio = 1.76 [95% CI, 1.00-3.09], p < 0.05). CONCLUSION: This study highlights the increased risk of hearing loss in patients carrying MT-RNR1 variants, particularly those with diabetes mellitus. Future research that integrates genetic and clinical data is crucial for developing more precise interventions to monitor and treat hearing loss in this vulnerable population.

Adolescent↗

PubMed2025

Integration of Gene Expression and Digital Histology to Predict Treatment-Specific Responses in Breast Cancer.

Deep learning models applied to digital histology can predict gene expression signatures (GES) and offer a low-cost, rapidly available alternative to molecular testing at the time of diagnosis. We optimized transformer-based models to infer GES results and applied this approach to pre-treatment H&E-stained biopsies from 1,940 breast cancer patients treated with neoadjuvant chemotherapy in clinical trial and real-world cohorts. The most predictive histology-derived GES for pathologic complete response (pCR) in the I-SPY2 trial was validated in four external cohorts: CALGB 40601, CALGB 40603, a trial of durvalumab plus CT, and standard-of-care CT-treated patients from the University of Chicago. Among HER2-negative patients, a transformer-based model trained using a signature composed of estrogen-regulated genes, proliferation, apoptosis, and interferon response genes predicted pCR with an AUC of 0.794, outperforming models based on clinical features alone (AUC 0.704, p = 0.001), pathologist TIL assessment, and a model trained directly to predict response from I-SPY2 cases. Tertiles of this signature stratify patients into clinically relevant groups with increasing likelihood of complete response, with pCR rates ≥50% in the top tertile regardless of treatment or hormone receptor status. Additional transformer-based signature models predicted response to specific therapies (but not chemotherapy alone), including a HER2 signaling signature in IO-treated patients, and a claudin-low signature in bevacizumab treated patients. In HER2- cohorts with available gene expression data and histology, models trained on expression data performed similarly to digital histology predictions, but the combination of gene expression and histology outperformed histology alone. These findings suggest that histology-based GES provides additive information to RNA sequencing data and can inform precision treatment selection across breast cancer subtypes.

Journal Article↗

PubMed2026

Predicting training outcomes for developmental dyslexia from EEG data.

Developmental dyslexia (DD) is characterised by lower-than-average reading abilities and is diagnosed in approximately 10% of individuals. The societal barriers may limit professional fulfilment and psychological wellbeing of individuals with DD, calling for the development of effective interventions to counteract them. As DD is associated with challenges in both phonological and visuo-attentional domains, different longitudinal training approaches were developed to strengthen them. However, they require a considerable amount of personal, social and economic resources and the outcomes may vary depending on individual differences in behavioural and neurophysiological functionality. Hence, predicting training outcomes might help in developing personalised treatment protocols and optimising the use of resources. In the present work we applied machine learning to resting-state EEG to predict longitudinal training outcomes in adults with DD enrolled in a randomized clinical trial. In particular, one group received a visuo-attentional training combined with transcranial alternating current stimulation (tACS), another group received visuo-attentional training with sham/placebo stimulation, and the third group received a phonological training with sham/placebo stimulation. The improvement in text reading speed was associated with spectral power in low-beta and individual frequencies in the alpha (IAF) and beta (IBF) bands, while the improvement in pseudoword reading was associated with IBF. The findings highlight the potential of capturing neural markers of treatment responsiveness in DD. Future studies should focus on the generalisability of predictive models to real-world settings, while investigating whether specific EEG markers predict responsiveness to distinct remediation protocols, thus supporting the development of personalised interventions.

Humans↗

PubMed2026

Diagnostic Performance of Machine Learning for Systemic Lupus Erythematosus: Systematic Review and Meta-Analysis.

BACKGROUND: Early and accurate diagnosis of systemic lupus erythematosus (SLE) and its organ involvement is essential. Previous reviews of machine learning (ML) in SLE combined heterogeneous tasks and validation strategies and may have overinterpreted model performance. OBJECTIVE: This study evaluated the diagnostic performance of ML and deep learning (DL) models for 3 clinically distinct SLE-related tasks: SLE classification or diagnosis, lupus nephritis (LN) diagnosis, and neuropsychiatric systemic lupus erythematosus (NPSLE) discrimination. We also assessed methodological quality and certainty of evidence. METHODS: PubMed, Embase, Cochrane Library, Web of Science, and IEEE Xplore were searched from January 2014 to April 2026. Eligible peer-reviewed diagnostic accuracy studies developed or validated ML or DL models for 1 of the 3 prespecified tasks, used an accepted reference standard, and provided data for a 2×2 contingency table. Bivariate random-effects meta-analyses with the Hartung-Knapp-Sidik-Jonkman adjustment were used to pool sensitivity and specificity. We reported 95% prediction intervals (PIs), assessed risk of bias using the Quality Assessment of Diagnostic Accuracy Studies for Artificial Intelligence tool (QUADAS-AI; Viknesh Sounderajah [Imperial College London]), and evaluated certainty of evidence using the Grading of Recommendations Assessment, Development, and Evaluation framework for diagnostic test accuracy. RESULTS: Twenty-nine studies were included: 17 for SLE classification, 5 for LN diagnosis, and 7 for NPSLE discrimination. In the primary task-stratified analysis, pooled sensitivity was 0.91 (95% CI 0.86-0.94; 95% PI 0.56-0.99), and pooled specificity was 0.94 (95% CI 0.91-0.96; 95% PI 0.69-0.99), with low heterogeneity (I²=23.9% and 22.9%, respectively). DL models showed a sensitivity of 0.93 and specificity of 0.95, compared with 0.88 and 0.94 for traditional ML models. Certainty of evidence was high for most analyses but low for LN diagnosis because of inconsistency and imprecision. All studies were retrospective, and only 9 of 29 (31%) performed independent external validation. Overall risk of bias was high or unclear in 22 of 29 (75.9%) studies. No study reported model calibration, decision-curve analysis, or net clinical benefit. CONCLUSIONS: ML models showed promising diagnostic accuracy across 3 distinct SLE-related tasks, but wide PIs, limited external validation, and pervasive risk of bias restrict conclusions about real-world generalizability. Prospective multicenter studies with standardized tasks and reference standards, independent external validation, and formal assessment of calibration and clinical utility are required before clinical implementation.

Humans↗

PubMed2026

Real-World Experience With TRBC1 Immunohistochemistry Across Cutaneous T-Cell Lymphoma Subtypes: A Large Cohort Study.

T-cell receptor β-chain constant region 1 (TRBC1) immunohistochemistry identifies clonal αβ T-cell populations on tissue sections, but its real-world performance across cutaneous T-cell lymphoma (CTCL) and related infiltrates is uncharacterized. The analytic cohort comprised 665 biopsies (566 patients) with paired T-cell receptor (TCR) clonality testing, classified clinicopathologically as mycosis fungoides (MF; MF-Patch, MF-Plaque, MF-Tumor, and MF-Folliculotropic); MF or Sézary syndrome; primary cutaneous small or medium T-cell lymphoproliferative disorders (LPDs); other CTCL-cutaneous LPDs; or reactive. At the primary <15%/>85% threshold, TRBC1 IHC achieved 85.8% sensitivity (337/393), 79.8% specificity (217/272), 86.0% positive and 79.5% negative predictive value, and 83.3% accuracy. Sensitivity was lowest in MF-Patch (84.2%). Three-reader agreement (Fleiss κ = 0.943) fell to κ = 0.776 in 176 reflexed biopsies, with disagreement concentrated on MF-Patch and CD30-positive LPDs. Monotypic TRBC1 predicted neoplasia, with odds rising with infiltrate density: MF-Patch (odds ratio, 5.41), MF-Plaque (10.40), MF/Sézary syndrome with MF-Tumor (15.19), and CTCL-cutaneous LPD (18.16). Polytypic TRBC1 was associated with reactive disease (odds ratio, 53.65), effectively excluded clonality (negative likelihood ratio, 0.18), and was uniformly observed in an independent 270-biopsy reactive cohort. For observer-independent validation, digital image analysis-derived TRBC1 quantification (QuPath) was applied to a stratified random subset of 250 biopsies representative of the cohort's tumor-burden distribution. The digital read-tracked molecular clonality (85.1% sensitivity, 80.1% specificity against TCR; area under the curve, 0.842) agreed with the dermatopathologist's manual read in 87.6% of cases (κ = 0.752), with the data-derived cutoff matching the prespecified <15%/>85% threshold value. Discordance was directional for both manual scoring and digital quantification: in MF-Patch, 25 of 38 (65.8%) and 12 of 17 (70.6%) cases were polytypic with monoclonal TCR (false-negative-dominant); in reactive biopsies, 34 of 45 (75.6%) and 19 of 20 (95%) were monotypic with polyclonal TCR (false-positive-dominant). These findings support a TRBC1-first approach, reserving reflex TCR testing for borderline expression or clinicopathologic discordance, preserving diagnostic accuracy while reducing molecular testing and reimbursement-based costs.

Sézary syndrome↗

PubMed2026

2024-2025 BNT162b2 KP.2 COVID-19 full season vaccine effectiveness from vaccine registries linked to administrative claims in two states: A cohort study in non-immunocompromised adults.

BACKGROUND: Data on effectiveness of COVID-19 vaccinations during the 2024-2025 respiratory season are limited, particularly among those with underlying medical conditions (UMC). We estimated BNT162b2 KP.2 vaccine effectiveness (VE) against COVID-19-associated hospital admission, emergency department (ED), and urgent care (UC) visits in two U.S. states. METHODS: Retrospective cohort study of non-immunocompromised adults living in Louisiana or California, with ≥1 year prior continuous enrollment in insurance plans contributing to the HealthVerity claims database beginning August 22, 2024. The effectiveness of BNT162b2 KP.2 vaccine (2024-2025 formulation, hereafter referred to as BNT162b2), measured as a time-varying exposure against hospital admission, ED, or UC encounters with International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) code U07.1 was calculated as 1 - adjusted hazard ratio using Cox proportional hazard models adjusted for age group, sex, state, insurance payor, presence or absence of UMCs, and pre-index healthcare utilization. Stratifications included those aged 65 years and older, those aged 18-64 years with UMCs, and those aged 18-64 years without UMCs. RESULTS: The cohort included 6,256,421 individuals (93% California, 7% Louisiana); 330,565 (5%) received the BNT162b2 vaccine. Vaccinated individuals were older and had more comorbidities, wellness visits, and prior influenza vaccination. Overall, 66% of the study population had ≥1 UMC; the most prevalent conditions were obesity (25%), history of immunocompromised conditions (23%), and mental health conditions (19%). COVID-19-related encounter rates for ED, UC or hospitalization were lower among vaccinated compared to unvaccinated persons (25.1 vs 36.3 per 100,000 person-months). Among all adults, VE was 37% against hospitalization, 12% against ED/UC encounters, and 16% against ED/UC/hospitalization encounters. Results were similar across age groups and UMCs. CONCLUSIONS: BNT162b2 provided protection against COVID-19-associated outcomes of ED, UC or hospitalization among non-immunocompromised U.S. adults, including those with UMCs, over the course of the 2024-2025 respiratory virus season, supporting continued vaccine recommendations. REGISTRATION: This study was posted on clinicaltrials.gov prior to analyses (NCT06923137).

Adolescent↗

PubMed2026

Real-world clinical utility of exome sequencing in pediatric drug-resistant epilepsy: Experience from a tertiary center in Thailand.

BACKGROUND: Genomic testing has increasingly contributed to the diagnosis and management of pediatric drug-resistant epilepsy (DRE), particularly in patients with suspected genetic etiologies. This study evaluated the diagnostic yield and real- world clinical utility of whole-exome sequencing (WES) in children with DRE. METHODS: Children with DRE and seizure onset before 15 years of age were enrolled between January 2020 and December 2023. Clinical data, including demographics, seizure characteristics, developmental history, electroencephalography (EEG), brain magnetic resonance imaging (MRI), and prior investigations, were reviewed. WES was performed in all probands and, when available, their parents. Variants were interpreted according to standard guidelines. Clinical utility and 1-year seizure and developmental outcomes were assessed from follow-up records. RESULTS: Fifty-six patients (23 males, 33 females) were included. The median age at seizure onset was 1 year (interquartile range [IQR] 0.3-4 years), and 96.4% had developmental comorbidities. Pathogenic or likely pathogenic variants were identified in 39% (22/56), with the highest diagnostic yield in children with seizure onset before 3 years of age. Channelopathies accounted for most genetically solved cases (68%), predominantly involving sodium channel genes. Genetic diagnoses provided clinical utility in 73% (16/22) of solved cases by guiding treatment and precision management. At 1-year follow-up, genetically solved patients showed more favorable seizure and developmental outcomes than those with genetically unsolved patients. CONCLUSION: WES achieved a 39% diagnostic yield and substantial clinical utility in pediatric DRE, particularly in early-onset and channelopathy-related disorders. These findings support early molecular diagnosis to facilitate genotype-informed management in appropriately selected children. However, the more favorable developmental and seizure outcomes observed in genetically solved patients should be interpreted with caution, as they may have been influenced by multiple factors beyond genetic diagnosis. In resource-limited settings, careful clinical phenotyping remains essential for treatment decisions and for prioritizing children for genomic testing.

Clinical utility↗