SEARCH · Search PubMed
Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Explore the source record for details and available documents.
BACKGROUND: Predictive modelling for cancer risk, treatment-related complications, and survival is central to precision oncology. Conventional logistic regression (LR) and Cox proportional hazards (CoxPH) regression remain widely used but are limited when modelling nonlinear interactions, high-dimensional imaging features, and multimodal clinical-metabolic predictors. Artificial intelligence (AI) and machine learning (ML) methods offer expanded capability through automated feature extraction, ensemble learning, and flexible survival modelling, but the evidence on when AI/ML adds value over conventional models across cancer sites and predictive tasks remains fragmented. OBJECTIVE: To systematically evaluate the methodological performance, validation strategies, and translational limitations of AI/ML models compared with conventional statistical models in published predictive-modelling studies for breast, colorectal, or pancreatic cancer. METHODS: PubMed, Scopus, and Web of Science were searched for studies published between January 2019 and March 2025. Two reviewers independently conducted title-and-abstract screening, full-text eligibility assessment, and PROBAST risk-of-bias assessment. Sixty-five studies (n = 907,567 participants) were narratively synthesised by cancer site, predictive task, model family, comparator, validation strategy, predictor modality, and calibration or explainability reporting. RESULTS: The 65 studies comprised breast cancer (n = 35), colorectal cancer (n = 21), and pancreatic cancer (n = 9). AI/ML superiority over LR and CoxPH was task- and data-dependent. CNN- and U-Net-based models predominated in imaging and body-composition tasks, tree-based ensembles consistently outperformed LR for tabular perioperative complication prediction, and CoxPH remained competitive, and in the largest pancreatic risk study, superior to XGBoost (C-index 0.802 vs 0.723) in well-structured datasets. PROBAST analysis-domain risk was moderate in 54 of 65 studies (83%), driven by limited external validation, sparse calibration reporting (11/65), and few decision-curve analyses (7/65). CONCLUSION: AI/ML adds the most methodological value in imaging-derived feature extraction and nonlinear perioperative prediction, while conventional regression remains preferable in large, structured datasets with linear predictors. Clinical translation requires standardised body-composition definitions, external validation, calibration assessment, decision-curve analysis, and explainability, in line with TRIPOD+AI and CLAIM standards.
AIM: To provide evidence for selecting and developing reliable clinical assessment tools for hypoglycemia in diabetic kidney disease patients during haemodialysis. DESIGN: Review. METHODS: Systematic searches were performed in 9 Chinese and English databases to collect literature regarding the development of hypoglycemia risk prediction models in haemodialysis patients with diabetic kidney disease. Two reviewers independently performed literature screening, data extraction, risk-of-bias assessment, and applicability evaluation. The Prediction Model Risk of Bias Assessment Tool was used to assess the risk of bias and applicability of the included studies. Meta-analysis was conducted using R software. DATA SOURCES: CNKI, Wanfang, VIP, CBM, PubMed, Cochrane Library, EMbase, Web of Science, and CINAHL. The search period covered from the establishment date of each database to December 2025. RESULTS: Six studies, comprising six prediction models, were included. Two studies performed internal validation, and three conducted external validation. All models reported the area under the curve, ranging from 0.813 to 0.866, and calibration measures. Four studies were rated as having a high risk of bias, while all six demonstrated good overall applicability. The meta-analysis showed that the pooled AUC value of the six studies was 0.846 (95% CI: 0.823-0.867). CONCLUSION: Research on hypoglycemia risk prediction models in haemodialysis patients with diabetic kidney disease remains in the developmental stage. Although the included prediction models exhibited satisfactory apparent discriminatory ability and clinical applicability, most of the original studies suffered from a high risk of bias and lacked adequate validation. The true predictive performance and clinical application value of these models remain to be further verified. Accordingly, routine and unconditional clinical application is not recommended at this stage. Future studies should include more high-quality, multicenter external validation and develop models with high generalizability, favourable clinical applicability, and robust predictive performance to facilitate early identification of hypoglycemia risk in this population. IMPACT: This study systematically evaluated the hypoglycemia risk prediction models for diabetic kidney disease patients during haemodialysis, and the research on hypoglycemia risk prediction models for maintenance haemodialysis patients during dialysis is still in the development stage. This study provides a reference for clinical medical staff to select or develop hypoglycemia risk prediction and assessment tools for diabetic kidney disease patients during haemodialysis. REPORTING METHOD: This study was conducted in accordance with the relevant guidelines of the EQUATOR Network and followed the TRIPOD-SRMA Checklist. PATIENT OR PUBLIC CONTRIBUTION: No patient or public contribution. TRIAL REGISTRATION: PROSPERO: CRD420251243352.
BACKGROUND: Early ACL injury risk identification in athletes is essential. This systematic review examines machine learning (ML) models for predicting ACL injuries, evaluating their methodological quality, performance, and reliability. METHOD: A comprehensive electronic search was conducted across PubMed, Scopus, Web of Science, and IEEE Xplore databases, supplemented by Google Scholar for grey literature, covering articles published between January 1, 2015, and August 30, 2025. Eligible studies were appraised using the Prediction Model Study Risk of Bias Assessment Tool (PROBAST) for methodological quality and risk of bias, and the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) guidelines for quality of evidence. RESULTS: Ten studies were included. PROBAST showed eight studies had moderate risk of bias and two low risk. TRIPOD found only two studies met quality criteria. ML models included logistic regression (n = 5), support vector machines (n = 4), k-nearest neighbor (n = 3), decision trees (n = 3), random forests (n = 5), neural networks (n = 2), linear discriminant analysis (n = 1), and pre-trained CNNs (n = 1). AUC ranged from 0.63 to 0.98. Accuracy (reported in six studies) ranged from 26% to 95%; however, these values should be interpreted with caution due to the absence of confidence intervals, lack of class imbalance handling, and limited external validation across studies. Tree-based ensemble methods such as random forest achieved competitive accuracy (74-86%), while SVM, a non-ensemble classifier, reported accuracy ranging from 71% to 95%; however, the highest values were obtained in studies with notably small sample sizes (n = 12 to n = 39), raising concerns about overfitting and generalizability. CONCLUSION: Current ML algorithms show promise for identifying athletes at high ACL injury risk and detecting relevant risk factors. Although study quality was generally satisfactory, future research should prioritize external validation and model interpretability to support clinical translation.
INTRODUCTION: Risk prediction models for gestational diabetes mellitus (GDM) offer potential for early identification and targeted prevention. External validation is crucial to assess model performance across diverse populations. Despite the availability of numerous GDM prediction models, limited evidence exists on their external validation frequency, methodological quality, and clinical applicability. This systematic review evaluated externally validated GDM prediction models, focusing on methodological rigor, reporting standards, and clinical relevance to inform future research and implementation. MATERIAL AND METHODS: Databases including Ovid MEDLINE, Embase, Scopus, Emcare, and CINAHL were searched up to May 1, 2025. Studies reporting external validation of GDM risk prediction models were included. Two reviewers independently screened studies. Data were extracted using the CHARMS framework, and risk of bias and applicability were assessed using PROBAST+AI. The study protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO; CRD420251125758). RESULTS: Twenty-six studies validated 33 models, with validation sample sizes ranging from 50 to 75 161. Over half used the IADPSG criteria to define GDM. Discrimination metrics were commonly reported, but calibration, overall performance, and clinical utility were often lacking. Meta-analysis was feasible for only four models: Teede et al., Nanda et al., Naylor et al., and Van Leeuwen et al., each showing fair discrimination. The Teede et al. model was the most widely validated, with 11 external validations across six continents and a pooled AUC of 0.72 (95% CI: 0.67-0.76). Despite fewer validations, the Nanda et al. model achieved the highest pooled discrimination (5 validations; pooled AUC 0.77, 95% CI: 0.74-0.80). The Naylor et al. and van Leeuwen et al. models also underwent meta-analysis, as sufficient external validation studies were available to support comparative performance assessment. Notably, 69.23% of studies had a high risk of bias. CONCLUSIONS: While many models showed acceptable predictive performance, most validations were methodologically weak. Future studies should follow best-practice guidelines and promote scalable validation strategies, such as algorithm sharing, to enhance clinical utility.
OBJECTIVE: To systematically review and evaluate published risk prediction models for perioperative blood transfusion in patients undergoing total hip or knee arthroplasty (THA/TKA). METHODS: We systematically searched PubMed, Web of Science, the Cochrane Library, and Embase from inception to May 31, 2025. Two researchers independently screened the literature, extracted data, and assessed the risk of bias and applicability using the Prediction model Risk Of Bias Assessment Tool (PROBAST). The area under the receiver operating characteristic curve (AUC) values were pooled via a meta-analysis using Stata 18.0. RESULTS: d Fourteen studies containing 36 prediction models were included. The incidence of blood transfusion among THA/TKA patients ranged from 3.2% to 30.8%. Preoperative hemoglobin (Hb) level, tranexamic acid (TXA) use, operative duration, intraoperative blood loss, and age were the most frequently incorporated predictors. Model sensitivity ranged from 58% to 94.5%, and specificity ranged from 71.3% to 94%. Meta-analysis showed that the pooled AUC value of the 13 validated models was 0.87 (95% CI: 0.85-0.90), suggesting good discriminatory performance. All models were rated as having a high risk of bias. The applicability of four studies was rated as unclear. CONCLUSION: Although the included studies demonstrated promising discriminative ability of prediction models for blood transfusion in THA/TKA, all were assessed as having a high risk of bias using the PROBAST tool. Therefore, future research should prioritize the development of models with larger sample sizes, rigorous study designs, and multicenter external validation.
INTRODUCTION: Accurate prediction of peri-operative cardiac complications is critical to optimise pre-operative decision-making. Traditional risk prediction scores, such as the Revised Cardiac Risk Index, show only modest discrimination. Machine learning can model complex, non-linear relationships but their predictive performance compared with traditional scores remains unclear. METHODS: We performed a systematic review and Bayesian network meta-analysis. The primary outcome was postoperative adverse cardiac events following non-cardiac surgery. Prediction models were assessed relative to the Revised Cardiac Risk Index. As many studies evaluated multiple versions of each model type, the highest performing ('best version') and lowest performing ('worst version') results were analysed. Models were ranked using the surface under the cumulative ranking curve (SUCRA). RESULTS: Thirteen studies evaluating 54 models and 927,113 patients were included. Machine learning approaches generally outperformed traditional risk scores. Automated machine learning ranked highest (SUCRA 96.6) showed the greatest improvement in the best version analysis (mean difference (MD) 0.28 (95%CrI 0.16-0.40)) and remained superior in the sensitivity analysis (MD 0.30 (95%CrI 0.14-0.45)). Gradient boosting models showed superior performance over the Revised Cardiac Risk Index across analysis (best version: MD 0.20 (95%CrI 0.14-0.26), worst version: MD 0.18 (95%CrI 0.12-0.25), SUCRA 82.4). The Gupta Perioperative Risk for Myocardial Infarction or Cardiac Arrest score outperformed the Revised Cardiac Risk Index in the best version analysis (MD 0.16 (95%CrI 0.01-0.32)). Between-study heterogeneity was low. None of the included studies externally validated their machine learning models and only six were judged to be at low risk of bias. DISCUSSION: Most machine learning models showed better discrimination than traditional risk scores, with automated machine learning and gradient boosting models ranking highest. However, study quality, calibration reporting and absence of external validation limit immediate clinical adoption. Prospective, multicentre evaluation is required before integration of these models into peri-operative practice.
PURPOSE: To identify risk factors associated with the development of high myopia following congenital cataract surgery and to establish a robust predictive model. DESIGN: Retrospective clinical cohort study. SUBJECTS: This retrospective study included 106 pediatric patients who underwent congenital cataract surgery with primary IOL implantation (mean follow-up 8.19 years). The model was externally validated in an independent cohort of 72 patients with a mean follow-up of 7.83 years. METHODS: Preoperative and postoperative ocular biometric parameters were collected. Risk factors for postoperative high myopia were analyzed using Cox proportional hazards regression, which served as the basis for model construction. The predictive performance of the model was rigorously evaluated for discrimination and calibration. Discriminative ability was quantified using Harrell's C-index and the area under the receiver operating characteristic curve (AUC). Model calibration was assessed via calibration plots by comparing predicted probabilities with actual observed outcomes. Internal validation was performed using a bootstrapping method (500 iterations) to ensure model stability and adjust for potential overfitting. RESULTS: An initial postoperative refraction of <+0.75D, and a higher IOL Power to Axial length Ratio (IOL/AL ratio) were identified as significant risk factors for the development of postoperative high myopia. Shorter preoperative axial length was associated with a greater magnitude of postoperative myopic shift. The predictive model demonstrated robust performance, achieving a C-index of 0.711 (internal validation C-index: 0.713). The area under the receiver operating characteristic curve (AUC) values for predicting high myopia at 5 and 10 years were 0.858 and 0.745, respectively. Furthermore, calibration curves demonstrated excellent agreement between the predicted and observed outcomes throughout the follow-up period. In external validation, the model achieved a C-index of 0.825, 5-year AUC of 0.833, and 10-year AUC of 0.713. CONCLUSIONS: Our analysis established that initial postoperative refraction <+0.75D, and an elevated IOL/AL ratio are key determinants of high myopia risk following surgery. Shorter preoperative axial length was associated with a greater magnitude of postoperative myopic shift. This predictive framework provides clinicians with a practical tool to optimize preoperative IOL selection and identify high-risk infants who require vigilant myopia prevention and balanced amblyopia management.
OBJECTIVE: To screen for independent risk factors for unplanned readmission in elderly patients with mild ischemic stroke, and to construct and validate an online risk prediction calculator based on an interpretable machine learning model, thereby providing a promising practical tool for accurate clinical assessment of 30‑day all‑cause unplanned readmission risk in this population. METHODS: A prospective cohort study was conducted, including 1050 patients aged ≥ 60 years with mild ischemic stroke admitted between August 2023 and September 2024. Participants were randomly divided into a training set (840 cases) and a test set (210 cases) at a ratio of 8:2. Risk factors were screened by univariate analysis and multivariable Logistic regression. Four machine learning models, namely LightGBM, XGBoost, Random Forest, and K‑Nearest Neighbors (KNN), were developed and their performance was evaluated using AUC, accuracy, sensitivity, and specificity as metrics. The SHAP framework was used for interpretability analysis, and an online calculator was subsequently developed based on the optimal model. RESULTS: Univariate analysis showed significant differences (P < 0.05) in 13 factors including age, smoking, AIP, TyG index, HALP score, etc. Multivariable Logistic regression identified age (OR = 9.752), smoking (OR = 5.171), AIP (OR = 6.691), TyG index (OR = 4.393), HALP score (OR = 2.831), and ≥ 2 comorbidities (OR = 3.664) as independent risk factors. All four machine learning models demonstrated good predictive performance. Based on a comprehensive evaluation of multiple metrics and computational efficiency, the LightGBM model exhibited the best predictive performance (AUC = 0.884, accuracy = 0.829, sensitivity = 0.812, specificity = 0.875). SHAP analysis showed that age, AIP, TyG index, smoking, and HALP score were key predictors. An online calculator developed based on this model enables individualized risk predictions. CONCLUSION: Key risk factors associated with 30‑day unplanned readmission in elderly patients with mild ischemic stroke were identified. The LightGBM model demonstrated high predictive accuracy, and together with the interpretability analysis and online calculator, offers a practical tool to support clinical risk assessment. However, this tool requires future external validation.
BACKGROUND: Due to the complex pathophysiology of non-traumatic subarachnoid hemorrhage (SAH), accurate risk prediction remains a challenge. Our aim is to develop and validate a comprehensive prognostic model that integrates demographic characteristics, vital signs, laboratory parameters, and more, to provide clinical decision-making support in real-world practice. METHODS: We conducted a retrospective cohort study of 785 Non-traumatic subarachnoid hemorrhage patients. The cohort was randomly divided into a training set (n = 549) and a validation set (n = 236). Feature selection was performed using LASSO regression, followed by backward stepwise Cox regression for optimization. A nomogram was constructed based on independent predictive factors, and model performance was assessed using discrimination, calibration, and decision curve analysis. To prevent immortal-time bias, all predictors were anchored to a fixed early (first-24-hour) measurement window, treatment variables were modelled as binary indicators rather than cumulative exposures, and a five-model sensitivity analysis with baseline-severity adjustment was performed. RESULTS: The development of our model followed a systematic approach: first, 15 potential predictive factors were selected via LASSO regression, which were then refined to 12 independent predictors using backward stepwise Cox regression. The final predictive factors included: Ventilation, AHT, Nimodipine 60 mg, Age, SAPS.II, Input amount, Calcium total, Platelet count, White blood cells, Anion gap, pH, and Chloride. The integrated model demonstrated excellent predictive ability for 7-day, 14-day, and 21-day mortality in both the training set (AUC: 0.972, 0.934, 0.898) and the validation set (AUC: 0.968, 0.948, 0.911). Calibration curves and decision curve analysis confirmed the model's reliability and clinical utility across different time points. We constructed a nomogram for individualized risk prediction. Univariate Kaplan-Meier survival analysis demonstrated significant stratification of survival outcomes by each predictor, while restricted cubic spline analysis revealed non-linear relationships between continuous variables and mortality risk. Random survival forest analysis identified the top three predictive factors (Nimodipine 60 mg, Ventilation, AHT) and compared them with our full 12-variable model, confirming superior performance of the integrated model at all time points. At the 28-day primary endpoint, the model achieved a time-dependent AUC of 0.898 (training) and 0.904 (validation); after restricting predictors to the early baseline window, the leakage-controlled model retained good discrimination (validation C-index 0.803). CONCLUSIONS: Our ICU 28-day mortality prognosis model demonstrated robust performance in predicting ICU 28-day mortality in non-traumatic subarachnoid hemorrhage. The model, through the nomogram, provides individualized risk assessment, aiding clinical decision-making and patient stratification.
BACKGROUND AND OBJECTIVES: Pseudouridine (Ψ) represents one of the most abundant and conserved RNA modifications. Ψ provides an additional hydrogen-bond donor that enhances RNA structural stability and modulates translation. It participates in diverse biological processes, including RNA-protein interactions, splicing, translational control, and stress responses. Aberrant pseudouridylation is implicated in cancer, neurodegenerative disorders, and autoimmune diseases. Despite its biological importance, experimental identification of Ψ sites remains time-consuming and costly, limiting the feasibility of transcriptome-wide profiling. Computational approaches have therefore become essential complements to experimental techniques. However, state-of-the-art machine-learning and deep-learning predictors often suffer from limited generalizability due to small training datasets. To overcome these issues, we aim at constructing new long-sequence datasets and developing a novel Ψ site predictor. METHODS: New long-sequence datasets were constructed as benchmarks for RNA Ψ-site prediction. The Ψ modification sites in RMBase 3.0 were mapped to the reference genomes across three species of human, mouse, and yeast, and the RNA sequences with a length of 201 were generated by extending the upstream and downstream from the mapped, central sites. To eliminate sequence redundancy, the sequences were clustered using CD-HIT with a 70% sequence identity threshold. We developed Meta-PseU, a logistic regression-based meta-classifier that considered 118 machine learning and deep learning classifiers. The datasets and programs are freely accessible at https://github.com/kuratahiroyuki/MetaPseU. RESULTS: By optimizing model configuration, we proposed the Meta-PseU model stacking 32 machine learning and deep learning classifiers out of 118 classifiers. Meta-PseU substantially improved model generalizability, overcoming a key limitation of existing approaches. It greatly outperformed state-of-the-art predictors and achieved increasing accuracy with increasing sequence length. CONCLUSIONS: Long-sequence datasets were newly constructed as benchmarks for RNA Ψ-site prediction. Meta-PseU offers a new framework for robust Ψ-site identification by using long sequences.
OBJECTIVES: To develop an individualized nomogram for predicting disease progression risk in systemic anaplastic large cell lymphoma (sALCL). METHODS: Independent predictors of progression-free survival (PFS) were identified using Cox regression in a multicenter retrospective cohort of 109 sALCL patients (2010-2022). These were incorporated into a three-factor nomogram, evaluated via bootstrapped internal validation (1000 resamples), ROC analysis, C-index, decision curve analysis (DCA), and clinical impact curve (CIC). RESULTS: A total of 29 PFS events occurred during a median follow-up of 31 months. Multivariable modelling selected serum β2-microglobulin elevation, extranodal disease, and front-line chemotherapy choice (CHOP versus CHOPE or BV+CHP) as autonomous progression drivers. Upon internal bootstrap validation, the nomogram yielded strong prognostic accuracy, achieving AUCs of 0.81, 0.85 and 0.87 for 1-, 3- and 5-year progression-free survival, alongside a corrected C-index of 0.779 (95% CI: 0.699 - 0.861). Calibration plots showed close agreement between predicted and observed outcomes, while DCA confirmed superior net clinical benefit versus conventional IPI or Ann Arbor stratification across multiple decision thresholds. CONCLUSION: This first sALCL-specific nomogram integrates clinical and treatment variables to provide personalized PFS risk estimation. While internally validated, this exploratory, observation-based tool requires external validation and recalibration in prospective cohorts before clinical implementation.
The rapid expansion of metagenomic sequencing has generated vast repositories of viral sequence data that far outpace our capacity to interpret them using conventional approaches. Highly divergent sequences, sparse functional annotation, and taxonomically uneven sampling present fundamental challenges for reference-dependent methods, which lose sensitivity precisely for novel and understudied viruses with high public health relevance. Artificial intelligence (AI) provides a new avenue to address these challenges by enabling predictive inference from viral genomes and proteins while reducing dependence on sequence similarity. In this Review, we discuss representative advances in AI for virus discovery, taxonomic classification and functional annotation, prediction of host range and zoonotic potential, and efforts toward forecasting emerging variants. These advances are transforming viral genomics from a largely descriptive discipline into one with increasing predictive capability. We also critically assess the major challenges that constrain current approaches, including the availability of high-quality and representative datasets, rigorous model evaluation, biological interpretability and responsible governance for increasingly capable AI models.
This research was undertaken to address environmental concerns associated with industrial solid waste and to reduce cement consumption in geotechnical engineering. It specifically investigates the feasibility of using steel slag (SS) and silica fume (SF) as partial substitutes for ordinary Portland cement (OPC) in soil stabilization. The effects of SS, SF, OPC, and initial moisture content on the unconfined compressive strength (UCS) of stabilized soil were investigated through single-factor experiments and response surface methodology (RSM). The results show that SS and SF can synergistically enhance the strength of stabilized soil, although their interaction effect was not statistically significant within the investigated ranges. Compared with soil stabilized solely with OPC, the addition of 18 % SS and 10 % SF reduced OPC consumption by 3 % without compromising strength. Microstructural and compositional analyses further revealed that SS mainly supplied calcium- and silica-bearing components, while SF provided highly reactive silica and micro-filling effects, jointly promoting hydration reactions and improving the compactness of the stabilized soil matrix. As a result, more hydration products were formed in the OPC/SS/SF-stabilized soil than in the OPC-stabilized soil, which contributed to pore filling and strength enhancement. This study provides useful guidance for the sustainable utilization of industrial solid waste and the low-carbon development of soil stabilization materials.
BACKGROUND: Accurately distinguishing malignant pleural effusion (MPE) from non-malignant pleural effusion is clinically important, but the generalisability and methodological quality of machine-learning (ML) models remain uncertain. METHODS: We searched eight databases to 23 April 2026. Diagnostic performance was pooled using random-effects and Reitsma bivariate models, and study quality was assessed using PROBAST+AI. RESULTS: Forty-two studies were included; 17 contributed to the AUC meta-analysis and 14 to the bivariate analysis. The pooled AUC was 0.90 (95 % CI 0.85-0.94; 95 % prediction interval 0.62-0.98), with sensitivity of 0.80 (95 % CI 0.77-0.83) and specificity of 0.87 (95 % CI 0.79-0.92). Only nine studies reported external, temporal or independent validation. Externally validated studies had a lower pooled AUC than studies without external validation (0.83 vs 0.92), with lower specificity observed in the two externally validated studies contributing sensitivity and specificity data. All 42 development assessments had high overall quality concerns, and all 42 model evaluations were judged at high risk of bias. CONCLUSIONS: ML models showed good apparent accuracy for distinguishing MPE from non-MPE, but the evidence was limited by substantial heterogeneity, high risk of bias and scarce external validation. The pooled estimates reflect the average performance of different selected models rather than the expected accuracy of a single clinical test. ML models should be regarded as adjuncts to existing diagnostic pathways until they are confirmed by rigorous multicentre prospective external validation and clinical-impact studies.
BACKGROUND: Accurate classification of missense variants remains a challenging task despite major advances in genomics. Numerous computational models have been developed to assist in variant classification, but often require repeated integration and benchmarking efforts. Ensemble methods have been proposed to overcome the limitations of single predictors, but mostly rely on fixed, predefined weights that constrain their ability to capture interactions among predictive signals. METHODS: We present GenixRL, a dynamic ensemble framework that reformulates model fusion as a reinforcement learning optimization problem. GenixRL uses a Q-learning agent to learn a policy that dynamically weights the probabilistic outputs of complementary predictors, including BayesDel (addAF and noAF), ClinPred, and MetaRNN. Replacing static weighting with policy learning allows GenixRL to adaptively identify optimal weightings and substantially improve classification accuracy. RESULTS: In benchmark evaluation against 25 state-of-the-art predictors, GenixRL achieved an AUROC of 0.9644 on an independent ClinVar dataset. On saturation genome editing assays for BRCA1 and BRCA2, GenixRL achieved the best performance and ranked highest on 14 of 17 clinically significant genes in a zero-shot evaluation. Applied to uncertain and conflicting ClinVar variants, GenixRL enabled tiered, evidence-based prioritization of hundreds of thousands of variants as likely pathogenic or pathogenic with high confidence, supported by orthogonal population evidence from gnomAD. CONCLUSION: GenixRL advances pathogenicity prediction for missense variants and provides an adaptive ensemble that sorts variants of uncertain significance into tiered candidates for expert curation and functional validation.
PURPOSE: To determine whether ATF3 polymorphisms can serve as genetic biomarkers for machine learning-based precision analgesia by establishing a genotype-phenotype association suitable for predictive modeling of postoperative opioid requirements. METHODS: In a prospective cohort of 167 adults undergoing abdominal surgery, ATF3 SNPs rs3122721 and rs3125293 were genotyped. A structured dataset architecture was developed to represent genetic profiles as input features for supervised learning models, enabling translational analysis of genotype‑dependent opioid consumption over 72 h. RESULTS: Patients with homozygous genotypes of the ATF3 SNPs had significantly higher opioid requirements than non‑carriers, despite reporting similar subjective pain scores. This consistent genotype‑dependent pattern provided a clinically relevant phenotype suitable for integration into predictive algorithms. CONCLUSION: ATF3 genotyping offers a promising biomarker for computationally informed precision analgesia. By linking genomic variability to clinically meaningful outcomes within a structured clinical and genomic framework, this approach supports the future development of risk-stratified clinical decision-support systems to optimize postoperative pain management.Trial registration ChiCTR1900021991, registered 30 April 2019. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at https://doi.org/10.1007/s13755-026-00480-9.
BACKGROUND: Substantial improvements in air quality have been recorded following the implementation of China's Clean Air Act (CCAA) in 2013. However, the association between CCAA implementation and individual-level cardiovascular disease (CVD) risk remains unclear. We aimed to examine the long-term association between CCAA implementation and individual-level predicted CVD risk. METHODS: In this prospective, quasi-experimental study, we used data from the China Kadoorie Biobank, a prospective cohort study that recruited participants from five urban and five rural areas across China between 2004 and 2008, with three resurveys conducted after the baseline survey (in 2008, 2013-14, and 2020-21). We included 34 862 individuals (mean age 51·3 years) who participated in at least one resurvey and had no history of CVD at baseline. Participants were classified into intervention (n=25 497) and control (n=9365) groups based on the local government's targets for particulate matter reduction. We estimated the 10-year risk of incident CVD morbidity or mortality using a validated risk prediction model. We used a difference-in-difference model to assess the long-term association between CCAA implementation and predicted risk, with adjustments made for regional confounders and individual-level characteristics, including demographics, lifestyle factors, medical history, and indoor air pollution exposure. The relationship between changes in long-term exposure to PM2·5, PM10, and O3 and predicted risk after CCAA implementation was analysed using a linear model. The estimated risk differences associated with air pollutant changes were estimated based on the magnitude of changes and their corresponding effect sizes. FINDINGS: After the CCAA was implemented, PM2·5 and PM10 concentrations declined in both groups, but O3 concentrations increased. The intervention group showed a 3·95% (95% CI 3·18-4·72%) lower increase in predicted risk than the control group, with larger estimated differences under stricter enforcement. Between 2013 and 2021, each 10 μg/m3 change in PM2·5 concentration was positively associated with a 1·80 (1·34-2·27) percentage point change in predicted CVD risk, whereas each 10 μg/m3 change in PM10 concentration was associated with a 1·24 (0·84-1·63) percentage point change and each 10 μg/m3 change in O3 concentration with a 0·58 (0·33-0·83) percentage point change. Overall, the observed changes in air pollutants during the study period were associated with an average 6·6 percentage point reduction in predicted CVD risk. INTERPRETATION: The CCAA and improved air quality were associated with a slower increase in predicted CVD risk, supporting the necessity for stricter, multipollutant air quality policies to maximise public health benefits. FUNDING: National Natural Science Foundation of China, Kadoorie Charitable Foundation, Noncommunicable Chronic Diseases-National Science and Technology Major Project, National Key R&D Program of China, Chinese Ministry of Science and Technology, and UK Wellcome Trust.