SEARCH · Search PubMed

Results for “Risk prediction model”

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.

SourceResearch connectionFrom yearThrough year

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

1,541 records · Page 3Linked to original sources

PubMed2026

A multi-scale fusion model based on multi-phase contrast-enhanced CT for predicting pancreatic cancer resectability.

Purpose.Develop a multi-scale fusion model (MSFM) based on multi-phase contrast-enhanced computed tomography (CECT) to predict pancreatic cancer (PC) resectability, thereby assisting expert decision-making.Methods.This retrospective study enrolled 280 patients with PC from four institutions, which were randomly divided into a training cohort (202 patients) and an independent test cohort (78 patients). Three-phase CECT images (arterial, venous, and delayed phases) were used for modeling. The MSFM comprises two sub-networks: (1) a multi-phase fusion network for extracting cross-phase shared fusion features, (2) a phase-specific branch network for capturing phase-specific features; and a post-fusion strategy to generate the final predictive score by integrating the shared fusion features and three groups of phase-specific features. Additionally, a human-machine fusion deep learning model (HMfDL) was constructed by fusing the predictive score of the MSFM with expert assessments.Results.In the independent test, the MSFM achieved an AUC (area under the receiver operating characteristic curve) of 0.8385 (95% CI: 0.7521-0.9249), accuracy of 84.62%, sensitivity of 72.00%, and specificity of 90.57%. This performance outperformed single-phase models (AUC range: 0.7638-0.7781), two-phase models (AUC range: 0.7826-0.7864), and ten states-of-the-art classifiers (AUC range: 0.7404-0.7796). The HMfDL further improved the performance, reaching an AUC of 0.8626 (95% CI: 0.7853-0.9400), accuracy of 91.03%, sensitivity of 80.00%, and specificity of 96.23%. Notably, the HMfDL corrected 58.82% of misdiagnosis made by experts.Conclusions. The MSFM effectively fuses multi-phase CECT to enable highly accurate predictions of PC resectability, and provides valuable support for expert decision-making through HMfDL.

Humans↗

PubMed2026

Multimodal alignment improves generalizability of genomic biomarker prediction in computational pathology.

Computational pathology models that use digitized histopathology whole-slide images have the potential to become a cost-effective and scalable alternative to molecular assays for the prediction of genomic biomarkers, a key task in precision oncology. However, as new genomic biomarkers are discovered or quantified, large, labeled datasets must be prospectively collected to train new models. To address this challenge, we developed multimodal alignment for biomarker learning and generalization (MARBLE), a multimodal contrastive pretraining strategy that integrates structured biomarker knowledge into representation learning of histopathology images. MARBLE aligns histopathology-derived representations with representations of genomic biomarkers generated by a large language model (LLM) and a protein language model (PLM). This biologically informed alignment enables data-efficient generalization to novel, out-of-distribution biomarkers. Using the MSK-IMPACT cohort of over 40,000 patients across multiple biomarker panel versions, we design experiments grounded in real-world data to demonstrate the value of our proposed approach.

CP: computational biology↗

PubMed2026

Non-linear predictive modeling and comprehensive meta-analysis of rectal temperature in Santa Inês sheep: a systematic review of thermal challenges and biometerological trends.

A systematic and bibliometric review, combined with a meta-analysis, was used to adjust an equation for estimating the physiological responses of Santa Inês sheep subjected to different thermal challenges. The systematic review compiled data on physiological responses and the thermal environment, which were then used in the meta-analysis to adjust regression models. The bibliometric analysis mapped the relationships among studies, highlighting their usefulness in interpreting research findings and biases. Addressing prior methodological critiques, the core of this study involves replacing the linear approach with a non-linear segmented regression model to accurately define the Thermal Neutral Zone (TNZ). The Segmented Regression Model was crucial, establishing the upper limit of the Thermal Neutral Zone (TNZ) at an air temperature (tair) of 34.64 °C, where trectal begins to increase abruptly. The model, while identifying a biologically significant breakpoint, exhibited a moderate Multiple R-squared of 0.3529, highlighting the high heterogeneity and methodological variability in the current Santa Inês literature. This non-linear approach offers a biologically superior tool for identifying the onset of thermal distress.

Animals↗

PubMed2026

Age at menopause and subjective cognitive symptoms predict digital cognitive outcomes at the gynecological Well-Woman visit.

INTRODUCTION: Women are at increased risk for Alzheimer's Disease (AD). Growing evidence suggests that the menopausal transition may represent a vulnerable window for development of AD-related pathology. Yet, women are diagnosed with AD later than men. Conducting routine cognitive screenings and integrating information about both cognitive symptoms and age at menopause may help address sex-based disparities in detection and prevention. This study investigated whether subjective cognitive symptoms, in combination with age at menopause, were associated with performance on a digital cognitive task in postmenopausal women. METHODS: 183 postmenopausal women (mean age = 63.8, range = 45-85) were recruited after their Well-Woman visit. Participants completed the Screener for Cognitive Problems in Everyday Life (SCoPE) to assess subjective cognitive symptoms, followed by a sensitive measure of objective cognition: the Linus Health Digital Clock and Recall (DCR™). Information was also collected on age at menopause. We examined associations of subjective cognitive symptoms and age at menopause with digital cognitive performance, adjusting for age, education and depression. Model fit was evaluated using adjusted R2, AIC, and BIC. RESULTS: 48.1% of women reported one or more cognitive symptoms on the SCoPE. On objective testing, 73.2% scored in the normal range, 20.8% in the borderline range, and 6.0% in the impaired range. SCoPE total score was negatively associated with objective cognitive performance in adjusted models (B = -.12, p = .03). Age at menopause showed a significant quadratic association with cognitive performance (B = -0.006, p<.001). SCoPE total was not associated with DCR subtests, while age at menopause predicted both Delayed Recall and Clock Drawing. CONCLUSION: Subjective cognitive symptoms and age at menopause were associated with lower performance on a sensitive, objective cognitive test. Findings support routine cognitive screening and suggest that subjective cognitive symptoms as well as age at menopause are associated with cognitive function.

Humans↗

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

Meta-PseU: A meta-classifier for robust prediction of RNA pseudouridine modification sites from long sequences.

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.

Pseudouridine↗

PubMed2026

ADAM10's combined influence on the diagnostic usefulness of IL 22, IL 10, IL-17 A, and IL-17D in autism spectrum disorders: Predicted role on gut leakiness as co-morbidity.

Autism spectrum disorder (ASD) is a complex neurodevelopmental disorder with increasing global prevalence but a lack of reliable diagnostic biomarkers. Emerging evidence suggests that immune dysregulation, gut-brain axis dysfunction, and increased intestinal permeability play key roles in ASD pathophysiology. This study investigated the combined diagnostic value of ADAM10 and cytokines (IL-10, IL-22, IL-17 A, and IL-17D). Multivariable logistic regression produces an improved ROC curve that improves diagnostic accuracy over individual markers by combining numerous predictors into a single risk score (linear predictor). The technique, which frequently raises individual marker AUCs, entails modelling a binary result, calculating the probability, and visualizing ROC based on the projected probabilities. In this case-control study, plasma levels of ADAM10, IL-10, IL-22, IL-17 A, and IL-17D were measured in 37 male children with ASD and 37 age-matched controls. Group comparisons, correlation analyses, and receiver operating characteristic (ROC) curve analyses, including combined ROC models, were performed. ADAM10, IL-22, and IL-17 A levels were significantly reduced in children with ASD compared to controls, whereas IL-10 and IL-17D showed no significant differences. ADAM10, IL-17 A, and IL-22 demonstrated good diagnostic performance, with AUC values of 0.886, 0.855, and 0.812, respectively. In contrast, IL-10 and IL-17D showed poor discriminatory ability, with AUC values of 0.524 and 0.599, respectively. Combined ROC analysis markedly improved diagnostic accuracy, with all panels including ADAM10 achieving AUC values above 0.90, and some reaching as high as 0.988, with high sensitivity and specificity. The combination of ADAM10 with selected cytokines significantly enhances diagnostic performance compared to individual markers, supporting a link between immune dysregulation, barrier dysfunction, and gut permeability in ASD.

Humans↗

PubMed2025

Predictive Validity of Violence Screening Tools in Emergency and Psychiatric Services: A Systematic Review.

Violence against healthcare staff, including a threat or an act of violence toward people during their work, poses a physical and psychological risk to workers internationally. Screening is an important strategy in preventing violence against healthcare professionals. The aim of this systematic review was to synthesize evidence on the predictive validity of risk assessment tools used to screen for violence and aggression risk toward healthcare workers in emergency and psychiatric departments (PD). Primary studies that examined the predictive validity of risk assessment tools for workplace violence were identified via a systematic search of Medline, PsycINFO, Embase, and the Cochrane databases. There were 62 eligible studies, ten of which had a lower risk of bias (RoB). Those studies with high RoB were primarily due to a failure to present calibration measures as part of the analysis. All included studies adopted a longitudinal design and were conducted in PDs. The ten highest-quality studies reported on eight different instruments, four of which showed acceptable to outstanding predictive performance. The Dynamic Appraisal of Situational Aggression and the Brøset Violence Checklist showed the best predictive performance; they were also validated in emergency departments and are best suited for short-term risk prediction. We recommend that the selection of a risk assessment tool should consider the following: (a) the target population, (b) the violence operationalization, and (c) the purpose of the monitoring. We note that the use of a screening tool should be a part of a multicomponent strategy to ensure staff safety.

Humans↗

PubMed2026

Comparison of the predictive performance of systemic immune-inflammation index and neutrophil-to-lymphocyte ratio for three-month poor functional outcome in ischemic stroke: a systematic review and meta-analysis.

INTRODUCTION: Ischemic stroke (IS) is a leading cause of global mortality and disability. Early and accurate prognosis is crucial for patient management. The neutrophil-to-lymphocyte ratio (NLR) and systemic immune-inflammation index (SII) are emerging inflammatory biomarkers; however, their relative predictive value for three-month poor functional outcome (modified Rankin Scale [mRS] > 2) remains uncertain. METHODS: We systematically searched PubMed, Embase, Web of Science, and the Cochrane Library up to 20 July 2025, adhering to PRISMA guidelines. Observational studies reporting the association of SII or NLR with three-month poor outcome were included. Study quality was evaluated using the Newcastle-Ottawa Scale. Area under the curve (AUC), odds ratios (OR), and standardized mean differences (SMD) were pooled using random-effects models in Stata 16.0. RESULTS: Twenty-one studies involving 7520 IS patients were analysed. NLR demonstrated marginally superior discriminative ability compared to SII (AUC 0.71, 95% CI: 0.67-0.76 vs. 0.68, 95% CI: 0.64-0.71), though this difference was not statistically significant. Elevated NLR was significantly associated with poor outcome (OR = 1.26, 95% CI: 1.17-1.37, p < .001), whereas SII was not (OR = 1.00, 95% CI: 1.00-1.00, p = .384). Both markers showed moderate effect sizes (SMD: NLR = 0.69, SII = 0.72; p < .001). NLR performed better in non-intervention and Chinese subgroups, while SII exhibited consistent AUC values across treatment and ethnic subgroups. CONCLUSION: NLR and SII are accessible prognostic markers in IS. NLR demonstrates superior accuracy and a significant association with poor outcome, while SII shows greater stability across patient subgroups. Both may assist in risk stratification, in resource-limited settings.

Humans↗

PubMed2026

AI-enabled viral genomics: from virus discovery to host prediction and emerging variant forecasting.

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.

Artificial Intelligence↗

PubMed2026

Meta-analysis of growth and inactivation kinetics of Legionella.

Quantitative risk assessments intended to inform evidence-based water management plans and public health targets for Legionella in engineered water systems are constrained by fragmented and heterogeneous growth and inactivation kinetics. We conducted a meta-analysis of 25 growth and 39 thermal- and chemical-inactivation studies, fitting microbial persistence models to harmonize parameters. Nonlinear models outperformed first-order formulations, indicating that lag phases and resistant or protected subpopulations are central to Legionella persistence. Random forest analysis identified environmental and methodological drivers of variability based on 226 growth rates and reduction times for thermal (209) and chemical (135) inactivation. Growth was primarily governed by temperature, nutrient availability, and compatible Legionella-host pairings; thermal inactivation by quantification method, temperature, and turbidity; and chemical inactivation by inoculum size, disinfectant type, concentration, and host-associations. Accordingly, temperature-dependent growth parameters and exposure metrics for heat, free-chlorine, and monochloramine, expressed as TT (Temperature×time) and CT (Concentration×time), were derived as condition-specific inputs for predictive models. Growth optima around 37-40 °C, together with lag-time estimates, indicate that hot-water temperature setbacks and energy-saving practices may favor Legionella proliferation under repeated or prolonged lukewarm exposure. Culture- and viability-based TT differences highlight the need to consider viable‑but-non-culturable persistence in monitoring programs. CT comparisons suggest monochloramine may be advantageous because of its lower apparent sensitivity to host-associated protection. Although limited by restricted experimental conditions, the findings show that predictive models should account for microbial ecology, water matrix effects, and quantification endpoints. Future kinetic studies should prioritize realistic multi-host systems, strain pre-adaptation, complementary viability measurements, and standardized protocols and reporting to ensure reproducibility and enable robust system-level predictive modeling.

Legionella↗

PubMed2026

Association of lipoprotein-associated phospholipase A2 with recurrence risk and its predictive value in large artery atherosclerotic stroke.

OBJECTIVE: To investigate the association of lipoprotein-associated phospholipase A2 (Lp-PLA2) with large artery atherosclerotic (LAA) stroke and its predictive value for recurrence. METHODS: We consecutively enrolled 412 acute LAA stroke patients. Using a cutoff of 200 ng/mL, patients were divided into high and low Lp-PLA2 groups, and into recurrence and non‑recurrence groups based on 1‑year follow‑up. Baseline characteristics, lipid profiles, National Institutes of Health Stroke Scale (NIHSS) scores, and vascular stenosis degree were compared. Binary logistic regression and Receiver Operating Characteristic (ROC) analysis were used to identify independent risk factors and evaluate predictive value. RESULTS: The high Lp-PLA2 group had significantly higher low-density lipoprotein cholesterol (LDL-C), small dense low-density lipoprotein cholesterol (sdLDL-C), prevalence of severe stenosis (≥70%), and proportion of NIHSS > 15 (all P < 0.05). The recurrence group showed elevated Lp-PLA2, higher LDL‑C and sdLDL-C, more severe neurological deficits, and more severe stenosis (all P < 0.001). Multivariable regression identified elevated Lp-PLA2 (per 10 ng/mL: OR = 1.139, 95% CI: 1.089-1.191), moderate (OR = 3.145) and severe (OR = 11.663) neurological deficits, and severe stenosis (OR = 9.390) as independent risk factors for recurrence (all P < 0.05). The Area Under the Curve (AUC) of Lp-PLA2 was 0.75 (95% CI: 0.69-0.82), with an optimal cutoff of 208.95 ng/mL. CONCLUSION: Elevated Lp-PLA2 is associated with adverse lipid profiles, more severe neurological deficits, and greater vascular stenosis in LAA stroke patients, and independently predicts 1‑year recurrence. Lp-PLA2 shows moderate predictive value, supporting its potential for risk stratification.

Humans↗

PubMed2026

Assessment of the Potential of Different Anthropometric Indices in Predicting the Risk of Diabetes and Associated Co-morbidities.

Diabetes, a chronic disorder, is showing a rapidly increasing trend globally. India holds the second position in the global diabetes epidemic. The present investigation is an assessment of different anthropometric measurements and their association with type 2 diabetes to determine their diagnostic potential for diabetes as well as its co-morbidities. In this cross-sectional study, we have measured anthropometric parameters and blood biomarkers in subjects with diabetes. We have presented the comparisons of cost- and time-effective anthropometric variable with costly and time-dependent biochemical variables in control and diabetic groups (n = 233/group). Correlations between anthropometric variables and biochemical measurements, as well as the diagnostic utility of anthropometric variables for diabetes, were evaluated. The diagnostic utility of anthropometric variables for diabetes was assessed through receiver operating characteristic (ROC) curves. Neck circumference, sagittal abdominal diameter (SAD), skinfold thickness, and body roundness index (BRI) displayed high specificity and diagnostic utility for diabetes, emphasizing their potential in predicting diabetes and the further development of metabolic syndrome. The study highlights the importance of cost- and time-effective anthropometric assessments in diabetes risk evaluation and calls for further research to elucidate this intricate relationship and develop personalized management strategies.

Humans↗

PubMed2026

Whole-Genome Deep Learning Predicts Chemotherapy Response in Colorectal Cancer.

Chemotherapy response in colorectal cancer (CRC) exhibits significant heterogeneity, with current clinical predictors failing to capture complex genomic determinants of resistance. We developed a hybrid deep learning framework integrating convolutional neural networks (CNNs) and bidirectional long short-term memory (BiLSTM) networks to analyze whole-genome somatic mutations, evolutionary conservation, chromatin accessibility, and 3D genome architecture in 2,546 TCGA patients. An attention mechanism identified predictive genomic regions. The model achieved an AUC of 0.92 (95% CI: 0.89-0.94) in cross-validation and 0.88 (95% CI: 0.85-0.91) in independent validation, outperforming clinical models (ΔAUC = +0.18, p < 0.001). Key predictors included non-coding variants in TP53, KRAS, and PIK3CA regulatory regions. Triple-positive patients (mutations in all 3 regions) had significantly worse progression-free survival (HR = 4.7, p < 0.001). Our framework enables accurate chemotherapy response prediction and reveals novel non-coding resistance mechanisms, advancing precision oncology in CRC.

Humans↗

PubMed2025

Quantitative Outcomes for Shared Assessment and Management in Forensic Mental Health: A Meta-Analysis and Systematic Review.

Despite leading models of mental health care encouraging user involvement, users in forensic mental health (FMH) report poor involvement given the difficulty in reconciling shared approaches with risk-averse and legally mandated settings. While previous research has demonstrated qualitative benefits to shared approaches in FMH and has led to a proliferation of self-rated assessment tools, there remains to quantify agreement on self-rated tools and to clarify the impact of shared approaches on care. This meta-analysis examines (1) the correlation between clinician and user ratings, (2) the predictive validity of self-ratings for violence, and (3) the effects of shared risk management on violence and restriction in FMH. Five databases were searched from inception to April 2024, selecting for adult FMH inpatients, shared risk assessment, needs assessment or violence management as interventions, and quantitative outcomes (correlation, agreement, predictive validity, and effect on violence or restriction rates). Fifteen quantitative evaluations were retained. One of three planned meta-analyses could be conducted, with seven records providing paired clinician-user t-tests. Eleven more records provided clinical recommendations on operationalizing shared approaches. Random-effects meta-analysis showed a significant and large paired standard difference of .95 (95% CI = [.49,1.42]) across tools, with significant differences in DUNDRUM-3, DUNDRUM-4, and CANFOR sub-models. While acknowledging between-study heterogeneity, results substantiate quantitative differences where clinicians generally rate more needs and lesser progress than users across tools, showing that self-ratings can and should be used to broach collaborative discussions on needs and progress during FMH treatment. There remains an evidence gap for quantitative benefits in care outcomes and a need to standardize agreement measures for future comparisons and clinical sub-group analyses.

Humans↗

PubMed2026

Unconfined compressive strength prediction for the ordinary Portland cement-steel slag-silica fume ternary system based on response surface methodology.

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.

Construction Materials↗

PubMed2026

Artificial Intelligence for Diagnosis, Risk Stratification, and Prognosis of Neuroblastoma - A Systematic Review and Meta-Analysis.

PURPOSE: To synthesizes evidence on artificial intelligence (AI) performance in neuroblastoma (NB) diagnosis, risk stratification, prognosis, and genomic characterization. MATERIALS AND METHODS: A systematic review and meta-analysis was conducted following PRISMA 2020 guidelines (PROSPERO: CRD42024539475) across five databases. Meta-analyses used random-effects models with logit-transformed Area Under the Curve (AUCs) and cluster-robust standard errors. AI models were classified as Machine Learning Models (MLM) or Hybrid Nomograms (HN) based on their construction methodology. RESULTS: Of 3,742 articles identified, 53 were included. MLMs demonstrated higher point estimates than radiologists in differential diagnosis (AUC: 0.87 vs. 0.83), though this difference was not statistically significant and carried substantial uncertainty. HNs achieved stronger performance in risk stratification (AUC: 0.87). AI-derived nomograms (AUC: 0.9) and gene signatures (AUC: 0.8) outperformed conventional prognostic markers descriptively. Chemotherapy response prediction remained below clinical utility thresholds across all model types. Only 33.9% of models reported calibration and 24.5% underwent external validation. CONCLUSIONS: AI demonstrates proof-of-concept across multiple NB clinical domains. However, clinical adoption remains premature given persistent gaps in external validation, calibration, dataset size, and pediatric-specific model development. Future studies should test these models prospectively in multicenter pediatric cohorts, ideally through COG or SIOPEN, using shared definitions for diagnosis, risk group, treatment response, and survival outcomes.

Humans↗

PubMed2026

From population to individual: advocating personalised digital tools for heat-health early warning in a changing climate.

Escalating heat extremes under climate change are imposing substantial health burdens, with 2023 and 2024 consecutively breaking global temperature records. Mounting evidence suggests that heatwaves elevate the risks of hospitalisation and mortality across multiple disease categories, including ischaemic heart disease, stroke, chronic obstructive pulmonary disease, and acute kidney injury. Nonetheless, most existing heat-health warning systems remain primarily reliant on population-level predictions, and considering individual differences and disease-specific considerations when defining warning levels would benefit the effectiveness of early prevention for high-risk groups. In this Viewpoint, which is based on the framework of precision public health-delivering the right intervention to the right population at the right time-we propose a framework for personalised digital heat-health early warning tools comprising three dimensions: individualised, risk-stratified prediction models that generate tiered early warnings; personalised health prompts coupled with theory-informed behavioural interventions; and adaptive, equity-oriented alert delivery mechanisms tailored to diverse populations. Such tools have the potential to bridge precision disease prevention and climate adaptation, thereby helping to mitigate heat exposure risks and disease burdens, particularly among high-risk populations. Future implementation research will be essential to address substantial challenges related to feasibility, validation, and equity.

Journal Article↗