SEARCH · Search PubMed

Results for “External beam radiation therapy”

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.

656 records · Page 3Linked to original sources

PubMed2026

Effectiveness of a digi-physical tool and working method for paediatric obesity treatment in Abu Dhabi: a non-inferiority intervention study using an external historical comparator.

BACKGROUND: Effective paediatric obesity treatment requires high intensity, scalable interventions. A digi-physical tool for paediatric obesity treatment has shown positive results in Stockholm, Sweden. This study evaluates whether the same treatment method is effective in a different cultural setting. METHODS: This non-inferiority intervention study, using an external historical comparator, included 60 consecutively recruited children aged 6-15.9 years with obesity who initiated treatment at Sheikh Shakhbout Medical City in Abu Dhabi between June and December 2023. Patients were treated with Evira, a digi-physical tool and working method enabling high intensity individualized care, real-time monitoring, and interactive patient-clinician communication. The primary outcome was BMI z-score change at 26 weeks. Non-inferiority was assessed using a predefined margin of 0.10 BMI z-score, with outcomes compared to a prior published trial in Stockholm (n = 107). RESULTS: A total of 112 children were included in the analysis (Abu Dhabi cohort, n = 35; Stockholm cohort, n = 77). The adjusted mean change in BMI z-score was - 0.20 (95% CI: - 0.28, - 0.12) in the Abu Dhabi cohort and - 0.20 (- 0.26, - 0.14) in the Stockholm cohort (p = 0.88). Non-inferiority was confirmed, (predefined margin 0.10 was not exceeded). A clinically significant BMI z-score reduction (≥ 0.20 units) was achieved by 45.7% of participants in Abu Dhabi and 36.4% in Stockholm (p = 0.35). Non-retention rates at 26 weeks were 41.7% vs. 28.0%, respectively (p = 0.07). CONCLUSIONS: The findings provide promising evidence that treatment outcomes achieved with the digi-physical treatment tool were comparable in the Abu Dhabi and Stockholm cohorts, supporting its feasibility in a second cultural and healthcare setting.

Humans↗

PubMed2026

RR-interval-based atrial fibrillation detection and burden estimation: cross-dataset validation and calibration-aware probability analysis.

Objective.Atrial fibrillation (AF) burden has become an increasingly important endpoint in long-duration rhythm monitoring, but reliable burden estimation requires more than accurate AF detection alone. In particular, when burden is derived by aggregating predicted AF probabilities over time, probability calibration may directly affect burden validity under external dataset shift.Approach.This study developed an interpretable-interval feature model for AF detection and evaluated it using record-wise cross-validation on a development cohort and independent cross-dataset external validation on public Holter electrocardiographic databases. Window-level performance was assessed using the area under the receiver operating characteristic curve (ROC-AUC), area under the precision-recall curve (PR-AUC), Brier score, expected calibration error (ECE), and calibration intercept and calibration slope. Recording-level AF burden was estimated using both probability-based and hard-label aggregation and evaluated using mean absolute error (MAE) and agreement analyses.Main results.The model showed high discrimination in both development and external evaluation, with external ROC-AUC ofand PR-AUC of. However, external calibration deteriorated despite preserved ranking performance, with Brier score of, ECE(15) of, calibration intercept of, and calibration slope of. In the external cohort, probability-based burden estimation preserved strong association with reference burden but showed weaker raw agreement than hard-label aggregation, with MAE ofversus, consistent with systematic probability underprediction. Repeated external recalibration across record-level splits substantially improved probability quality and probability-based burden estimation. Median probability-burden MAE decreased fromwithout recalibration toafter Platt recalibration andafter isotonic recalibration, while median ECE(15) decreased fromtoand, respectively.Significance.These findings indicate that-interval-based AF detection maintained strong ranking performance in the tested external cohort, but probability calibration should be evaluated explicitly when predicted probabilities are aggregated into AF-burden estimates.

Atrial Fibrillation↗

PubMed2026

Advanced/Novel Stenting for Pediatric Dynamic Airway Collapse.

Pediatric dynamic airway collapse is a complex condition that can impact all levels of the pediatric airway. These conditions can pose life threatening risk to pediatric patients and carry lasting impacts. While traditionally, tracheostomy has been used to address all levels of dynamic collapse, recent advances have allowed for more individualized, anatomy-specific stenting and splinting strategies for treatment. This article covers pathophysiology and the latest evidence on strategies to address nasopharyngeal, oropharyngeal, proximal trachea, and tracheobronchial dynamic collapse.

Humans↗

PubMed2026

Data-centric, robust, and explainable multimodal deep learning for clinical decision support: A systematic review.

PURPOSE: Multimodal deep learning is increasingly proposed for clinical decision support (CDS) under a "data-centric" framing that prioritizes label quality, missing-modality robustness, distribution shift, calibration, and explainability. Prior reviews have examined multimodal medical AI, CDS, and data-centric methods separately, but none address their intersection. We mapped the modalities, fusion strategies, and data-centric and explainability techniques used in this recent literature, quantified how often each is implemented rather than merely mentioned, assessed deployment-relevant evidence (external validation, clinical-outcome measurement, equity), and formally appraised study-level risk of bias. METHODS: Following the PRISMA 2020 statement (PROSPERO CRD420261427815; registered retrospectively), we screened 150 records and included primary, clinical, multimodal studies that applied machine or deep learning to a decision-support task and reported at least one quantitative result. Two reviewers screened and extracted data with consensus adjudication. Each study was coded against pre-specified operational definitions, separating implemented or empirically evaluated techniques from those only mentioned. Study-level risk of bias was assessed with PROBAST + AI. Synthesis was narrative. RESULTS: Thirty-one studies met inclusion; 30 (97%) were published between 2024 and 2026, with a median of three modalities (range 2-6), most commonly structured EHR (71%) and imaging (39%). Data-centric techniques were frequently reported (74-84% across label-noise, distribution-shift, calibration, missing-modality and class-imbalance handling; equity 61%). However, external validation was reported in only 4/31 studies (13%), a clinical or provider outcome in 3/31 (10%), and no study reported routine deployment. Overall risk of bias was high in 27/31 studies (87%), driven by the analysis domain. CONCLUSION: Within this recent, self-selected slice of the field, technical robustness and explainability techniques are widely reported but rarely validated out-of-distribution or against clinical outcomes, and the underlying evidence is at high risk of bias. Progress requires external multi-site validation, clinical-outcome measurement, formal bias appraisal, and adherence to AI reporting standards (e.g., TRIPOD + AI) before deployment can be justified.

Deep Learning↗

PubMed2026

Diagnostic performance of machine learning models for malignant and non-malignant pleural effusion: Systematic review and meta-analysis.

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.

Humans↗

PubMed2025

Online Risk Behavior in Adolescents: A Systematic Review.

Identifying and categorizing online risk behaviors is crucial for assessing their impact on adolescents. Despite extensive research, previous studies have not provided a clear classification of these behaviors. This systematic review synthesizes the quantitative literature on adolescent online risk behaviors from the inception of research to September 2023, aiming to: (a) offer a comprehensive overview of the types of online risk behaviors and the specific actions encompassed within each category among adolescents; (b) summarize the adverse outcomes associated with these behaviors; and (c) discuss the implications and future research directions. Utilizing key terms, this study sourced studies from four electronic databases (Scopus, PubMed, Web of Science, and EMBASE), ultimately including 22 English-language quantitative studies. The review reveals that online risk behaviors are primarily categorized into content risk behaviors, contact risk behaviors, and conduct risk behaviors. Adolescents engaging in these behaviors are at an increased risk of experiencing physical health issues, mental health problems, externalizing behaviors, and even self-harm and suicidal thoughts or actions. Further research is needed to develop and validate an online risk behavior scale and conduct longitudinal and experimental studies to establish causal relationships and examine the long-term effects of these behaviors on adolescent well-being. The review concludes with implications for future research and potential prevention, intervention, and policy strategies to mitigate online risk behaviors in adolescents.

Humans↗

PubMed2026

Effectiveness of Internet-Delivered Cognitive Behavioural Therapy (ICBT) in Improving Weight-Loss and Psychosocial Well-Being Among Adults With High BMI: A Systematic Review.

AIM: To examine the effectiveness of Internet-delivered Cognitive Behaviour Therapy (ICBT) in improving psychosocial well-being and promoting weight-loss in adults ≥ 18 years with BMI ≥ 25 kg/m2. BACKGROUND: Global obesity engenders significant physical and psychosocial health consequences. Second-wave ICBT focused on restructuring negative thoughts and behaviours has been explored as a potential intervention for elevated BMI and mental health concerns, but its effectiveness remains to be fully established, making further evaluation essential. METHODS AND DATA SOURCES: Eight databases were searched from inception to January 2025 for randomised controlled trials (RCTs), including participants ≥ 18 years with BMI ≥ 25 kg/m2 and second-wave ICBT evaluating BMI, weight, depression, eating behaviours, and self-esteem. This review followed PRISMA 2020. Study quality was assessed using the Cochrane Risk of Bias (ROB 2) and GRADE. Data were extracted using a modified Cochrane form. Random-effects meta-analysis calculated Standardised Mean Differences (SMD) with 95% confidence intervals, with subgroup analyses exploring heterogeneity. RESULTS: Nine trials with 2278 participants were included. Significant improvements were seen in BMI, weight, and depressive symptoms while self-esteem effects were small and non-significant. Compared with passive controls, ICBT showed greater improvements in BMI and weight, whereas differences versus active control were smaller and inconsistent. Face-to-face CBT demonstrated superior outcomes for depression and self-esteem. Male-tailored interventions showed greater improvements. Shorter programmes yielded larger short-term weight loss, while longer programmes supported more sustained effects. Narrative synthesis indicated improvements in emotional and external eating, with increased mindful and restrictive eating behaviours. CONCLUSION: ICBT improved weight, BMI, and depressive symptoms, with limited evidence for self-esteem. Male-tailored interventions and longer programmes may enhance sustainable outcomes. IMPACT: Future ICBT programs should integrate strategies targeting sustainable weight loss and psychosocial well-being to support long-term outcomes. NO PATIENT OR PUBLIC CONTRIBUTION: Patients or members of the public were not involved, as this study synthesised previously published data. TRIAL REGISTRATION: PROSPERO registration number: CRD42024497961.

Humans↗

PubMed2026

Chemoradiotherapy versus short-course radiotherapy for response-adapted organ preservation in early-stage and intermediate-stage rectal cancer (STAR-TREC): 12-month results of an international, multicentre, open-label, parallel-group, randomised, phase 2/3 trial.

BACKGROUND: Total mesorectal excision (TME) is the standard treatment for most early-stage and intermediate-stage rectal cancer but can cause substantial perioperative morbidity, functional impairment, and reduced quality of life. We assessed whether long-course chemoradiotherapy (LCCRT) or short-course radiotherapy (SCRT) could increase organ preservation and reduce surgery, toxicity, and quality-of-life harms without compromising oncological outcomes. METHODS: STAR-TREC is an international, multicentre, open-label, parallel-group, randomised, phase 2/3 trial in five European countries. Eligible patients were aged 16 years or older in the UK or aged 18 years or older elsewhere, had an Eastern Cooperative Oncology Group (ECOG) performance status of 0-1, and rectal adenocarcinoma (≤40 mm staged as mrT1-T3bN0). In phase 2, participants were randomly assigned (1:1:1) to LCCRT-based organ preservation (LCCRT-OP; 50 Gy in 25 fractions plus oral capecitabine 825 mg/m2 twice daily), SCRT-based organ preservation (SCRT-OP; 25 Gy in five fractions), or primary TME. Phase 2 assessed feasibility, with recruitment at months 12 and 24 as the primary endpoint and feasibility thresholds of four or more and six or more randomisations per month, respectively. Phase 3 adopted a partially randomised patient-preference design, allowing participants to choose either organ preservation or TME. Participants that chose organ preservation were randomly assigned (1:1) to receive LCCRT-OP or SCRT-OP using centralised, computer-generated assignment, with stratification by country and MRI T category (≤T3a vs T3b) using minimisation. The phase 3 primary endpoint was organ-preservation 30 months after treatment initiation, defined as absence of TME, stoma, or local recurrence, which was assessed in the modified intention-to-treat population, which included participants in phase 2 and phase 3. After a planned interim analysis of unmasked phase 2 data, the trial steering committee and independent data monitoring committee recommended reporting a 12-month, modified intention-to-treat analysis of implementation outcomes for participants recruited before Aug 8, 2023. This study is registered with ISRCTN (14240288) and is closed. FINDINGS: Between June 14, 2017, and April 8, 2024, 503 participants were enrolled at 37 sites. Phase 2 enrolled 120 participants, with recruitment rates of three and six participants per month at months 12 and 24, respectively. Overall, 12-month TME-free survival was 60% (47 of 78 participants). After phase 3 recruitment ended, interim analysis of unmasked phase 2 data showed an early TME-free survival benefit with LCCRT versus SCRT (12-month median TME-free survival not reached [95% CI not reached-not reached] vs 7·6 months [95% CI 6·4-not reached]; hazard ratio [HR] 3·7 [95% CI 1·7-8·0]; posterior probability of superiority >99·5%). The trial steering committee and independent data monitoring committee therefore recommended expanded analysis of 426 participants recruited before Aug 8, 2023: 120 from phase 2 and 306 from phase 3. 17 participants withdrew before treatment, leaving 409 in the modified intention-to-treat population: 163 allocated to LCCRT, 168 to SCRT, and 78 to primary TME. 116 (28%) participants were female and 293 (72%) were male. Among participants who opted for organ preservation, 12-month TME-free survival was 78·5% (95% CI 72·4-85·1) with LCCRT and 60·6% (53·6-68·4) with SCRT (HR 1·90 [95% CI 1·29-2·81]). The most common grade 3-4 serious adverse events were gastrointestinal disorders (four [2%] with LCCRT vs six [4%] with SCRT vs six [8%] with TME) and procedural complications (three [2%] with LCCRT vs five [3%] with SCRT vs five [6%] with TME). One participant allocated to primary TME died after an anastomotic leak. INTERPRETATION: These early results support a response-adapted organ-preservation approach, with LCCRT appearing more effective than SCRT at 12 months. Organ-preservation might also reduce treatment-related toxicity compared with primary TME. Longer follow-up is needed for the prespecified 30-month endpoint and definitive functional and oncological outcomes. FUNDING: Cancer Research UK, Stand Up to Cancer, Dutch Cancer Society, Danish Cancer Society, Kom Op Tegen Kanker, Cancerfonden, ALF Region Stockholm, RCC Region Stockholm.

Humans↗

PubMed2026

Can't see the forest for the trees: The influence of marker type on inferred phylogenetic relationships in a cosmopolitan bat genus.

Fine-resolution information on species relationships and biological diversity is critically needed to guide conservation efforts amidst rapid environmental changes. Systematics, which forms the foundation of this knowledge, has been revolutionized by phylogenomics, utilizing genome-scale datasets. However, the use of diverse marker types, non-comparable taxon sampling, and outgroup selection can lead to conflicting phylogenetic hypotheses. These inconsistencies complicate study comparisons and hinder our ability to assess marker-specific impacts on phylogenetic resolution. The phylogenetic reconstruction of the bat genus Myotis, encompassing over 140 species and characterized by a rapid radiation in the last 20 million years, has been particularly influenced by these challenges. Achieving phylogenetic resolution in Myotis is particularly complex due to subtle interspecific differences in both morphological and molecular traits. Mitochondrial and nuclear markers often produce discordant trees, influenced by hybridization, introgression, and methodological variations. In this study, we employed a consistent taxonomic sample set of 44 Myotis taxa to evaluate the impact of five different genetic marker types on phylogenetic reconstruction. We observed significant discordance between topologies derived from conserved nuclear and mitochondrial markers and found that transposable elements were inadequate for resolving relationships across the entire genus. Our results also clarify the placement of previously problematic taxa within the genus. These findings emphasize the importance of aligning genetic marker choice with specific phylogenetic questions and highlight the influence of taxonomic and methodological variation on phylogenomic outcomes. This work provides a framework for improving phylogenetic inference in rapidly radiating groups and enhances our understanding of evolutionary history in Myotis.

Animals↗

PubMed2026

Multimodal intervention benefits: Responder analysis of J-MINT PRIME Kanagawa trial.

INTRODUCTION: The J-MINT PRIME Kanagawa trial was an 18-month multimodal intervention (incorporating exercise, nutrition, and metabolic management) for dementia prevention. Because the primary analysis showed no significant benefits, we performed an exploratory responder analysis to identify responsive subpopulations. METHODS: We analyzed the Full Analysis Set comprising 188 participants. Classification and regression tree (CART) analysis, applied to the intervention arm, identified baseline predictors of cognitive improvement. These rules were then applied to the entire cohort to evaluate treatment effects on the Mini-Mental State Examination (MMSE) using fully adjusted mixed-effects models for repeated measures (MMRM). RESULTS: CART identified a "Target Group" (N = 108) characterized by baseline profiles such as an MMSE score < 28 or specific metabolic ranges (e.g., LDL-C < 135 mg/dL). Within this target group, the intervention significantly preserved MMSE trajectories compared with the control group (group × time interaction, P = 0.022). In contrast, the Non-Target Group (N = 80), consisting of high-functioning individuals (MMSE ≥ 28), exhibited no significant group × time interaction. DISCUSSION: Multimodal interventions may effectively preserve global cognition in older adults with sub-threshold cognitive decline. Careful targeting of appropriate populations, while considering potential longitudinal measurement artifacts (e.g., practice effects), is essential. These findings provide a hypothesis-generating framework that warrants external validation in future prevention trials.

Humans↗

PubMed2026

Beyond predictive performance: A systematic review and critical methodological appraisal of AI/ML and conventional modelling strategies in breast, colorectal, and pancreatic Cancer.

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.

Humans↗

PubMed2026

Radiographic assessment and orthodontic intervention effects on orthodontically induced root resorption: a systematic review and meta-analysis of clinical trials.

The relative contribution of radiographic methods and characteristics of the orthodontic intervention to orthodontically induced root resorption (OIRR) remains unknown. The aims of this systematic review and meta-analysis were to (1) estimate the pooled OIRR effect across orthodontic intervention versus comparator contrasts, (2) compare pooled estimates by radiographic method (2D [two-dimensional] vs. 3D/CBCT [three-dimensional/cone-beam computed tomography]), and (3) explore whether force mechanics (intrusive versus nonintrusive) modified OIRR magnitude. Seven randomized controlled trials and one prospective study (January 2010-October 2025) were included. Only OIRR was the outcome, reported as correlation coefficients (r). The primary analysis combined within-study intervention-versus-comparator estimates. Subgroup analysis of 2D versus 3D/CBCT imaging was prespecified, whereas post-hoc analysis of intrusive versus nonintrusive mechanics was performed. The pooled analysis for the primary outcome showed a small, nonsignificant OIRR effect (r = 0.07; 95% confidence interval [CI]: -0.12 to 0.27; p = 0.372) with high heterogeneity (I2 = 84.0%). Radiographic method did not change the pooled estimates significantly (p = 0.331). Force-mechanics analysis showed that intrusive mechanics was related to significantly higher root resorption than nonintrusive mechanics (r = 0.40; 95% CI = 0.15 to 0.65 versus r = -0.03; 95% CI = -0.16 to 0.10; p < 0.001). This accounted for 87.1% of the between-study variance. The average orthodontic intervention effect on OIRR was small and not significant; however, the OIRR magnitude was strongly affected by force mechanics, particularly by intrusive forces. There was no significant difference in pooled estimates by radiographic method; however, 3D/CBCT provides superior volumetric quantification and should be used judiciously according ALARA (as low as reasonably achievable) principles.

Root Resorption↗

PubMed2026

Ventriculostomy-Related Infections by Country-Income Level: A Systematic Review and Bayesian Hierarchical Meta-analysis.

Our objective was to perform a systematic review and meta-analysis of published literature on ventriculostomy-related infection (VRI) and evaluate temporal and global trends. We conducted a systematic review and Bayesian hierarchical random-effects meta-analysis of VRI rates in adults, stratified by country-income level (high-income countries [HIC]; low- or middle-income countries [LMIC]), study design, sample size, enrollment period, VRI intervention, and VRI definition. We identified 159 articles published between 1989 and 2025 that included 523,704 patients with 7293 VRIs. The pooled VRI rate was 8.64% [95% CI: 7.44-9.97], with moderate heterogeneity and good model fit. The leave-one-out sensitivity analysis showed a mean absolute change of 0.06% and a maximum change of 0.2%, indicating robust analysis. Five of the 33 represented countries had VRI rates below the global pooled rate of 8.64%. Four were HICs: Singapore (VRI rate 3.3% [0.8-7]), the United States (VRI rate 4.6% [3.4-5.9]), Germany (VRI rate 6.1% [1.1-18.9]), Norway (8.3% [0.3-68.4]), with 1 LMIC: China (8.5% [5.4-12.4]). VRI was significantly higher in studies using definitions beyond CSF culture alone for VRI (+3.16% [0.11- 6.52]) and in those from Europe (+7.29% [4.62-10.10]) and the Western Pacific (+4.09% [1.55-6.98]). No other subgroup demonstrated significant differences. This Bayesian meta-analysis provides global estimates and factors associated with VRI. Standardization of VRI definitions is critical for future benchmarking of VRI rates.

Humans↗

PubMed2026

The role of artificial intelligence in the diagnosis and prognosis of traumatic brain injury based on brain CT scans: a systematic review.

Traumatic brain injury (TBI) is a leading cause of emergency department visits and a major contributor to injury-related mortality and long-term neurological disability. Non-contrast computed tomography (CT) is the gold-standard imaging modality for the rapid diagnosis of TBI. Clinical outcomes depend strongly on early detection and prompt acute management. Artificial intelligence (AI)-based models may support faster automated identification of traumatic findings and early prediction of patient prognosis. A systematic literature search was conducted in PubMed/MEDLINE, Scopus, IEEE Xplore, ACM Digital Library, and the Cochrane Library in accordance with PRISMA 2020 guidelines to evaluate AI-based models for automated detection of TBI-related findings on CT and for prediction of clinical outcomes. Risk of bias and applicability were assessed using QUADAS-2 for diagnostic accuracy studies and PROBAST + AI for prediction model studies. Twenty-two studies were included. Sixteen studies evaluated diagnostic tasks and 10 evaluated prognostic outcomes, with four studies contributing to both categories. Diagnostic performance was generally high, with many studies reporting AUC values approaching or exceeding 0.90, particularly for larger lesion volumes.Prognostic performance was more variable, with moderate to high discrimination and substantial heterogeneity. Only 9 studies incorporated independent external validation, and performance was frequently lower in external cohorts. All prognostic model studies were judged to be at high overall risk of bias using PROBAST + AI, and most diagnostic accuracy studies also demonstrated high or unclear risk of bias in at least one QUADAS-2 domain, most frequently in patient selection. AI-based models applied to brain CT demonstrate strong technical performance for both diagnostic and prognostic tasks in TBI. However, most studies relied on retrospective designs and lacked independent external validation which limits models generalizability and raises concern for potential overfitting. Prospective, multicenter studies with standardized methodologies and rigorous external validation are required before widespread clinical implementation.

Humans↗

PubMed2026

Quo vadis, BGA? A collaborative EDNAP exercise on the challenges and progress in forensic biogeographical ancestry inference.

There is a broad consensus that forensic tests for the prediction of externally visible characteristics (EVC) and analysis of biogeographic ancestry (BGA) of an individual are technically reliable. However, interpretation of the results and population-specific genotype distribution patterns remains challenging. EVC and BGA analyses provide valuable information for population genetics studies and as investigative leads for criminal cases, as well as for historical and contemporary identification tests. However, inaccurate or incorrect predictions, for example, from subjective bias in the interpretations made, have the potential to misdirect police investigations. The legal situation regarding EVC and BGA testing varies by country: ranging from countries where it is explicitly prohibited, to those without specific regulations on biogeographic ancestry prediction, and others that have already enacted laws governing its use. The reluctance to utilize these analyses is not only due to legal restrictions and data protection concerns, but also to initial limited sets of sufficiently comprehensive forensic DNA assays. Forensic BGA marker panels typically contain up to ∼300 SNPs. This relatively small number of genetic markers, along with limited reference population data, complicates the interpretation of results from donors of unknown origin. This paper presents the results of a collaborative EDNAP study, which, for the first time, evaluated the approach to reporting EVC and BGA data between international laboratories. For the study, DNA from nine individuals with self-reported ancestry was collected and analysed using various forensic panels differing in the number and composition of ancestry-informative markers genotyped, comprising: the Precision ID mtDNA Whole Genome Panel, the VISAGE Basic Tool and the VISAGE Enhanced Tool for Appearance and Ancestry Prediction, and the Ion AmpliSeq™ PhenoTrivium Panel. To ensure full data protection, all SNP genotypes and uniparental marker haplotypes obtained were not shared with third parties. Instead, the genetic data were analysed using a range of commonly used population analysis software packages. These analysis outcomes were then distributed to twelve European forensic laboratories (both academic and law enforcement institutions), who were asked to prepare reports based on their interpretation of the phenotypes and ancestry they inferred from the analysis data. A questionnaire sent alongside the genetic information, aimed to evaluate which difficulties were encountered by the participants in processing the BGA analysis data they were given.

Humans↗

PubMed2026

Diagnostic performance of machine learning models versus established risk stratification for intracranial aneurysm rupture: a systematic review and bivariate meta-analysis.

BACKGROUND: Machine learning (ML) models have been proposed to improve the discrimination of intracranial aneurysm rupture status beyond established clinical risk stratification tools. However, reported performance is heterogeneous and the relative contribution of model architecture and feature dominance remains unclear. METHODS: We performed a Preferred Reporting Items for Systematic Reviews and Meta-Analyses-diagnostic test accuracy systematic review and diagnostic meta-analysis of studies evaluating ML models for intracranial aneurysm rupture discrimination. PubMed, Embase and CENTRAL were searched to February 2026. Sensitivity and specificity were pooled using a bivariate random-effects model, with summary receiver operating characteristic curves generated across training, internal testing and external validation datasets. Models were compared with regression-based approaches and Population, Hypertension, Age, Size of aneurysm, Earlier subarachnoid haemorrhage, Site of aneurysm (PHASES) scores. Subgroup and meta-regression analyses explored associations between algorithm family and feature domain. RESULTS: Sixty-two retrospective cohorts (29 709 patients 209 models) met the inclusion criteria. In training datasets, pooled sensitivity and specificity for ML were 0.81 (95% CI 0.75 to 0.85) and 0.83 (0.80-0.86), with an area under the curve (AUC) of 0.878, exceeding PHASES (AUC 0.667). In testing datasets, ML retained higher discrimination (AUC 0.837) than regression models (0.806) and PHASES (0.646). In external validation, sensitivity was preserved (0.82), but specificity declined (0.66). Deep learning demonstrated the highest AUCs (training and testing). Incorporation of haemodynamic or radiomic features improved pooled discrimination relative to morphology alone. Evidence of small-study effects and mostly unclear Prediction Model Risk Of Bias Assessment Tool ratings were observed. CONCLUSIONS: ML approaches demonstrate higher pooled discrimination for aneurysm rupture status than conventional risk scores in retrospective datasets, but reduced external validation specificity and heterogeneity limit confidence for clinical translation. Prospective, externally validated, calibrated models are required before integration into routine cerebrovascular risk stratification.

Humans↗

PubMed2026

Cognitive Behavioral Therapy vs Media Literacy for Prevention of Gaming Disorder and Unspecified Internet Use Disorder: A Randomized Clinical Trial.

IMPORTANCE: Mitigating the harmful effects of digitalization on youth mental health is essential. OBJECTIVES: To evaluate the effectiveness of cognitive behavioral therapy (CBT) compared with a media literacy-based intervention for gaming disorder and unspecified internet use disorder and to determine whether universal or indicated prevention yields stronger benefits. DESIGN, SETTING, AND PARTICIPANTS: PROTECTconfirm was a randomized clinical trial conducted from July 1, 2020, to August 31, 2024, across 44 secondary schools in Baden-Württemberg, Germany, with 1-, 4-, and 12-month follow-up. All students were included in universal prevention analyses, whereas indicated prevention analyses included a subgroup of adolescents at high risk meeting 2 or more Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) criteria for gaming disorder or unspecified internet use disorder at baseline. INTERVENTIONS: Students were individually randomized within 90 classes to PROTECTtraining, a CBT-based program; or PROTECTinfo, a structurally parallel media literacy program. Both consisted of four 90-minute sessions delivered weekly during school hours by trained local prevention professionals under live supervision and adherence protocols. MAIN OUTCOMES AND MEASURES: The primary outcome was gaming disorder symptom severity and unspecified internet use disorder symptom severity, assessed at 12 months with a modified version of the Video Game Dependency Scale. Secondary outcomes included internalizing, externalizing, and transdiagnostic symptoms. Primary and secondary outcomes were analyzed according to the intention-to-treat principle. RESULTS: A total of 1793 students were randomized to PROTECTtraining (n = 896; mean [SD] age, 13.1 [1.5] years; 498 male students [56.1%]) or PROTECTinfo (n = 897; mean [SD] age, 13.1 [1.5] years; 484 male students [54.8%]). Participants in the subsample of 205 adolescents at high risk showed significantly lower 12-month severity of gaming disorder and unspecified internet use disorder symptoms with PROTECTtraining than with PROTECTinfo (mean [SD] Modified Video Game Dependency Scale score: 14.7 [9.7] vs 17.7 [10.3]; within-group Cohen d = -0.91 [95% CI, -1.10 to -0.72] vs Cohen d = -0.61 [95% CI, -0.80 to -0.42]; between-group Cohen d = -0.30 [95% CI, -0.55 to -0.05]), but not in the total sample (mean [SD] Modified Video Game Dependency Scale score: PROTECTtraining, 8.8 [8.3] vs PROTECTinfo, 9.4 [8.4]; within-group Cohen d = -0.11 [95% CI, -0.18 to -0.05] vs Cohen d = -0.04 [95% CI, -0.11 to 0.02]; between-group Cohen d = -0.07 [95% CI, -0.16 to 0.01]). Secondary outcomes did not differ significantly between groups. CONCLUSIONS AND RELEVANCE: In this randomized clinical trial of 1793 adolescents, the CBT-based intervention reduced gaming disorder symptoms and unspecified internet use disorder symptoms more effectively than the media literacy-based intervention among adolescents at high risk. These findings support the use of CBT-based approaches primarily within indicated, rather than universal, prevention. TRIAL REGISTRATION: German Clinical Trials Register (DRKS) trial registration: DRKS00033989.

Humans↗

PubMed2026

Artificial Intelligence for Diagnosing Meibomian Gland Dysfunction: A Systematic Review and Meta-Analysis of Diagnostic Test Accuracy Studies.

PURPOSE: To identify, appraise, and synthesize the performance of artificial intelligence-based meibography reading as compared with human graders in diagnosing meibomian gland dysfunction. METHODS: We followed Cochrane methodology and reporting guidelines for diagnostic test accuracy reviews. To assess potential risk of bias and applicability, we used a modified Quality Assessment of Diagnostic Accuracy Studies-2 checklist. We applied bivariate logistic models to estimate summary sensitivity and specificity when appropriate and used the GRADE framework to rate the certainty of the evidence. RESULTS: We identified 14 eligible studies involving 5511 predominantly middle-aged participants (average age: 27-55 years) who were primarily female (≥54.5%). A total of 18,926 meibography images were obtained through noncontact infrared (11 studies) or in vivo confocal microscopy (three studies). Two studies reported external validation of deep learning models, 12 reported internally validated models, and one reported both. All but one study had high risk of bias in at least one domain; 12 studies raised high or intermediate concern about applicability. Based on three external evaluations, the summary sensitivity and specificity for diagnosing meibomian gland dysfunction from normal glands were 97.5% (95% confidence interval: 77.5%-99.8%) and 85.5% (95% confidence interval: 47.3%-97.5%). Sources of heterogeneity in internally validated models included study population, case mix, and others. The overall evidence was very low to low certainty because of imprecision, high risk of bias, and concerns about applicability. CONCLUSIONS: Artificial intelligence-based meibography grading appears less accurate than human graders. Future studies should adopt rigorous designs, including a more diverse participant pool (or image set), and external validation.

Humans↗