Tag: pubmed

Diabetes Obes Metab. 2026 Aug 25. doi: 10.1111/dom.71260. Online ahead of print.

ABSTRACT

AIMS: We aimed to examine whether offspring exposed to maternal gestational diabetes mellitus (GDM) had different fat-lean balance trajectories from unexposed offspring and whether the estimated associations were attenuated after adjustment for maternal pre-pregnancy adiposity.

METHODS: Using UK Millennium Cohort Study data, we analysed body composition outcomes, including fat mass (FM), body fat percentage (BF%), fat mass index (FMI), fat-free mass index (FFMI), and the FMI/FFMI ratio. Linear mixed-effects models characterised body composition trajectories from ages 7-17 years, with additional adjustment to assess the attenuation of the estimates after accounting for maternal pre-pregnancy adiposity.

RESULTS: The analysis included 14 329 offspring, of whom 170 were exposed to maternal GDM. In the primary adjusted model, GDM-exposed offspring had higher model-predicted values for all body composition indices than unexposed offspring across the observed ages. GDM-by-age interaction terms indicated age-varying between-group differences for FM, FMI, and the FMI/FFMI ratio (all p < 0.05), with borderline evidence for BF% (p = 0.050) and no clear evidence for FFMI. After additional adjustment for maternal pre-pregnancy adiposity, the estimated between-group differences were attenuated, and some age-specific contrasts, particularly at age 7 years, were no longer statistically significant. Nevertheless, age-varying between-group patterns remained evident for several body composition indices.

CONCLUSION: Maternal GDM exposure was associated with age-varying differences in offspring fat mass and fat-lean balance during childhood and adolescence, and these associations were attenuated after adjustment for maternal pre-pregnancy adiposity.

PMID:42642339 | DOI:10.1111/dom.71260

J Matern Fetal Neonatal Med. 2026 Dec;39(1):2721085. doi: 10.1080/14767058.2026.2721085. Epub 2026 Aug 25.

ABSTRACT

BACKGROUND: Maternal and neonatal diseases pose major global public health challenges with disproportionate burdens across socioeconomic settings. This study systematically evaluates global epidemiological trends and future projections of these diseases.

METHODS: Using Global Burden of Disease (GBD) data from 1990 to 2021 across 204 countries, we analyzed age-standardized prevalence, incidence, mortality, and disability-adjusted life years (DALY) rates. Stratification was based on the Socio-Demographic Index (SDI). Medium-term trends (2022-2031) were forecasted using ARIMA models.

RESULTS: Globally, age-standardized prevalence increased (EAPC = 0.48), while incidence, mortality, and DALY rates declined significantly (EAPC: -0.97, -1.64, and -1.42, respectively). Low SDI regions, particularly sub-Saharan Africa, bore the heaviest mortality and DALY burdens-exceeding high SDI regions by more than tenfold. South Asia showed the highest prevalence. Males under 5 years and females aged 20-34 were identified as high-risk groups. Forecasts indicate continued prevalence rise and substantial mortality reduction by 2031, with persistent inequalities.

CONCLUSION: Despite global progress, marked inequalities persist. Differentiated health policies, strengthened primary care in low-resource settings, and multi-sector collaboration are urgently needed to reduce maternal and neonatal health inequities.

PMID:42642335 | DOI:10.1080/14767058.2026.2721085

J Matern Fetal Neonatal Med. 2026 Dec;39(1):2713330. doi: 10.1080/14767058.2026.2713330. Epub 2026 Aug 25.

ABSTRACT

OBJECTIVE: International guidelines recommend decision-to-delivery intervals (DDIs) of ≤30 minutes (min) for emergency cesarean deliveries (ECDs), whereas the German Society recommends ≤20 min. However, the feasibility and relevance of these thresholds vary across and within healthcare systems. This study assessed the association between DDI and perinatal outcomes in ECDs performed during and outside regular working hours at a tertiary obstetric care center in Turkey.

METHODS: This single-center, retrospective study included 835 ECDs with complete medical records out of 11,014 deliveries conducted between January 2021 and December 2024 at the Department of Obstetrics and Gynecology, Umraniye Training and Research Hospital, Istanbul, Turkey. Cases were categorized according to timing (working vs. outside working hours) and DDI duration (≤15, 16-30, 31-45, 46-60, and ≥61 min). Demographic data and perinatal outcomes were analyzed across these groups. Subgroup analyses were also performed according to gestational age at birth and type of anesthesia.

RESULTS: The overall mean DDI was 24.4 ± 12.0 min, with a median of 22 (8-68) min. 76.2% of ECDs had a DDI ≤30 min. Mean DDI was similar between working hours (24.2 min) and outside working hours (24.4 min) (p = 0.892). While the median 1- and 5-minute Apgar scores were comparable between groups (p = 0.236 and p = 0.645, respectively), neonatal intensive care unit (NICU) admission was higher during working hours (p = 0.021). The lowest Apgar scores were observed in the ≤15-minute DDI group, particularly in term pregnancies, whereas no consistent association was seen in preterm cases. In early preterm deliveries and cases under general anesthesia, prolonged DDI was associated with intrapartum fetal death. In multivariable analysis, gestational age was the only independent predictor of NICU admission, while in a Firth-penalized model, gestational age and general anesthesia were independently associated with a 5-minute Apgar score <7; DDI was not independently associated with adverse neonatal outcomes in either model.

CONCLUSIONS: In this tertiary obstetric care center, the recommended 30-minute DDI target was achieved in the majority of ECDs. Although lower Apgar scores were observed in cases with the shortest DDIs, prolonged DDI was not independently associated with adverse neonatal outcomes after adjustment for potential confounders. Gestational age was the main determinant of neonatal outcome, whereas DDI showed no independent effect. Nevertheless, early preterm deliveries and cases performed under general anesthesia may represent high-risk subgroups in which prolonged DDI is associated with adverse fetal outcomes.

PMID:42642333 | DOI:10.1080/14767058.2026.2713330

Acad Radiol. 2026 Aug 25:S1076-6332(26)00618-5. doi: 10.1016/j.acra.2026.08.028. Online ahead of print.

ABSTRACT

RATIONALE AND OBJECTIVES: To address the challenge of preoperative prediction of synchronous liver metastasis (LM) in pancreatic cancer (PC), we developed and validated machine learning models integrating clinical and computed tomography (CT) radiomics features, and compared the performance and interpretability of linear (linear discriminant analysis [LDA]) versus nonlinear (multilayer perceptron [MLP]) architectures.

MATERIALS AND METHODS: This retrospective study enrolled 340 patients with pancreatic ductal adenocarcinoma (190 with LM, 150 without). Five radiomics features were selected using the minimum redundancy maximum relevance and least absolute shrinkage and selection operator methods. Based on independent clinical predictors (age, carcinoembryonic antigen [CEA], and clinical N stage) and the selected radiomics features, 26 machine learning models were constructed. The optimal models (LDA and MLP) were evaluated in an independent validation cohort (n = 102) using receiver operating characteristic curves, calibration, decision curve analysis, and SHAP (SHapley Additive exPlanations) analysis.

RESULTS: In the validation cohort, LDA achieved an area under the curve (AUC) of 0.828 (95% confidence interval [CI]: 0.735-0.897), and MLP achieved 0.822 (95% CI: 0.732-0.895). Both models demonstrated good calibration, with Hosmer-Lemeshow test P values of 0.551 (Brier score = 0.168) for LDA and 0.682 (Brier score = 0.172) for MLP, alongside high net clinical benefit. SHAP analysis revealed that clinical N stage dominated the LDA model, whereas CEA and radiomics features played more prominent roles in the MLP model. Friedman’s H-statistic identified five significant feature interactions. Subgroup analysis showed that MLP maintained stable performance (AUC 0.66-0.77), whereas LDA performed poorly in the lymph-node-negative (AUC = 0.50) and pancreatic head tumor (AUC = 0.62) subgroups. The clinical nomogram achieved an AUC of 0.788 in the validation cohort.

CONCLUSION: The integrated model enables accurate preoperative prediction of synchronous LM in PC. Rather than replacing pathological diagnosis, this tool is designed to assist clinical decision-making in the common dilemma of indeterminate subcentimeter liver lesions on CT: a low-risk score supports surveillance and spares unnecessary biopsy, whereas a high-risk score prompts confirmatory biopsy or a shift toward neoadjuvant systemic therapy. Both architectures exhibit robust discrimination and clinical utility; however, the nonlinear MLP model shows superior stability across clinical subgroups, offering a promising tool for individualized risk assessment and treatment planning.

PMID:42642322 | DOI:10.1016/j.acra.2026.08.028

Acad Radiol. 2026 Aug 26:S1076-6332(26)00588-X. doi: 10.1016/j.acra.2026.08.001. Online ahead of print.

ABSTRACT

BACKGROUND: Point-of-care ultrasound (PoCUS) is increasingly integrated into clinical practice and medical curriculum, with potential educational benefits beyond sonographic skill acquisition. This randomized study evaluated whether a short PoCUS training program could improve medical students’ ability to interpret cross-sectional images.

METHODS: We conducted a randomized post-test study at Université Paris Cité. Final-year medical students (n = 417) participated in a PoCUS course consisting of three half-day hands-on ultrasound practice on peers sessions. Students were randomized into a pretest or post-test group to complete a 37-item survey assessing five cognitive domains: visuospatial ability, anatomical knowledge, spatial orientation, anatomical relationships, and pathology recognition. Items covered three imaging modalities (ultrasound, computed tomography, and magnetic resonance imaging) and four anatomical regions. Scores were expressed as accuracies (%). Statistical comparisons used univariate models and multivariate analyses.

RESULTS: Post-test participants achieved significantly higher total scores than pre-test participants (median 76/100 vs. 73/100, P = 0.015), with improvements in anatomical knowledge (P = 0.013), spatial orientation (P = 0.005), and ultrasound-related items (P < 0.001). Enhanced performance was observed in cardiac anatomy questions (P = 0.005). A correlation was found between visuospatial ability and total performance (ρ = 0.17, P = 0.0006), and gender had a significant impact on scores, with male students performing better than female students (P < 0.05).

CONCLUSION: PoCUS training enhances medical students’ understanding of cross-sectional anatomy and multimodal image interpretation, with effects mediated by visuospatial abilities. It represents a valuable pedagogical approach in the context of the expanded role of imaging in medical education and clinical practice.

PMID:42642318 | DOI:10.1016/j.acra.2026.08.001

Injury. 2026 Aug 24:113626. doi: 10.1016/j.injury.2026.113626. Online ahead of print.

ABSTRACT

BACKGROUND: Older adults account for an increasing proportion of emergency department visits for mild traumatic brain injury (mTBI). Although age is incorporated as a high-risk criterion in major clinical decision rules, the operational impact of risk stratification on computed tomography (CT) utilization in very elderly patients remains uncertain.

METHODS: We conducted a retrospective cohort study including 3088 consecutive patients aged ≥ 14 years presenting with mTBI (Glasgow Coma Scale 13-15) between 2019 and 2022. The primary outcome was the presence of an acute intracranial lesion on head CT. Multivariable logistic regression models were developed to identify independent predictors. Model performance was assessed using optimism-corrected estimates of the area under the receiver operating characteristic curve (AUC), with 95% confidence intervals, after bootstrap internal validation. In patients aged ≥ 75 years, probability thresholds were selected to achieve predefined sensitivity targets (100% and 99%) to estimate potential reductions in CT utilization.

RESULTS: CT was performed in 2657 patients (86.0%), identifying 155 acute intracranial lesions (5.8%). Patients aged ≥ 75 years represented 54.7% of the cohort and underwent CT in 98.2% of cases. In this subgroup, the prediction model showed moderate discrimination (optimism-corrected AUC 0.693; 95% CI, 0.610-0.775). At a threshold ensuring 100% sensitivity, CT utilization could be reduced by only 1.5% (specificity 1.6%). Allowing sensitivity to decrease to 99.0% increased potential CT reduction to 10.6% (specificity 11.2%).

CONCLUSIONS: In patients aged ≥ 75 years with mTBI, multivariable risk stratification demonstrates moderate discrimination but minimal capacity to safely reduce CT imaging. High baseline risk and low tolerance for missed lesions substantially limit the operational benefit of rule-out strategies in this population.

PMID:42642284 | DOI:10.1016/j.injury.2026.113626

Int J Gynecol Cancer. 2026 Aug 7:104948. doi: 10.1016/j.ijgc.2026.104948. Online ahead of print.

ABSTRACT

OBJECTIVE: We aimed to evaluate the association between KELIM and achievement of complete cytoreduction (R0; no macroscopic residual disease) in patients with high-grade serous ovarian carcinoma and to determine whether its performance and candidate thresholds vary according to homologous recombination status.

METHODS: This single-center retrospective study included patients with high-grade serous ovarian carcinoma treated with neoadjuvant chemotherapy between January 2010 and December 2025. KELIM was calculated for each patient and categorized as favorable (≥1) or unfavorable (<1). Associations with residual disease (R0, no macroscopic residual disease; R1, residual tumor <1 cm; or R2, residual tumor ≥1 cm) were assessed. Receiver operating characteristic curves identified candidate KELIM cutoffs. We evaluated the association of KELIM and homologous recombination status with progression-free survival; survival outcomes were analyzed using Kaplan-Meier estimates and Cox proportional hazards models. Sub-group analyses were performed according to homologous recombination status.

RESULTS: Overall, 145 patients were included. Median age was 64 years (interquartile range 57 to 73), and median KELIM was 0.93 (interquartile range 0.91 to 1.02). Patients with favorable KELIM (n = 58) had significantly higher R0 rates than those with unfavorable KELIM (n = 87) (77.6% versus 47.1%, p =.0002). Receiver operating characteristic analysis demonstrated moderate predictive performance for R0 (area under the curve 0.668; candidate cutoff 0.995). Candidate KELIM thresholds were lower in homologous recombination-deficient tumors (0.74) and breast cancer gene mutated tumors (0.745) than in homologous recombination-proficient tumors (0.94). Favorable KELIM was not significantly associated with progression-free survival (hazard ratio 0.78, 95% confidence interval 0.54 to 1.13, p =. 20). Median progression-free survival was numerically longer among patients with favorable KELIM compared with unfavorable KELIM (16.6 vs 13.7 months), although this difference was not statistically significant.

CONCLUSIONS: KELIM was associated with complete cytoreduction in patients with high-grade serous ovarian carcinoma, with differences observed according to homologous recombination status. These findings highlight the need for further investigation of candidate KELIM thresholds across distinct molecular sub-types to determine whether subtype-specific approaches may improve its clinical applicability.

PMID:42642281 | DOI:10.1016/j.ijgc.2026.104948

Int J Gynecol Cancer. 2026 Aug 6:104893. doi: 10.1016/j.ijgc.2026.104893. Online ahead of print.

ABSTRACT

BACKGROUND: Folate receptor-α is an established therapeutic target in ovarian cancer and identifies candidates for mirvetuximab soravtansine. However, co-expression patterns between folate receptor-α and other actionable biomarkers, and their therapeutic implications, remain incompletely characterized.

METHODS: We conducted a single-institution cohort study of 110 patients with platinum-resistant ovarian cancer treated with mirvetuximab soravtansine (2019-2025). Tumor biomarkers were assessed using immunohistochemistry, fluorescence in situ hybridization, and next-generation sequencing. Co-expression among 10 biomarkers, including folate receptor-alpha, estrogen receptor, progesterone receptor, human epidermal growth factor receptor-2, BRCA, homologous recombination/loss of heterozygosity, programmed death-ligand 1, claudin-6, and tumor mutational burden, were evaluated using Jaccard similarity indices, Cohen’s kappa statistics, and correlation analyses. Exploratory progression-free survival and overall survival analyses were performed by folate receptor-α and estrogen receptor sub-groups.

RESULTS: Folate receptor-α expression was high in 80% of tumors and intermediate/low in 20%. Co-expression analyses demonstrated largely independent biomarker patterns, reflecting biological heterogeneity. A hormone receptor-associated cluster emerged, including strong co-expression between folate receptor-α and estrogen receptor (Jaccard = 0.76), and moderate overlap between estrogen receptor and claudin-6 (0.62), programmed death-ligand-1 and estrogen receptor (0.56), and programmed death-ligand-1 and folate receptor-α (0.50). Survival outcomes were comparable between folate receptor-α-high and folate receptor-α-intermediate/low tumors. Among folate receptor-α-high tumors, estrogen receptor expression (≥10%) did not meaningfully stratify progression-free or overall survival outcomes.

CONCLUSIONS: Biomarker co-expression in advanced ovarian cancer demonstrates substantial heterogeneity, with patterns that may inform future biomarker-guided treatment combinations and sequencing strategies. Prospective validation is needed to refine multi-marker frameworks and determine whether biologically driven treatment approaches based on co-expression biomarker patterns may have clinical utility. Exploratory analyses support further investigation of folate receptor-α as a potentially graded rather than strictly binary biomarker; however, these findings are hypothesis-generating.

PMID:42642279 | DOI:10.1016/j.ijgc.2026.104893

Lancet Digit Health. 2026 Aug 25:101039. doi: 10.1016/j.landig.2026.101039. Online ahead of print.

ABSTRACT

BACKGROUND: Electroencephalogram (EEG) interpretation is essential for neurological diagnosis, but expert interpretation is limited globally, and existing AI methods address narrow tasks. This study aimed to develop and externally validate a broadly applicable foundation model capable of expert-level performance across diverse EEG tasks and clinical settings.

METHODS: In this multicentre study we developed a multidomain omnibus for reading and generalising over thorough EEG interpretation (MORGOTH), a foundation model that supports broad clinical interpretation across all major settings. We developed MORGOTH using EEGs from 18 677 patients across Massachusetts General Hospital, Brigham and Women’s Hospital, Beth Israel Deaconess Medical Center, and Boston Children’s Hospital, collected between Jan 1, 2003, and Feb 1, 2025, and validated it internally on 13 334 patients and externally on 1573 patients from 48 institutions, spanning diverse clinical settings and ages (0 years to >90 years). Test datasets annotated by six to 30 experts enabled inter-rater reliability (IRR) analysis by comparing model-expert and expert-expert agreement. MORGOTH was compared against both human experts and state-of-the-art models using area under the curve (AUC) and the percentage of experts’ operating points under the curve (EUC) for receiver operating characteristic (ROC) and precision-recall curves, as well as IRR and statistical calibration.

FINDINGS: MORGOTH achieved expert-level performance with AUC-ROC scores of 0·86 to 0·98 across 17 EEG findings. MORGOTH outperformed at least 90% of experts on three of seven multi-expert-annotated datasets and exceeded at least 20% of experts on each of the 17 tasks. Event-level performance was especially strong for seizure and ictal-interictal-injury continuum detection (EUC=96·6%) and spike detection (EUC=100%). IRR analysis showed that MORGOTH matched or exceeded expert consensus. External validation confirmed consistent performance with modest declines from internal to external test sets (event-level AUC -1·21%, EUC -3·33%; EEG-level AUC -2·12%, and EUC -9·52%). MORGOTH performance also remained robust across age, sex, and moderate channel loss, with lower age sensitivity (20·90% vs 30·90%) and fewer sex-related differences (33·33% vs 44·00%) than SPaRCNet.

INTERPRETATION: MORGOTH advances automated EEG interpretation with expert-level performance across clinical settings, offering improved diagnostic accuracy in low-resource environments and greater efficiency in high-volume centres.

FUNDING: US National Institutes of Health.

PMID:42642264 | DOI:10.1016/j.landig.2026.101039

J Voice. 2026 Aug 26:S0892-1997(26)00420-0. doi: 10.1016/j.jvoice.2026.08.004. Online ahead of print.

ABSTRACT

Professional singers represent a uniquely vulnerable population whose livelihoods depend on optimal vocal function, yet the specific impact of COVID-19 infection on their voice health remains insufficiently characterized. We conducted a cross-sectional observational study through an anonymous online survey distributed via Wenjuanxing among professional singers in mainland China between January and June 2025. A total of 154 participants (mean age 35.8 years, 59.2% female) with at least 2 years of professional singing experience and confirmed COVID-19 infection completed a structured questionnaire that collected demographic information, infection details, self-reported vocal symptoms, and standardized voice handicap assessments using the Mandarin Chinese version of the Voice Handicap Index-10 (VHI-10). Statistical associations were analyzed using prevalence ratios with 95% confidence intervals calculated via generalized linear models. Vocal alterations were reported by 62% of singers during the acute phase and by 53% after recovery, while new-onset throat clearing emerged in 44% of participants exclusively in the postinfection period. Self-perceived voice impairment was present in 65.6% of participants, with a mean VHI-10 score of 14.8 (SD = 7.2). COVID-19 infection was significantly associated with vocal impairment (prevalence ratio = 2.33, 95% CI: 1.04-5.46, P = 0.03) and throat clearing (prevalence ratio = 9.13, 95% CI: 2.47-64.42, P = 0.001). After adjustment for age, gender, and years of professional experience, these associations remained clinically meaningful at 2.29 and 8.97, respectively. Correlation analysis revealed strong positive relationships between symptom severity and VHI-10 scores (r = 0.68 for voice impairment, r = 0.74 for throat clearing). Our findings demonstrate that COVID-19 infection more than doubles the risk of voice impairment and increases the likelihood of new-onset throat clearing nearly ninefold among professional singers. This study provides novel evidence from a Chinese cohort, highlighting the critical need for targeted voice care protocols and routine postinfection voice assessment in this high-demand occupational group.

PMID:42642245 | DOI:10.1016/j.jvoice.2026.08.004