Category: Statistics

Eur Spine J. 2026 May 18. doi: 10.1007/s00586-026-09981-3. Online ahead of print.

ABSTRACT

OBJECTIVE: To assess whether preoperative spinal angiography with endovascular embolization followed by microsurgical resection is technically feasible and safe for hypervascular spinal tumors, and to examine how cervical, thoracic, and lumbar vascular anatomy influences feasibility.

METHODS: We retrospectively reviewed 25 patients treated from 2017 to 2025 at two centers (14 men, 11 women; mean age 53 years) with hypervascular tumors involving the cervical (n = 3), thoracic (n = 15), lumbar and sacral (n = 7) spine. All underwent preoperative angiography (DSA)-embolization (within the same procedure if technically possible). We recorded whether embolization could be safely performed and any embolization-related complications; we also graded angiographic devascularization (devascularization grade, DG) and operative estimated blood loss (EBL). Selective or super-selective embolization was performed when arterial anatomy allowed safe catheterization, and surgery followed within 24-48 h.

RESULTS: Embolization was feasible in 23/25 patients (92%): thoracic 15/15 (100%), lumbar 6/6 (100%), sacral 1/1 (100%), and cervical 1/3 (33%). Two cervical RCC (renal cell carcinoma) metastases were not embolized because multiple short feeders shared trunks with radiculomedullary (spinal cord-supplying) arteries, precluding safe super-selective access. Among embolized cases, super-selective feeder occlusion was performed in 5/23 (21.7%) and segmental-vessel embolization in 18/23 (78.3%) using precipitating hydrophobic injectable liquid (PHIL), coils, and n-butyl cyanoacrylate (NBCA). High-grade devascularization (DG2-3) was achieved in all embolized cases, and completion angiography showed no or minimal residual tumor blush. There were no permanent embolization-related neurological deficits (0/23); one transient ischemic event occurred (1/23, 4.3%; vertebrobasilar ischemia in a C2 aneurysmal bone cyst) with full recovery. Mean estimated blood loss was 650 mL (median 600; range 400-800 mL. Differences in mean EBL between subsets of RCC and non-RCC tumors were statistically unsignificant, Correlations between DG and EBL was not identified.

CONCLUSION: A coordinated endovascular-surgical pathway achieved high embolization feasibility with no permanent embolization-related neurological complications, and no postoperative neurological deterioration in this cohort. Feasibility limitations were concentrated in the cervical spine, where short, shared feeders involving spinal cord-supplying arteries can preclude safe embolization. Limitations include retrospective design, small cohort with few cervical cases, and heterogeneous tumors/embolic agents; prospective studies with standardized devascularization and functional outcome metrics are warranted.

PMID:42144462 | DOI:10.1007/s00586-026-09981-3

Eur J Pediatr. 2026 May 18;185(6):404. doi: 10.1007/s00431-026-07084-1.

ABSTRACT

The diagnosis of Clostridioides difficile infection (CDI) in children is challenging due to high rates of asymptomatic colonization and the limited ability of available assays to differentiate colonization from true infection. The relative utility of polymerase chain reaction (PCR) versus stool toxin A/B detection by enzyme immunoassay (EIA) remains uncertain in pediatric practice. This study aimed to evaluate whether stool toxin A/B detection adds incremental clinical value among children with gastrointestinal multiplex PCR positivity for C. difficile toxin genes (tcdA/tcdB) by comparing risk factors, treatment decisions, and outcomes according to toxin status. This retrospective observational study included 103 pediatric patients (< 18 years) with PCR positive for C. difficile toxin genes (tcdA/tcdB) results at a tertiary children’s hospital between October 2022 and April 2025. Patients were categorized into three groups: stool toxin A/B not evaluated, stool toxin A/B negative, and stool toxin A/B positive. Demographics, risk factors, clinical characteristics, laboratory findings, treatment decisions, and outcomes were compared across stool toxin A/B groups. Of the 103 patients, 26 (25.2%) had no stool toxin A/B, 63 (61.2%) were stool toxin A/B negative, and 14 (13.6%) were stool toxin A/B positive. More than half of the cohort (54.3%) received CDI treatment, with no significant difference in treatment initiation rates among stool toxin A/B groups. The risk factors-including underlying disease (e.g., malignancies, inflammatory bowel disease, and immunodeficiencies), recent hospitalization, antibiotic exposure, and proton pump inhibitor (PPI) or enteral tube use-were similarly distributed. Clinical severity, laboratory parameters, imaging findings, recurrence, complications (1.0%), ICU admission (4.9%), and mortality (3.9%) did not differ significantly between stool toxin A/B positive and negative patients. Although a statistically significant difference in diarrhea severity was observed across groups, this was driven by the toxin-not-evaluated group. Importantly, stool toxin A/B positivity did not correlate with clinical severity, underlying risk factors, laboratory abnormalities, or outcomes.

CONCLUSION: In this pediatric cohort, stool toxin A/B detection did not provide additional clinical value beyond PCR positivity for C. difficile toxin genes (tcdA/tcdB). PCR alone, when interpreted in the context of compatible symptoms and epidemiological risk factors, may be sufficient to guide treatment decisions for suspected pediatric CDI. Given the limited incremental value of stool toxin A/B testing, optimized diagnostic algorithms and further multicenter pediatric studies are warranted.

WHAT IS KNOWN: • Pediatric CDI diagnosis is complicated by high asymptomatic colonization; PCR for C. difficile toxin genes (tcdA/tcdB) is sensitive but cannot distinguish colonization from infection. • Stool toxin A/B EIAs are more specific but less sensitive, and guidelines often recommend multistep algorithms.

WHAT IS NEW: • In our cohort of pediatric patients with PCR positivity for C. difficile toxin genes (tcdA/tcdB), stool toxin A/B status was not associated with clinical severity, risk factors, laboratory findings, outcomes, or treatment initiation. • When clinical features are compatible, PCR positivity for C. difficile toxin genes (tcdA/tcdB) alone may be sufficient to guide treatment, with limited incremental value of stool toxin A/B testing.

PMID:42144455 | DOI:10.1007/s00431-026-07084-1

Sci Rep. 2026 May 17. doi: 10.1038/s41598-026-53149-4. Online ahead of print.

ABSTRACT

During pregnancy, mental health conditions are a major health concern across the globe. Common mental disorders have complex effects on the health of the mother and fetus, as well as long-lasting social and economic repercussions. Hence, this study aimed to assess the worldwide pooled prevalence and risk factors of common mental disorders in pregnant women. We searched published papers using major databases, including Embase, PubMed, PsycINFO, Web of Science, Google Scholar, and HINARI, for studies on the prevalence and risk factors of common mental disorders published up to January 2024. The research teams used the Newcastle-Ottawa Quality Assessment Scale to assess the quality of each study. Two independent researchers screened and extracted the data. The analysis conducted using STATA statistical software version 11. The global estimated pooled prevalence and risk factors of common mental disorders in pregnant women were assessed using a random effect model. The heterogeneity of the included studies was evaluated using the I² statistic. The researchers assessed the publication bias using the funnel plot and Egger’s statistical test. The studies comprised eighteen studies with 17,380 pregnant women. The global pooled prevalence of common mental disorders in pregnant women was 31.59% (95% CI: 23.74-39.43). Subgroup analysis estimated the prevalence of common mental disorders in Africa (30.30%; 95% CI: 20.95-39.65), Asia (22.96%; 95% CI: 13.12-32.78), and South America (40.30%; 95% CI: 23.92-59.15). The risk factors considered included a family history of mental illness, chronic medical conditions, intimate partner violence, unplanned pregnancy, emotional violence, and a history of abortion. However, only intimate partner violence (POR = 2.63; 95% CI: 1.12, 6.17) was found to be significantly associated with common mental disorders among pregnant women. This study reveals a high global burden of common mental disorders among pregnant women, with regional variations. Intimate partner violence was found to be a significant risk factor. To mitigate its impact, integrating routine mental health screening and intimate partner violence prevention should be made into maternal healthcare services.

PMID:42144451 | DOI:10.1038/s41598-026-53149-4

EJNMMI Phys. 2026 May 17. doi: 10.1186/s40658-026-00887-z. Online ahead of print.

ABSTRACT

PURPOSE: Lesion segmentation is important in targeted radionuclide therapy dosimetry. In this study, we proposed a novel 2D multi-modal Segment Anything Model (mmSAM) for lesion segmentation in whole-body multi-tracer PET/CT images.

METHODS: The AutoPET 2024 dataset, including ¹⁸F-PSMA (369 subjects), ¹⁸F-FDG (170 subjects) and ⁶⁸Ga-PSMA (170 subjects) PET/CT images and their corresponding tumor masks, was used in this study. The ¹⁸F-PSMA PET/CT dataset was used as the primary dataset and was divided into 233: 36: 100 for training, validation and testing the mmSAM, using both PET and CT input. The transferability of the primary model was evaluated on 100 patients on other 2 datasets, without and with fine-tuning using 70 cases of ¹⁸F-FDG and ⁶⁸Ga-PSMA respectively. Standard single modal 2D SAM with only PET input, 3D nnUNet with two-channel PET/CT input and the thresholding-based method were also implemented for comparison. Mean Dice, the 95th percentile Hausdorff distance (mean HD95), mean standardized uptake value (mean |SUV_mean| error), metabolic tumor volume (mean |MTV| error), true positive rate (TPR), positive predictive value (PPV), and false discovery rate (FDR) were computed for all tumors. Statistical significance among different segmentation methods was evaluated using the Wilcoxon test.

RESULTS: mmSAM achieved the best performance as compared to other segmentation methods for the primary ¹⁸F-PSMA dataset (mean Dice/HD95/|SUV_mean| error/|MTV| error/TPR/PPV/FDR = 0.76/1.92 mm/5.10%/14.60%/100%/96.51%/3.49%, all p < 0.05), without fine-tuning (mean Dice/HD95/|SUV_mean| error/|MTV| error/TPR/PPV/FDR = 0.61/2.28 mm/ 15.83%/27.71%/100%/78.91%/21.09% for ¹⁸F-FDG; 0.77/1.33 mm/5.55%/17.96%/100% /97.78%/2.22% for ⁶⁸Ga-PSMA, p < 0.05), as well as with fine-tuning (mean Dice/HD95/|SUV_mean| error/|MTV| error/TPR/PPV/FDR = 0.65/1.81 mm/6.62%/13.93%/100%/ 80.50%/19.50% for ¹⁸F-FDG; 0.81/1.15 mm/4.62%/14.30%/100%/99.30%/0.70% for ⁶⁸Ga-PSMA, p < 0.05) on the cross-tracer datasets.

CONCLUSION: The proposed mmSAM is promising for lesion segmentation in multi-tracer oncologic PET/CT images. Fine-tuning significantly enhances segmentation accuracy for cross-tracer studies.

PMID:42144442 | DOI:10.1186/s40658-026-00887-z

Skeletal Radiol. 2026 May 18. doi: 10.1007/s00256-026-05252-w. Online ahead of print.

ABSTRACT

OBJECTIVES: To assess current trends and impacts of hybrid (HYB) and work from home (WFH) practice models on the musculoskeletal (MSK) radiology workforce and explore implications for recruitment and the future of the subspecialty.

MATERIALS AND METHODS: A 38-question voluntary, anonymous survey was distributed to the Society of Skeletal Radiology (SSR) membership (n = 1060). Descriptive statistics and chi-squared test were performed to analyze demographics, trends, and preferences for HYB-WFH practice models. Thematic analysis of open-ended responses identified perceptions and impacts of HYB-WFH.

RESULTS: The survey response rate was 30% (314 respondents). Most (89%) work in practices that support HYB-WFH. Most (84%) felt HYB-WFH was extremely (54%) or somewhat important (30%) but would not consider the job if it meant reduced compensation (56%) or higher productivity requirements (52%). HYB was the most desired work model (84%), followed by WFH (13%) and in-person (4%). Preference for HYB-WFH was associated with geographic residence, job location (urban/suburban/rural), and commute time (p < 0.05), but not age, gender, dependents, caregiver status, or career stage. Perceived benefits included improved work-life balance, job satisfaction, and productivity, with possible tradeoffs in career development, collegiality, and trainee education. Among 59% planning to change jobs in the next 5 years, motivating factors included workload (66%), compensation (54%), and HYB-WFH (39%). Most preferred HYB for their next job.

CONCLUSION: HYB-WFH practice models are strongly desired by the MSK workforce, regardless of age, gender, and life and career situation. Embracing HYB-WFH trends is key to supporting workforce sustainability and adapting to an evolving job market.

PMID:42144439 | DOI:10.1007/s00256-026-05252-w

Sci Rep. 2026 May 17. doi: 10.1038/s41598-026-53357-y. Online ahead of print.

ABSTRACT

This research presents a HashCode-based security abstraction that implements privacy-by-design in customer churn statistics, safeguarding identity while maintaining analytical integrity. A set of machine learning and deep learning models, including Logistic Regression, Random Forest, XGBoost, and a Multilayer Perceptron (MLP), is ideal for assessing churn while maintaining stringent privacy standards. Experimental results show that models that merely use behavioral, transactional, and temporal aspects can participate with and balance each other. The SGD Logistic Regression has an accuracy of 77.5%, a precision of 0.759, a recall of 0.812, an F1-score of 0.785, and an AUC of 0.815. This shows that it is quite sensitive to churners. The Random Forest has an accuracy of 77.0%, a precision of 0.767, a recall of 0.782, an F1-score of 0.775, and an AUC of 0.794. This means that it can generalize well. XGBoost has an accuracy of 74.0%, a precision of 0.733, a recall of 0.762, an F1-score of 0.748, and an AUC of 0.769. This shows that aggressive boosting doesn’t work very well with this structured dataset. The MLP (Keras) has the best overall performance, with an accuracy of 80.0%, a precision of 0.802, a recall of 0.802, an F1-score of 0.802, and the highest AUC of 0.825. This shows that it is better at learning non-linear representations and balancing categorization. These results provide strong churn prediction without relying on rich identifiers or assumptions that need a lot of infrastructure, which is different from previous studies. The study offers a deployable, regulation-compliant architecture demonstrating that accurate churn prediction is attainable by behavior-driven analytics within a consistent security enforcement paradigm.

PMID:42144427 | DOI:10.1038/s41598-026-53357-y

Sci Rep. 2026 May 18. doi: 10.1038/s41598-026-48616-x. Online ahead of print.

ABSTRACT

Micronutrients play essential roles in metabolic regulation, neurotransmission, and immune function. Disturbances in elements such as zinc, iron, magnesium, and vitamin C have been linked to cognitive impairment, mood dysregulation, and altered immune responses in psychiatric populations. This study aimed to compare serum levels of vitamin C, iron, magnesium, and zinc in patients with mood and psychotic disorders and healthy controls, and to examine their associations with peripheral immune indices, including neutrophil and lymphocyte counts. In this case-control study, 60 psychiatric patients (mood and psychotic disorders) and 20 healthy controls were recruited from 5 Azar Hospital in Gorgan. Serum micronutrient levels were measured using biochemical methods, and complete blood counts were used to determine neutrophil and lymphocyte percentages. Statistical analyses, including group comparisons and correlation tests, were performed using SPSS version 23. The mean age of participants was 40.87 ± 12.58 years. Serum zinc levels were significantly higher in psychiatric patients compared with healthy controls (P < 0.001), with no significant differences between mood and psychotic disorder groups. Serum iron, magnesium, and vitamin C levels did not differ significantly among the three groups. Correlation analyses revealed significant associations between serum zinc and magnesium (r = 0.286, P = 0.027), zinc and iron (r = 0.366, P = 0.004), and iron with both WBC (r = 0.291, P = 0.024) and neutrophil counts (r = 0.313, P = 0.015). The findings suggest that micronutrient-immune interactions may contribute to the biological profile of mood and psychotic disorders. Although causal relationships cannot be inferred, routine assessment of key micronutrients and immune indices may provide additional insight into the metabolic and inflammatory status of psychiatric patients. Further research is warranted to clarify the mechanistic pathways linking micronutrient balance, immune function, and psychiatric symptomatology.

PMID:42144425 | DOI:10.1038/s41598-026-48616-x

Sci Rep. 2026 May 18. doi: 10.1038/s41598-026-51012-0. Online ahead of print.

ABSTRACT

Academic performance (AP) prediction is crucial for recognizing at-risk students and enhancing learning outcomes. Traditional statistical models often fail to capture temporal and behavioral patterns. Deep learning (DL) approaches offer improved accuracy and adaptability by leveraging multi-dimensional student data for predictive analysis. The objective is to advance a robust predictive model that predicts students’ AP using multi-dimensional data, integrating temporal, behavioral, and demographic features. Students’ learning performance data for n = 2000 is collected from multiple sources, including student grades, attendance, learning management system (LMS) interactions, psychometric surveys, and demographic records. Collected data undergoes preprocessing steps, including handling missing values using K Nearest Neighbor Imputation (KNNI), outlier removal, and normalization. Principal Component Analysis (PCA) is employed to decrease dimensionality and extract relevant characteristics from high-dimensional datasets. A novel Gated Long Short-Term Memory Unit is optimized with Dove (GateLSTMU-Dove) to capture temporal dependencies and student engagement patterns. GateLSTMU identifies time-dependent patterns in educational data to support accurate performance forecasting. Dove optimizes model parameters efficiently, enhancing convergence speed and predictive accuracy of the GatedLSTMU. Python 3.10-based experiments demonstrate the model’s superior performance. GateLSTMU-Dove achieved lower error metrics and higher classification accuracy (98.85%) compared to baseline methods. Visualization of predictions confirmed accurate forecasting and interpretable temporal patterns in student performance. The GateLSTMU-Dove effectively predicted academic outcomes using multi-dimensional student data. It provides interpretable insights, supports early intervention strategies, and demonstrates a scalable, reproducible approach for data-driven AP management.

PMID:42144405 | DOI:10.1038/s41598-026-51012-0

Sci Rep. 2026 May 17. doi: 10.1038/s41598-026-51033-9. Online ahead of print.

ABSTRACT

Quantum Random Number Generators (QRNGs) exploit intrinsic quantum mechanical randomness and are promising candidates for higher-quality randomness than classical RNGs. However, QRNGs on Noisy Intermediate-Scale Quantum (NISQ) platforms are constrained by decoherence, gate and measurement errors, and limited circuit depth, and the resulting impact on statistical randomness is not yet systematically characterized. This work introduces parameterized quantum random number generator (PQRNG) architectures built from parameterized quantum circuits (PQCs) to improve tunability and expressive power under realistic noise. We study three architectures, PQC-H-CH, H-PQC-CH, and H-CH-PQC, and evaluate their randomness under Transpiler optimization and circuit-level error mitigation as preprocessing steps, using NIST SP 800-22 for statistical validation. Utilizing default NIST settings, PQC-H-CH attains the largest number of fully passing preprocessing configurations (126 combinations passing all 15 tests), compared with 110 for H-PQC-CH and 69 for H-CH-PQC, indicating strong robustness across Transpiler optimization and mitigation settings, and we find that preprocessing choices influence randomness quality more strongly than adjusting NIST test parameters. Overall, these results demonstrate that the proposed approach provides a solid foundation for developing more reliable and practical QRNGs using PQCs in NISQ devices.

PMID:42144404 | DOI:10.1038/s41598-026-51033-9

Sci Rep. 2026 May 18. doi: 10.1038/s41598-026-42351-z. Online ahead of print.

ABSTRACT

Selenium nanoparticles (SeNPs) have gained increasing attention due to their favorable biological properties and potential applications in cancer research. In this study, the ability of Bacillus sp. strain STG-83 to biosynthesize SeNPs was systematically investigated. This study comprehensively investigated how Bacillus sp. strain STG-83 can biosynthesize SeNPs. Response surface methodology (RSM) was utilized to optimize the bioreduction of selenite and to identify the key process parameters that play a significant role. The developed quadratic model showed a strong correlation with experimental data (R²=0.92). Statistical analysis demonstrated that time and selenium concentration significantly affected selenite reduction efficiency (P < 0.05), whereas bacterial inoculum percentage were not significant (P > 0.05). Increasing selenium concentration from 0.5 to 25 mM reduced the bioreduction efficiency from 100% to 29.37%, while extending time from 8 to 96 h increased efficiency from 42.03% to 61.51%. The biosynthesized SeNPs were characterized using UV-Vis’s spectroscopy, FTIR, EDX, SEM, TEM, and XRD analyses. The nanoparticles were predominantly spherical, with sizes ranging from 80 to 140 nm, and were coated with a bioorganic surface layer. Biological evaluation revealed that SeNPs induced dose-dependent cytotoxicity in U-87 line while exerting lower toxicity toward normal fibroblast cells. Flow cytometry analysis further demonstrated a significant increase in intracellular reactive oxygen species (ROS) levels following SeNP exposure, suggesting that oxidative stress plays a central role in the observed anticancer effects. The ROS generation triggered by SeNPs suggests they might serve as effective radiosensitizing agents. Future studies that combine radiation and in vivo approaches should confirm this potential.

PMID:42144400 | DOI:10.1038/s41598-026-42351-z