Bootstrap Resampling Research Articles

Optimizing prostate biopsy decision-making for patients with Prostate Imaging-Reporting and Data System (PI-RADS) ≥3 lesions: novel magnetic resonance imaging (MRI)-based nomograms.

Current protocols endorse biopsies for men with Prostate Imaging-Reporting and Data System (PI-RADS v2.1) scores ≥3. However, the subjective nature of PI-RADS can lead to increased false positives and unnecessary biopsies. Synthetic magnetic resonance imaging (MRI), which quantifies multiple relaxation parameters, and apparent diffusion coefficient (ADC), which is the most commonly used quantitative metric, have not yet been combined with a predictive tool. This study aimed to develop and validate novel nomograms using multiparametric MRI, including synthetic MRI, to forecast the risk of prostate cancer (PCa) and clinically significant prostate cancer (csPCa), and to assess the potential of these nomograms to reduce unnecessary biopsies in PI-RADS ≥3 cases. Between August 2020 and August 2022, 323 patients suspected of PCa were enrolled from two centers (cohort 1: 243; cohort 2: 80). All participants underwent multiparametric MRI, including synthetic MRI, before targeted biopsy. Univariable and multivariable logistic regression identified risk factors for PCa and csPCa. Internal validation was conducted using bootstrap resampling, and nomogram performance was evaluated through receiver operating characteristic (ROC) curve analysis, calibration plots, and decision curve analysis (DCA). External validation was performed with cohort 2 data. The impact of the nomograms on biopsy decisions was measured by the avoidance rate and the risk of missed diagnoses. The predictive nomogram for PCa incorporated four risk factors: age, quantitative transverse relaxation time (T2 value) from synthetic MRI, ADC value, and PI-RADS score. The csPCa nomogram included age, ADC value, and PI-RADS score. The nomograms showed high diagnostic accuracy with the area under the curves (AUCs) of 0.916 [95% confidence interval (CI): 0.901-0.974] and 0.947 (95% CI: 0.900-0.994) for PCa prediction in training and external datasets, and 0.884 (95% CI: 0.840-0.928) and 0.935 (95% CI: 0.871-0.998) for csPCa. Calibration curves confirmed the accuracy of predictions. DCA indicated that the nomograms possessed significant net benefit. For PCa detection, biopsy strategy combining our nomogram reduced biopsy procedures by 20.2% and 13.8% in the training and external cohorts, respectively, with a PCa miss rate of 4.5% for both cohorts. The csPCa-targeted biopsy strategy also provided clinical benefits, with biopsy avoidance rates of 20.2% and 10.0%, and csPCa miss rates of 4.8% and 1.7% for PI-RADS ≥3 patients in the two cohorts. The nomograms integrating multiparametric MRI and synthetic MRI are highly effective in predicting PCa and csPCa, concurrently, reducing unnecessary biopsies for patients with PI-RADS ≥3 lesion.

Molecular correlates of neural health associate with cardiovascular states

AbstractBackgroundSuboptimal cardiovascular health poses a significant risk for dementia, yet biological links between cardiovascular health and brain function are poorly understood. We examined how plasma markers of astrocytic activation (GFAP), neuronal axon breakdown (NfL), and Alzheimer’s disease (AD) pathology (pTau181) relate to several behavioral and risk indicators of systemic cardiovascular health in older adults along the AD continuum.Method209 older adults (59% female; 88% clinically normal; 15% amyloid PET+) completed 30‐day FitbitTM Flex2 monitoring (average daily steps), blood pressure (BP) and heart rate quantification, and plasma assayed for GFAP, NfL, and pTau181 (Quanterix Simoa). Regression models evaluated associations among indicators of cardiovascular health (pulse pressure [systolic BP‐diastolic BP], systolic BP, physical activity, resting heart rate) and plasma markers. Post‐hoc mediation models (1000 bootstrap resampling) evaluated mediational relationships among vascular health indicators and NfL. Models adjusted for age, sex, and BMI.ResultElevated pulse pressure associated with higher plasma pTau181 (β = 0.21, p<0.01), NfL (β = 0.19, p<0.01), and GFAP concentrations (β = 0.16, p = 0.01). Higher systolic BP associated with higher plasma pTau181 (β = 0.18, p = 0.03) and NfL concentrations (β = 0.19, p<0.01). Greater average daily steps associated with lower plasma NfL (β = ‐0.15, p = 0.04, but not pTau181 (β = ‐0.10, p = 0.27) or GFAP (β = ‐0.13, p = 0.057). Resting heart rate did not associate with plasma biomarker concentrations (ps>0.50). Given physical activity is a cardiovascular behavior that may support vessel health, we tested a post‐hoc mediation model examining if systolic BP or pulse pressure mediated the association between step count and NfL. Systolic BP, but not pulse pressure, fully mediated the relationship between daily steps and plasma NfL (Indirect Pathway: β = ‐0.04, p = 0.03; Direct Pathway: β = ‐0.11, p = 0.14).ConclusionElevated pulse pressure and systolic BP associate with higher levels of AD pathology, astrocytic activation, and axonal degeneration (pTau181, GFAP, NfL), while greater average daily steps associated with lower levels of neurodegeneration (NfL). Systolic BP may drive the association between physical activity and axonal degeneration, highlighting the importance of BP management in mitigating the effects of neurodegeneration. Studying links among cardiovascular health indicators and molecular brain outcomes could inform precise intervention targets for optimal brain aging, such as promoting the importance of blood pressure control and physical activity engagement.

Differential diagnosis of ovarian endometriosis cyst versus ovarian cystadenoma based on serum lactate dehydrogenase combined with CA-125 and CA19-9: A retrospective cohort study.

This study aims to construct and validate a nomogram for the differential diagnosis of ovarian endometriosis cyst versus ovarian cystadenoma. We retrospectively studied the clinical characteristics of patients with ovarian endometriosis cysts and ovarian cystadenomas from January 1, 2021, to June 1, 2022. Independent risk factors for differential diagnosis were investigated using univariate and multivariate logistic regression analyses. Based on these factors, a differential diagnosis of ovarian endometriosis cyst versus ovarian cystadenoma was established. The performance of the nomogram model was assessed by internal validation using bootstrapping resampling. Decision curve analysis (DCA) was performed to evaluate the net clinical benefit of the model. Immunohistochemistry showed that lactate dehydrogenase (LDH) A was overexpressed in ectopic endometrial tissues compared to that in normal endometrial tissues. In multivariate analysis, LDH, CA-125, and CA19-9 were identified as independent risk factors for the differential diagnosis of ovarian endometriosis cyst versus ovarian cystadenoma. LDH levels >135.50 U/L combined with CA-125 levels >25.20 U/mL and CA19-9 levels >13.59 U/mL as single covariates had a high value in the differential diagnosis of ovarian endometriosis cysts versus ovarian cystadenoma. The area under the receiver operating characteristic curve (ROC) of the nomogram constructed using LDH, CA-125, and CA19-9 expression data was 0.873 (95% CI, 0.827-0.920), and the bootstrap-validated concordance index (C-index) was 0.871. Decision curve analysis confirmed that the nomogram model had excellent clinical utility. Based on serum lactate dehydrogenase combined with CA-125 and CA19-9, we constructed and validated a nomogram for the differential diagnosis of ovarian endometriosis cyst versus ovarian cystadenoma to help physicians formulate the optimal treatment strategy.

Enhancing Transcriptomic Insights into Neurological Disorders Through the Comparative Analysis of Shapley Values.

Neurological disorders such as Autism Spectrum Disorder (ASD), Schizophrenia (SCH), Bipolar Disorder (BD), and Major Depressive Disorder (MDD) affect millions of people worldwide, yet their molecular mechanisms remain poorly understood. This study describes the application of the Comparative Analysis of Shapley values (CASh) to transcriptomic data from nine datasets associated with these complex disorders, demonstrating its effectiveness in identifying differentially expressed genes (DEGs). CASh, which combines Game Theory with Bootstrap resampling, offers a robust alternative to traditional statistical methods by assessing the contribution of each gene in the broader context of the complete dataset. Unlike conventional approaches, CASh is highly effective at detecting subtle but meaningful molecular patterns that are often missed. These findings highlight the potential of CASh to enhance the precision of transcriptomic analysis, providing a deeper understanding of the molecular mechanisms underlying these disorders and establishing a solid basis to improve diagnostic techniques and developing more targeted therapeutic interventions.

Impact of intensity standardisation and ComBat batch size on clinical-radiomic prognostic models performance in a multi-centre study of patients with glioblastoma.

To assess the effect of different intensity standardisation techniques (ISTs) and ComBat batch sizes on radiomics survival model performance and stability in a heterogenous, multi-centre cohort of patients with glioblastoma (GBM). Multi-centre pre-operative MRI acquired between 2014 and 2020 in patients with IDH-wildtype unifocal WHO grade 4 GBM were retrospectively evaluated. WhiteStripe (WS), Nyul histogram matching (HM), and Z-score (ZS) ISTs were applied before radiomic feature (RF) extraction. RFs were realigned using ComBat and minimum batch size (MBS) of 5, 10, or 15 patients. Cox proportional hazards models for overall survival (OS) prediction were produced using five different selection strategies and the impact of IST and MBS was evaluated using bootstrapping. Calibration, discrimination, relative explained variation, and model fit were assessed. Instability was evaluated using 95% confidence intervals (95% CIs), feature selection frequency and calibration curves across the bootstrap resamples. One hundred ninety-five patients were included. Median OS = 13 (95% CI: 12-14) months. Twelve to fourteen unique MRI protocols were used per MRI sequence. HM and WS produced the highest relative increase in model discrimination, explained variation and model fit but IST choice did not greatly impact on stability, nor calibration. Larger ComBat batches improved discrimination, model fit, and explained variation but higher MBS (reduced sample size) reduced stability (across all performance metrics) and reduced calibration accuracy. Heterogenous, real-world GBM data poses a challenge to the reproducibility of radiomics. ComBat generally improved model performance as MBS increased but reduced stability and calibration. HM and WS tended to improve model performance. Question ComBat harmonisation of RFs and intensity standardisation of MRI have not been thoroughly evaluated in multicentre, heterogeneous GBM data. Findings The addition of ComBat and ISTs can improve discrimination, relative model fit, and explained variance but degrades the calibration and stability of survival models. Clinical relevance Radiomics risk prediction models in real-world, multicentre contexts could be improved by ComBat and ISTs, however, this degrades calibration and prediction stability and this must be thoroughly investigated before patients can be accurately separated into different risk groups.

Recycling Space Beverage Packaging into LDPE-Based Composite Materials

Long-term space missions require careful resource management and recycling strategies to overcome the limitations of resupply missions. In this study, we investigated the potential to recycle space beverage packaging, which is typically made of low-density polyethylene (LDPE) and PET-aluminum-LDPE (PAL) trilaminate, by developing a LDPE-based composite material with PAL inclusions. Due to the limited availability of space beverage packaging, we replaced it with LDPE powder and commercial coffee packaging for the experiments. Fourier transform infrared spectroscopy (FTIR) was employed to thoroughly analyze the composition of the commercial coffee packaging. The simulant packaging was reduced to a filler, and its thermal properties were characterized by differential scanning calorimetry (DSC), while the particle size was analyzed via scanning electron microscopy (SEM) and the bootstrap resampling technique. Composite specimens were then fabricated by incorporating the filler into the LDPE matrix at loadings of 5 wt% and 10 wt%, and their mechanical and thermal properties were assessed through dynamic mechanical analysis (DMA) and thermal conductivity measurements. The 10 wt% corresponds approximately to the radio between PAL and PE in space beverage packaging and is, therefore, the maximum usable percentage when considering a single package. The results indicate that, as the filler loading increased, the mechanical performance of the composite material decreased, while the thermal conductivity was significantly improved. Finally, 10 wt% LDPE/PAL filaments, with a diameter of 1.7 mm and suitable for the fused filament technique, were produced.

Computational fluid dynamics and shape analysis enhance aneurysm rupture risk stratification

PurposeAccurately quantifying the rupture risk of unruptured intracranial aneurysms (UIAs) is crucial for guiding treatment decisions and remains an unmet clinical challenge. Computational Flow Dynamics and morphological measurements have been shown to differ between ruptured and unruptured aneurysms. It is not clear if these provide any additional information above routinely available clinical observations or not. Therefore, this study investigates whether incorporating image-derived features into the established PHASES score can improve the classification of aneurysm rupture status.MethodsA cross-sectional dataset of 170 patients (78 with ruptured aneurysm) was used. Computational fluid dynamics (CFD) and shape analysis were performed on patients’ images to extract additional features. These derived features were combined with PHASES variables to develop five ridge constrained logistic regression models for classifying the aneurysm rupture status. Correlation analysis and principal component analysis were employed for image-derived feature reduction. The dataset was split into training and validation subsets, and a ten-fold cross validation strategy with grid search optimisation and bootstrap resampling was adopted for determining the models’ coefficients. Models’ performances were evaluated using the area under the receiver operating characteristic curve (AUC).ResultsThe logistic regression model based solely on PHASES achieved AUC of 0.63. All models incorporating derived features from CFD and shape analysis demonstrated improved performance, reaching an AUC of 0.71. Non-sphericity index (shape variable) and maximum oscillatory shear index (CFD variable) were the strongest predictors of a ruptured status.ConclusionThis study demonstrates the benefits of integrating image-based fluid dynamics and shape analysis with clinical data for improving the classification accuracy of aneurysm rupture status. Further evaluation using longitudinal data is needed to assess the potential for clinical integration.

Clinical characteristics and prognosis analysis of pseudomonas aeruginosa bloodstream infection in adults: a retrospective study

Pseudomonas aeruginosa bloodstream infections (PA BSIs) in adults, especially those complicated by sepsis, are associated with high rates of morbidity and mortality. Early identification of risk factors for both mortality and sepsis-induced coagulopathy (SIC) is critical to optimizing patient management and improving outcomes. We conducted a retrospective analysis of 118 adult patients diagnosed with PA BSIs at the Affiliated Hospital of Xuzhou Medical University from January 2022 to February 2024. Univariate analysis was employed to identify significant clinical factors, followed by multivariate stepwise logistic regression to determine independent predictors of mortality and SIC. Based on these findings, nomogram models were constructed and evaluated using the area under the receiver operating characteristic curve (AUC), Bootstrap resampling, and calibration plots to assess model performance. Empiric sensitive antibiotic therapy (ESAT) (OR = 0.039, P < 0.001), coronary artery disease (CAD) (OR = 10.315, P = 0.010), and invasive mechanical ventilation (OR = 3.926, P = 0.020) emerged as significant predictors of mortality. In contrast, elevated C-reactive protein (CRP) (OR = 1.011, P = 0.003), procalcitonin (PCT) (OR = 1.030, P = 0.005), and lower hemoglobin levels (OR = 0.963, P = 0.004) were independently associated with SIC. The AUC of mortality prediction model is 0.908, while the SIC prediction model yielded an AUC of 0.817. The predictive models developed in this study demonstrate early identification of mortality rates and SIC risk in PA BSI patients, which may have the potential to guide timely therapeutic interventions and improve clinical outcomes in this high-risk population.

Demonstrating almost half of cotton fiber quality variation is attributed to climate change using a hybrid machine learning-enabled approach

Understanding the effects of climate change on cotton fiber quality will reduce the risks to production caused by global warming. Machine learning algorithms are effective for forecasting climate impacts on crops. However, the impact of climate change on cotton fiber quality is unclear. Hence, a hybrid machine learning-enabled approach, the Bayesian model average (BMA) method with multiple machine learning algorithms (linear regressor, SVR, RFR, GBDT, LightGBM, and XGBoost) and bootstrap resampling, was developed to explore the impact and screen the important climatic factors affecting various traits of fiber quality. On the basis of fiber quality data from 1033 test stations across Xinjiang, China, from 2016 to 2022, the explained variance for climate change in the hybrid machine learning model was as follows: 44.72 %–50.55 % for white cotton grade, 44.06 %–53.95 % for length, 51.72 %–56.81 % for micronaire, 32.70 %–49.50 % for length uniformity, and 45.66 %–53.09 % for strength in the 1000 bootstrapping samples. In addition, recursive feature elimination with cross-validation (RFECV) was used to select the optimal feature set and calculate the contribution of each feature. The variability in micronaire in the hybrid model was affected primarily by climate factors, such as the daily minimum temperature, rainfall, and wind speed, whereas the other quality traits were affected mainly by radiation-related climatic indicators. The climate during the harvest stage in October had a significant effect on cotton quality, explaining 33.0 % of the variance in white cotton grade, 32.1 % in length, and 48.3 % in fiber strength. Conversely, the climate during the boll opening and early harvest stages in September had a greater influence on micronaire and length uniformity, accounting for 21.4 % and 37.2 % of the variance, respectively. This study highlights that climate change explains nearly 50 % of the variation in fiber quality, with the influence being notably more considerable during the later stages of the cotton growth period. These findings clarify the uncertainty in the impact of climate change on cotton fiber quality considering the uncertainty of the single machine model and model errors. Equally important, this information can be valuable for farmers and growers seeking to improve fiber quality under climate change.

Development and validation of a risk prediction model for postpartum urinary incontinence in primiparas with singleton pregnancies: a multicenter clinical investigation.

Postpartum urinary incontinence (UI) is a serious condition that significantly affects the quality of life. Several studies have demonstrated that it is associated with pelvic floor dysfunction. This study aimed to develop and validate a UI risk prediction model to identify primiparas with singleton pregnancies at high risk. A multistage stratified random sampling process was used. UI was measured using the International Standard Consultation on Incontinence Questionnaire Form (a modified Bristol questionnaire, ICIQ-FLUTS). Records of 1,340 primiparas with singleton pregnancies were reviewed, and data were collected from January 2014 to December 2014 in multiple centers. A univariate logistic regression analysis was performed, followed by a multivariable logistic regression analysis of the data. Using bootstrap resampling, we constructed a nomogram to assess postpartum UI risk. A total of 1,340 patients were enrolled, including 345 with postpartum UI and 995 with non-postpartum UI. The occurrence of postpartum UI was significantly related to the mode of delivery, family history of UI, coffee or tea consumption, antenatal UI, and frequent cough. The nomogram exhibited good discriminatory ability with a C-index of 0.718 (95% confidence interval: 0.684-0.752) and a bootstrap-corrected C-index of 0.716. Additionally, the calibration curve demonstrated that the predicted outcomes aligned well with the actual observations. Ultimately, the decision curve analysis indicated that the nomogram exhibited favorable clinical applicability. The decision curve analysis suggests that the nomogram could provide clinical value. The clinician will then feel more confident about making clinical recommendations regarding postpartum UI screening for primiparous women with singleton pregnancies.

Cost-effectiveness of ultrahigh-resolution photon-counting detector coronary CT angiography for the evaluation of stable chest pain

BackgroundThe increased specificity of ultrahigh-resolution (UHR) photon-counting detector (PCD)-CT over energy-integrating detector (EID)-CT for coronary CT angiography (CCTA) could defer unwarranted downstream tests. The objective of the study was to simulate the cost-effectiveness of UHR CCTA in stable chest pain patients with coronary calcifications. MethodsA decision and simulation model was developed using Monte Carlo simulations with 1000 bootstrap resamples to estimate the costs associated with PCD-CT in lieu of EID-CT for CCTA and the referral for subsequent testing. The model was constructed using the diagnostic accuracy metrics of 55 coronary lesions in patients who underwent CCTA on both CT systems and subsequent invasive coronary angiography (ICA). Sensitivity and specificity were defined for each Coronary Artery Disease Reporting and Data System category. The aggregate healthcare expenditures were derived from the hospital billing system. ResultsAssuming a projected cohort of 15,000 patients over the lifetime of the PCD-CT, its implementation resulted in a 18.9 ​% reduction in the number of functional follow-up tests (6330.3 ​± ​59.5 vs. 5135.7 ​± ​60.6, p ​< ​0.001), a 6.0 ​% reduction in performed ICAs (1447.7 ​± ​36.2 vs. 1360.2 ​± ​34.7, p ​< ​0.001), and a 9.4 ​% decrease in major procedure-related complications. Over a 10-year expected life expectancy, PCD-CT led to an average cost saving of $794.50 ​± ​18.50 per patient and an overall cost difference of $11,917,500 ​± ​4,350,169. ConclusionsPCD-CT has the potential to reduce the financial burden on healthcare systems and procedure-related complications for stable chest pain patients with coronary calcification when compared to EID-CT.

Development of a Nomogram Based on Transcriptional Signatures, IFN-γ Response and Neutrophils for Diagnosis of Tuberculosis.

Tuberculosis (TB) is a major global health threat and its diagnosis remains challenging. This study aimed to develop a nomogram that incorporated peripheral blood transcriptional signatures and other blood tests for the diagnosis of tuberculosis. Patients with TB, patients with other definite pulmonary diseases (OPD), individuals with latent tuberculosis infection (LTBI), and healthy controls (HC) were retrospectively enrolled between May 2017 and April 2018. The results of the interferon-γ release assay (IGRA) and blood counts were obtained from medical records, and the transcripts of 10 genes were detected using reverse transcription polymerase chain reaction (RT-PCR). Variable selection was performed using least absolute shrinkage and selection operator regression (LASSO) and multivariate logistic regression was performed for the optimal prediction model with backward direction. The model was displayed as a nomogram, and its performance was evaluated for discrimination ability, calibration ability, and clinical usefulness. Internal validation of the prediction model was conducted using bootstrap resampling. A total of 185 participants were enrolled, including 84 patients with TB and 101 controls. A prediction nomogram composed of IGRA, percentage of neutrophils, and expression levels of CD64, granzyme A (GZMA), and PR/SET domain 1 (PRDM1) was established. The nomogram demonstrated good discrimination, with an unadjusted area under the curve (AUC) of 0.914 (95% CI: 0.875-0.954) and a bootstrap-corrected AUC of 0.914 (95% CI: 0.874-0.947). With a cutoff value of 0.519, the sensitivity and specificity for discriminating PTB from controls were 0.81 and 0.871, respectively. The nomogram also showed good calibration with the Hosmer-Lemeshow test (P=0.58) and good clinical practicality displayed by the decision curve analysis. A nomogram composed of IGRA, percentage of neutrophils, and expression of CD64, GZMA, and PRDM1 was established. The nomogram demonstrated a sensitivity and specificity of 81% and 87%, respectively, for differentiating TB from controls.

Nomogram model for predicting medication adherence in patients with various mental disorders based on the Dryad database

ObjectiveTreatment compliance among psychiatric patients is related to disease outcomes. How to assess patient compliance remains a concern. Here, we established a predictive model for medication compliance in patients with...

Validation of lymphovascular invasion as a predictor of lymph-node invasion in squamous cell carcinoma of the penis.

The aim of this study was to validate lymphovascular invasion (LVI) as a predictor of lymph-node invasion (LNI) in squamous cell carcinoma of the penis (SCCP). Within the Surveillance, Epidemiology, and End Results database (2010-2020), we identified SCCP patients who underwent lymphadenectomy with known LVI status. Univariable logistic regression models (LRMs) addressed LNI. Harrell's concordance index (c-index) quantified accuracy after 2000 bootstrap resamples for internal validation. Multivariable LRMs included the most informative, statistically significant predictors. Subgroup analyses were repeated in organ-confined (T1b-T2) and non-organ confined (T3-T4) stages. Of 586 SCCP patients, 219 (37%) had LVI. LVI was associated with higher rate of LNI (66 vs. 43%; P<0.001). Positive predictive value of LVI was 66 vs. 57% for negative predictive value. In multivariable LRMs, LVI independently predicted LNI (Odds ratio [OR]: 2.41; P<0.001). Bootstrap-adjusted c-index of multivariable model was 0.570 without LVI vs. 0.639 with LVI. In subgroup analyses, LVI independently predicted LNI in organ-confined (OR: 2.23; P<0.001) and in non-organ confined stages (OR: 3.10; P<0.001). In subgroup analyses, addition of LVI increased c-index from 0.530 to 0.595 in organ-confined and from 0.599 to 0.682 in non-organ confined. The current study validates LVI as an independent predictor of LNI in SCCP. LVI increases the accuracy of LNI predictions in the overall cohort as well as in organ-confined and non-organ confined stages. However, stage and grade even with the added consideration of LVI are not accurate enough to provide LNI prediction in individual patients.

A nomogram for predicting neurogenic lower urinary tract dysfunction in patients with acute ischemic stroke: A retrospective study.

To investigate the risk factors for neurogenic lower urinary tract dysfunction (NLUTD) in patients with acute ischemic stroke (AIS), and develop an internally validated predictive nomogram. The study aims to offer insights for preventing AIS-NLUTD. We conducted a retrospective study on AIS patients in a Shenzhen Hospital from June 2021 to February 2023, categorizing them into non-NLUTD and NLUTD groups. The bivariate analysis identified factors for AIS-NLUTD (p < 0.05), integrated into a least absolute shrinkage and selection operator (LASSO) regression model. Significant variables from LASSO were used in a multivariate logistic regression for the predictive model, resulting in a nomogram. Nomogram performance and clinical utility were evaluated through receiver operating characteristic curves, calibration curves, decision curve analysis (DCA), and clinical impact curve (CIC). Internal validation used 1000 bootstrap resamplings. A total of 373 participants were included in this study, with an NLUTD incidence rate of 17.7% (66/373). NIHSS score (OR = 1.254), pneumonia (OR = 6.631), GLU (OR = 1.240), HGB (OR = 0.970), and hCRP (OR = 1.021) were used to construct a predictive model for NLUTD in AIS patients. The model exhibited good performance (AUC = 0.899, calibration curve p = 0.953). Internal validation of the model demonstrated strong discrimination and calibration abilities (AUC = 0.898). Results from DCA and CIC curves indicated that the prediction model had high clinical utility. We developed a predictive model for AIS-NLUTD and created a nomogram with strong predictive capabilities, assisting healthcare professionals in evaluating NLUTD risk among AIS patients and facilitating early intervention.

Asymptotic Results for the Estimation of the Quadratic Score of a Clustering

In cluster analysis one often finds several partitions of a data set using different clustering methods and algorithms set with a variety of hyperparameters and tunings. The number of clusters K is one of the most relevant of such hyperparameters. Cluster selection is the task of choosing the desired partitions. The Bootstrap Quadratic Scoring is a recently introduced method where the cluster selection is performed by optimizing a score attached to a partition that is based on the quadratic discriminant function. Previously, we proposed the estimation of this cluster score via bootstrap resampling and investigated the proposed estimator based on numerical experiments and real data applications. However, that earlier work did not provide theoretical guarantees. In this paper, we fill that gap. We study the asymptotic behavior of the scoring method and show that the proposed estimator converges to well-defined population counterparts.

Factors associated with low resilience among French university students during COVID-19 lockdown

Abstract Background COVID-19 and subsequent lockdowns had an impact on the mental health of students, who had already been identified as a vulnerable population prior to this pandemic. At the end of the first COVID-19 lockdown, 15% of students reported suffering from moderate to severe anxiety. Psychological resilience is defined as the ability to bounce back from a stressful event and has been shown to have a positive impact on mental health parameters. This study aimed to identify factors associated with low resilience in the student population of the Grand Est region of France, at the end of the first national containment related to COVID-19. Methods The PIMS-CoV19 study: A cross-sectional online survey was conducted among students at the University of Lorraine to collect sociodemographic data, living and learning conditions, as well as levels of resilience and perceived health. Bivariable and multivariable logistic regression analyses were performed. Model robustness was assessed using bootstrap resampling. Results Among 3708 students, 50.6% had normal resilience, while 37.3% reported low resilience. Female gender (OR = 2.1, 95% CI: 1.8-2.6) and low social support (OR = 1.7, 95% CI: 1.1-2.6) were the main risk factors for low resilience. Negative thoughts (OR = 1.4, 95% CI: 1.4-1.5), lower quality of relationships with people in general (OR = 1.5, 95% CI: 1.3-1.8), and studying arts, humanities, or languages (OR = 1.4, 95% CI: 1.0-1.8) were identified as risk factors. Increased age (OR = 0.9, 95% CI: 0.9-1.0) and flat sharing (OR = 0.6, 95% CI: 0.4-0.9) were protective factors against low resilience. Conclusions Resilience seems to be impacted primarily by internal and micro-environmental factors. Consolidating levels of individual resilience of at-risk populations by acting on these factors could be the key to improving their mental health. Key messages • Resilience seems to be impacted primarily by internal and micro-environmental factors. • Consolidating levels of individual resilience of at-risk populations by acting on these factors could be the key to improving their mental health.

The INFLUENCE 3.0 model: Updated predictions of locoregional recurrence and contralateral breast cancer, now also suitable for patients treated with neoadjuvant systemic therapy

BackgroundIndividual risk prediction of 5-year locoregional recurrence (LRR) and contralateral breast cancer (CBC) supports decisions regarding personalised surveillance. The previously developed INFLUENCE tool was rebuild, including a recent population and patients who received neoadjuvant systemic therapy (NST). MethodsWomen, surgically treated for nonmetastatic breast cancer, diagnosed between 2012 and 2016, were selected from the Netherlands Cancer Registry. Cox regression with restricted cubic splines was compared to Random Survival Forest (RSF) to predict five-year LRR and CBC risks. Separate models were developed for NST patients. Discrimination and calibration were assessed by 100x bootstrap resampling. ResultsIn the non-NST and NST group, 49,631 and 10,154 patients were included, respectively. Age, mode of detection, histology, sublocalisation, grade, pT, pN, hormonal receptor status ± endocrine treatment, HER2 status ± targeted treatment, surgery ± immediate reconstruction ± radiation therapy, and chemotherapy were significant predictors for LRR and/or CBC in non-NST patients. For NST patients this was similar, but excluding (y)pT and (y)pN status, and including presence of ductal carcinoma in situ, axillary lymph node dissection and pathologic complete response.For non-NST patients, the Cox and RSF models were integrated in the online tool with 5-year AUCs of 0.77 (95%CI:0.77–0.77) and 0.68 (95%CI:0.67–0.68)] for LRR and CBC prediction, respectively. For NST patients, the RSF model performed best (AUCs 0.77 (95%CI:0.76–0.78) and 0.73 (95%CI:0.69–0.76) for LRR and CBC, respectively). Regarding calibration, observed-predicted differences were all <1 %. ConclusionThis INFLUENCE 3.0 models showed moderate performance in LRR and CBC prediction. The models have been made available as online tool to enable clinical decision support regarding personalised follow-up.

A Global-in-Time Neural Network Approach to Dynamic Portfolio Optimization

ABSTRACT We discuss a neural network approach, which does not rely on dynamic programming techniques, to solve dynamic portfolio optimization problems subject to multiple investment constraints. The approach allows for objectives of a very general form encompassing both time-consistent and time-inconsistent objectives, as well as objectives requiring multi-level optimization. The number of parameters of the neural network remains independent of the number of portfolio rebalancing events. Compared to reinforcement learning, this technique avoids the computation of high-dimensional conditional expectations. The approach remains practical when considering large numbers of underlying assets, long investment time horizons or very frequent rebalancing events. We prove convergence of the numerical solution to the theoretical optimal solution of a large class of problems under fairly general conditions, and present ground truth analyses for a number of popular formulations, including mean-variance, mean-semi-variance, and mean-conditional value-at-risk problems. Numerical experiments show that if the investment objective functional is separable in the sense of dynamic programming, the correct time-consistent optimal investment strategy is recovered, otherwise we obtain the correct pre-commitment (time-inconsistent) investment strategy. This method is agnostic as to the underlying data generating assumptions, and results are illustrated using (i) parametric models for underlying asset returns, (ii) stationary block bootstrap resampling of empirical returns, and (iii) generative adversarial network (GAN)-generated synthetic asset returns.

Foundational Segmentation Models and Clinical Data Mining Enable Accurate Computer Vision for Lung Cancer.

This study aims to assess the effectiveness of integrating Segment Anything Model (SAM) and its variant MedSAM into the automated mining, object detection, and segmentation (MODS) methodology for developing robust lung cancer detection and segmentation models without post hoc labeling of training images. In a retrospective analysis, 10,000 chest computed tomography scans from patients with lung cancer were mined. Line measurement annotations were converted to bounding boxes, excluding boxes < 1cm or > 7cm. The You Only Look Once object detection architecture was used for teacher-student learning to label unannotated lesions on the training images. Subsequently, a final tumor detection model was trained and employed with SAM and MedSAM for tumor segmentation. Model performance was assessed on a manually annotated test dataset, with additional evaluations conducted on an external lung cancer dataset before and after detection model fine-tuning. Bootstrap resampling was used to calculate 95% confidence intervals. Data mining yielded 10,789 line annotations, resulting in 5403 training boxes. The baseline detection model achieved an internal F1 score of 0.847, improving to 0.860 after self-labeling. Tumor segmentation using the final detection model attained internal Dice similarity coefficients (DSCs) of 0.842 (SAM) and 0.822 (MedSAM). After fine-tuning, external validation showed an F1 of 0.832 and DSCs of 0.802 (SAM) and 0.804 (MedSAM). Integrating foundational segmentation models into the MODS framework results in high-performing lung cancer detection and segmentation models using only mined clinical data. Both SAM and MedSAM hold promise as foundational segmentation models for radiology images.